International
Coqfetence on
Environ men tal
Sensing aqd
Assessment
1
^TtS O* *
^t0SX
£
\
as
% PRO^°
4
NASA
National
Aeronautics ai*id
Space
Administration
'IBP
AIAA
r.\ METEOROLOGICAL
SOCIETY
murW:
Volume 1
A Joint Conference comprising the
INTERNATIONAL SYMPOSIUM ON
ENVIRONMENTAL MONITORING
co-sponsored by
World Health Organization
U.S. Environmental Protection Agency
University of Nevada, Las Vegas
and the
THIRD JOINT CONFERENCE ON
SENSING OF ENVIRONMENTAL
POLLUTANTS
co-sponsored by
American Chemical Society
American Institute of Aeronautics
& Astronautics
American Meteorological Society
Institute of Electrical & Electronics
Engineers
Instrument Society of America
U.S. Department of Transportation
U.S. Environmental Protection Agency
National Aeronautics & Space
Administration
National Oceanic & Atmospheric
Administration of the Department
of Commerce
-------
J.S.
HHP
-------
Ii\tetnatioi\al CoqfeiSnce
on
Eijvirdnrqental Seqsing
aqd
Assessment
Volume 1
September 14—19, 1975
Las Vegas, Nevada
i
-------
.Aaafev,
S*ATli °
S&> sr4fr
w
NASA
National
Aerd aut'csanr)
Space
Atfirw iisiialion
A Joint Conference comprising the
INTERNATIONAL SYMPOSIUM ON
ENVIRONMENTAL MONITORING
co-sponsored by
World Health Organization
U. S. Environmental Protection Agency
University of Nevada, Las Vegas
and the
THIRD JOINT CONFERENCE ON
SENSING ENVIRONMENTAL POLLUTANTS
co-sponsored by
American Chemical Society
American Institute of Aeronautics & Astronautics
American Meteorological Society
Institute of Electrical & Electronics Engneers
Instrument Society of America
U. S. Department of Transportation
U. S. Environmental Protection Agency
National Aeronautics & Space Administration
National Oceanic & Atmospheric Administration
of the Department of Commerce
Library of Congress Card Catalogue #75^37494
Copyright 1976 by the
Institute of Electrical & Electronics Engineers, Inc.
345 East 47 Street, New York, NY 10017
IEEE Catalogue #75-CH 1004-1 ICESA
AIAA
AMERICAN
HlHOKOlOCKAl
S0CU1T
ii
-------
PREFACE
The International Conference on Environmental
Sensing and Assessment (ICESA) was held in Las Vegas,
Nevada, September 14-19, 1975. This meeting was co-
sponsored by the 3rd Joint Conference on Sensing of
Environmental Pollutants and the International Sympo-
sium on Environmental Monitoring. The 3rd Joint Con-
ference on Sensing of Environmental Pollutants was
sponsored by the American Chemical Society, the Ameri-
can Institute of Aeronautics and Astronautics, the
American Meteorological Society, the Institute of
Electrical and Electronics Engineers, the Instrument
Society of America, the United States Department of
Transportation, the United States Environmental Pro-
tection Agency, the National Aeronautics and Space
Administration and the National Oceanic and Atmospher-
ic Administration of the United States Department of
Commerce. The International Symposium on Environmental
Monitoring was sponsored by the United States Environ-
mental Protection Agency, the University of Nevada,
Las Vegas and the World Health Organization.
As a result of the efforts of representatives
from each of the sponsoring organizations and the en-
thusiastic response of the scientific community an ex-
ceptionally strong scientific program was developed.
The 241 papers selected for inclusion in the ICESA
program were chosen front over 450 abstracts which were
submitted by scientists from 26 countries for considera-
tion for presentation. Approximately 1000 people from
40 countries attended and participated in the ICESA
conference and 55 industrial, governmental and pro-
fessional organizations exhibited their products and
services.
As the technical program developed and the re-
sponse to this program became evident, the Coordina-
ting Committee of the ICESA determined that a confer-
ence record would be a valuable and important compo-
nent of the Information exchange resulting from this
meeting. In the interest of providing the ICESA at-
tendees and others involved in environmental monitor-
ing and assessment with a complete and timely confer-
ence proceedings, all speakers were requested to submit
by meeting date a manuscript covering the information
they were to present at the conference. It was the
opinion of the Coordinating Committee that the confer-
ence date deadline specified to authors would insure
as up-to-date and accurate a rendition of the presented'
material Ab possible. Also, the responsibility for
technical accuracy and clarity has been placed solely
upon the authors of manuscripts included in this
record. The ICESA conference organizers and sponsors
were only responsible for the selection of speakers
who made presentations at this conference.
Those involved with the ICESA are most proud of
the program presented and feel that the information
contained within this record will be of value for many
years to come. The response of the speakers to the
request for prepared papers was overwhelming, and
over 90% (225) of the authors making presentations at
this conference submitted completed manuscripts for
inclusion in this record. The credit for the success
of this conference must be given to a large number
of individuals who gave so willingly of their time
and talents. The individuals of the program committees
worked long and hard to develop and help organize the
various program elements. The session and panel
chairpersons must be credited with putting together
and efficiently running the individual- sessions. The
response of the environmental monitoring and assessment
scientific community to the call for papers, excellent
presentations and prompt submission of manuscripts
was overwhelming and contributed greatly to the pro-
gram success. Finally, and most importantly, the
ultimate success of the conference must be given to
the large number of attendees who participated so
effectively in the information and idea exchange
which took place in each of the scientific, panel
and plenary sessions.
S. H. Melfi, EPA; J. L. Moyers, ACS; G. Ozolins, WHO; E. A. Schuck, EPA
Coprogram Chairmen
iii
-------
INTERNATIONAL CONFERENCE ON
ENVIRONMENTAL SENSING
AND ASSESSMENT
Coordinating Chairperson, Dr. D. S. Barth, Director,
Environmental Monitoring and Support Laboratory, EPA,
Las Vegas, Nevada
General Chairpersons: General Vice Chairpersons:
Dr. D. S. Barth, EPA Dr. R. E. Stanley, EPA
Dr. 8. H. Dieterich, WHO Mr. B. H. Manhei-mer, HUD/IEEE
Dr. R. B. Smith, UNLV
Mr. G. B. Morgan, EPA
Program: J. L. Moyers, ACS; S. H. Melfi, EPA; E. A. Schuck,
EPA; G. Ozolins, WHO.
Local Arrangements: E. Emett, IEEE; R. Jaquish, EPA.
Publicity: H.J. Perlis, IEEE; G. S. Douglas, EPA.
Publication: L. Nagel, IEEE; D. Tate, EPA.
Exhibits: H. Thacke, IEEE.
Treasurer: K. McNeil, UNLV;0. Marsh, IEEE.
Staff Contacts: P. Edmonds, IEEE; G. S. Douglas, EPA
INTERNATIONAL SYMPOSIUM ON
ENVIRONMENTAL MONITORING
Chairpersons: Dr. B. H. Dieterich
Director, Division of Environmental Health
World Health Organization
Geneva, Switzerland
Dr. Robert B. Smith
Dean, College of Science, Mathematics, and
Engineering
University of Nevada, Las Vegas
Las Vegas, Nevada
Mr. George B. Morgan
Director, Monitoring Systems Research &
Development Division
U. S. Environmental Protection Agency
Las Vegas, Nevada
Vice Chairperson:
R. E. Stanley, EPA, Las Vegas, Nevada
Program Committee Chairpersons.
S. H. Melfi, EPA, Las Vegas, Nevada
E. A. Schuck, EPA, Las Vegas, Nevada
G. Ozolins, World Health Organization,
Geneva, Switzerland
Members: I. J. Selikoff, Mt. Sinai School of Medicine,
New York, New York
H. G. Hanson, Pan American Health Org.,
Washington, DC
F. Clarke, USGS, Washington. DC
R. H. Langford, USGS, Reston, Virginia
J. E. Thompson, EPA, Research Triangle
Park, North Carolina
E. Robinson, Washington State University,
Pullman, Washington
Clayton Jensen, NOAA, Rockville, Maryland
J. R. McNesby, National Bureau of Standards
Washington, DC
Members: ]. N. Pitts, Jr., University of California,
Riverside, California
D. Ballinger, EPA, Cincinnati, Ohio
R,. E. Stanley, EPA, Las Vegas, Nevada
F. Butrico, Pan American Health Org.,
Washington, DC
Finance Committee Chairperson:
Keith McNeil, UNLV, Las Vegas, Nevada
Members: O. Berroyer, EPA, Las Vegas, Nevada
). Rock, EPA, Las Vegas, Nevada
Protocol Committee Chairperson:
M. Carpenter, EPA, Las Vegas, Nevada
Exhibits Committee Chairpersons:
J. Coogan, EPA, Las Vegas, Nevada
J. Hodgeson, EPA, Las Vegas, Nevada
Local Arrangements Committee Chairpersons:
R. Jaquish, EPA, Las Vegas, Nevada
]. Payne, EPA, Las Vegas, Nevada
Printing Committee Chairperson:
D. Tate, EPA, Las Vegas, Nevada
Members: B. K. Spavin, EPA, Las Vegas, Nevada
J.J. Lorenz, EPA, Las Vegas, Nevada
D. L. Stevenson, EPA, Las Vegas, Nevada
Publicity Chairperson:
G. Douglas, EPA, Las Vegas, Nevada
Members: J. V. Behar, EPA, Las Vegas, Nevada
A. E. Peckham, EPA, Las Vegas, Nevada
R. A. Hussey, EPA, Washington, DC
C. J. Rizzardi, EPA, Las Vegas, Nevada
P. J. Wunder, EPA, Las Vegas, Nevada
Executive Secretary :
R. Kinnison, EPA, Las Vegas, Nevada
iv
-------
THIRD JOINT CONFERENCE ON SENSING
OF ENVIRONMENTAL POLLUTANTS
Steering Committee
Chairperson: D. S. Barth, EPA, Las Vegas, Nevada
Organizing Vice - Chairperson:
B. H. Manheimer, HUD, Washington, DC/IEEE
Members: S. T. Quigley, ACS, Washington, DC
K. M. Foreman, Grumman Aerospace Corp.,
Bethpage, New York/AIAA
K. C. Spengler, AMS, Boston, Massachusetts
S. C. Coroniti, DOT, Washington, DC
P. Vestal, ISA, Pittsburgh, Pennsylvania
R. A. Schiffer, NASA, Washington, DC
C. E". Jensen, NOAA, Rockville, Maryland
V. E. Derr, NOAA, Boulder, Colorado
T. R. Ellis, Dynatrend, Inc., Burlington,
Massachusetts
G.Turner, Beckman Instruments, Inc.
Fullerton, California /ISA
P. D. Edmonds, IEEE, New York, New York
Program Committee
Chairperson: J. L. Moyers, University of Arizona, Tucson
Arizona /ACS
Members: T. R. Ellis, Dynatrend, Inc., Burlington,
Massachusetts
G. B. Morgan, EPA, Las Vegas, Nevada
M. E. Ringenbach, NOAA, Rockville,
Maryland /IEEE
V. E. Derr, NOAA, Boulder, Colorado
R. A. Schiffer, NASA, Washington, DC
H. W. Baynton, NCAR, Boulder, Colorado/AMS
G. Turner, Beckman Instruments, Inc.,
Fullerton, California/ISA
Managing Society. IEEE
IEEE ORGANIZING COMMITTEE
Organizing Vice - Chairperson:
B. H. Manheimer, HUD, Washington, DC
Las Vegas Arrangements Chairperson:
E. D. Emett, Nevada Power, Las Vegas,
Nevada
Program Committee Representative:
M. E. Ringenbach, NOAA, Rockville,
Maryland
Exhibits Chairperson:
H. C. Thacke, Central Telephone Co., Las
Vegas, Nevada
Members: E. L. R. Corliss, NBS, Washington, DC
R. G. Gentile, Babcock & Wilcox, Alliance,
Ohio
Publicity Chairperson:
H. J. Perlis, New Jersey Institute of Technology
Newark, New Jersey
Treasurer / Finance Chairperson:
O. Marsh, Nevada Power Co. (Ret.), Las Vegas,
Nevada
Members: J. C. Alter, Ketron, Inc., Arlington,
Virginia
Publications Chairperson:
L. L. Nagel, EPA, New York, New York
Members: F. Louden, Nevada Power, Las Vegas,
Nevada
Staff:
P. D. Edmonds, IEEE, New York, New York
The above named are members of the IEEE Environmental
Quality Committee or the IEEE Las Vegas Section.
-------
CONTENTS
Pagination in contents and body of proceedings is organized
according to sequential order of" Sessions and paper number.
Volume I begins with Plenary Session I followed by the papers
in Technical Sessions I to 19. Volume 2 continues with the
papers in Technical Sessions 20 to 38 followed by Plenary
Sessions II and III.
**
Manuscript not received in time for publication.
PREFACE iii
ACKNOWLEDGMENTS iv
SESSION INDEX vi
ABBREVIATED SESSION TITLE INDEX xxi
AUTHOR INDEX xxii
LIST OF EXHIBITORS xxvi
PLENARY SESSION I INTRODUCTION AND KEYNOTE
INTRODUCTION BY D. S. BARTH Pl-I
REMARKS BY THE HONORABLE RUSSELL E. TRAIN PI-2
REMARKS BY DR. B. N. DIE.TERICH PI-5
OPENING STATEMENT BY DR. A. S. PAVLOV PI-6
ADDRESS BY DR. EDUARDO ECHEVERRIA ALVAREZ PI-8
INTERDEPENDENCE EQUALS ENVIRONMENTAL
IMPROVEMENT BY FITZHUGH GREEN Pl-II
LUNCHEON ADDRESS BY CHRISTIAN A. HERTER, JR PI-13
SESSION 1: TOXIC MATERIALS WITH SPECIAL REFERENCE TO HEAVY METALS - 1
Chairperson: Lars Friberg
The Karolinska Institute
Stockholm, Sweden
INTRODUCTORY REMARKS, Lars Friberg 1-0
THE FLOW OF METALS IN THE ENVIRONMENT, Dale Jen Kins I-I
SAMPLING AND ANALYSIS OF METALS IN AIR, WATER AND WASTE PRODUCTS, James Morgan 1-2
METAL CONCENTRATIONS OF MOSSES AND OTHER PLANT MATERIALS AS
INDICATORS OF ATMOSPHERIC METAL DEPOSITION, Germund Tyler 1-3
METAL CONCENTRATIONS IN BLOOD, URINE, HAIR AND OTHER TISSUES AS
INDICATORS OF METAL ACCUMULATION IN THE BODY, Thomas Clarkson 1-5
THE ROLE OF EPIDEMIOLOGY IN ASSESSING HEALTH EFFECTS OF METALS,
Kenneth Bridboard, Joseph K. Wagoner, Hector P. Blejer,
Philip J. Landrigan and Richard A. Lemen 1-6
METHYLMERCURY: FORMATION IN PLANT TISSUES, Don D. Gay 1-7
SESSION 2: HALOGENATED ORGANICS
Chairperson: James McNesby
Office of Air and Water Measurement
National Bureau of Standards
Washington, DC
MONITORING - THE TRIGGER FOR ACTION, I. E. Wallen 2-1
MONITORING VINYL CHLORIDE IN THE VICINITY OF POLYVINYL
CHLORIDE FABRICATION PLANTS, Lawrence A. Elfers and Harold G. Richter 2-2
ENVIRONMENTAL LEVELS OF PCB'S, Doris J. Ruopp and Vincent J. DeCarlo 2-3
HALOGENATED HYDROCARBONS AND THE ENVIRONMENT, John A. Zapp, Jr 2-4
vi
-------
PREPARATION AND EVALUATION OF VINYL CHLORIDE STANDARDS,
E. E. Hughes, W. D. Dorko, S. M. Freund and D. M. Sweger 2-5
IMPROVED METHODS OF SAMPLING AND ANALYSIS OF VINYL CHLORIDE
AND OTHER GASEOUS CARCINOGENS, R. C. Lao, R. S. Thomas and J. L. Monkman 2-6
^^COMPREHENSIVE AREA MONITORING SYSTEM FOR VINYL CHLORIDE,
K. R. Lindelin, R. E. Gillespie and D. Mattson 2-7
SESSION 3: DESIGN FOR ENVIRONMENTAL MONITORING SYSTEMS - 1
Chairperson: R. Hal Langford
Office of Water Data Coordination
U.S. Geological Survey
Reston, Virginia
BASIC NEEDS FOR MONITORING AIR POLLUTANTS IN A DEVELOPING
COUNTRY, i. M. Dave
A PRELIMINARY DESIGN FOR A NATIONAL ENVIRONMENTAL CENSUS, Richard H. Rosen 3-2
DESIGN OF NATIONWIDE WATER-QUALITY MONITORING NETWORKS,
R. J. Pickering and J. T. Ficke 3-3
SOME INTERNATIONAL ACTIVITIES IN ENVIRONMENTAL MONITORING, G. Ozolins 3-4
3-5
EARTHWATCH - SENTINEL FOR THE FUTURE, Clayton E. Jensen
and Da I I* W. Brown
^NETWORK DESIGN CONSIDERATIONS FOR THE GLOBAL ENVIRONMENTAL MONITORING
SYSTEMS (GEMS) OF THE UNITED NATIONS, Robert Citron . . . . . . ... 3_6
SESSION 4: THE EVALUATION AND ASSESSMENT OF PROBLEMS ASSOCTATFn
WITH WASTE DISPOSAL PROCESSES fwunftitu
Chairperson: G. Cox
Raytheon Company
Portsmouth, Rhode Island
POLYCHLORINATED BI PHENYLS OFF SOUTHERN CALIFORNIA, DavId R Youna
Deirdre J. McDermott and Theadore C. Heesen . . . . ' 9'
WATER CONTAMINATION DETECTION, Ihor Lysyj 4 2
TRACE METALS IN THE COMBUSTIBLE FRACTION OF URBAN REFUSE
Harold E. Marr, Stephen L. Law and David L. Neylan. ...... 4 3
SPARK SOURCE MASS SPECTROMETRY! THE OPTIMUM REFEREE MFTHOD FOR
MULTIELEMENT ANALYSIS, Charles E. Taylor 4
FREQUENCY DISTRIBUTIONS FOR COL I FORM BACTERIA IN WATER
Wesley 0. Pipes, Pamela Ward and S. H. Ahn. . 4 ^
TRACE METAL ANALYSIS BY ATOMIC EMISSION USING A DC ARGON PLASMA
Walter G. Cox '
W EVALUATION OF AUTOMATIC WASTEWATER COMPOSITORS, Daniel J.'Harris
and Willi am J. Kef fer
4-7
SESSION 5: MONITORING AND EVALUATION OF ATMOSPHERIC PARTICULATE MATTER
Chairperson: James Lodge
Boulder, Colorado
MICROSCOPICAL IDENTIFICATION OF ATMOSPHERIC PARTICLES, Walter C. McCrone 5-1
MONITORING OF ATMOSPHERIC AEROSOL MASS AND SULFUR CONCENTRATION
E. S. Macias, R. B. Husar and J. D. Husar ' 5 2
MEASUREMENTS WITH A PROTOTYPE MASS DISTRIBUTION MONITOR FOR
PARTICULATE AIR POLLUTION, W. Stober and F. J. MSnig 5_3
THE APPLICATION OF PATTERN RECOGNITION TECHNIQUES TO THE
CHARACTERIZATION OF ATMOSPHERIC AEROSOLS, S. P. Perone, M. Pilcher
P. jSaarenstroom and J. L. Moyers ' ' ^
vii
-------
THE USE OF A DUAL BEAM LASER TRANSMISSOMETER AS A MEANS OF
MONITORING AIR QUALITY, W. C. Malm, K. O'Del I and J. S. Hall 5-5
DESIGN CONSIDERATIONS AND FIELD PERFORMANCE FOR AN IN SITU,
CONTINUOUS FINE PARTICULATE MONITOR BASED ON RATIO-TYPE LASER
LIGHT SCATTERING, Gerhard Kreikebaum and Frederick M. Shofner 5-6
NEW ANALYTICAL PROSPECTS FOR AN OLD DETECTOR, L. L. Altpe+er, Jr 5-7
SESSION 6: TOXIC MATERIALS WITH SPECIAL REFERENCE TO HEAVY METALS - 2
Chairperson: Lars Friberg
The Karolinska Institute
Stockholm, Sweden
MODIFYING FACTORS IN METAL TOXICOLOGY, David H. Groth 6-1
THE DEPOSITION OF PB-CONTAINING PARTICULATE FROM THE
LOS ANGELES ATMOSPHERE, C. I. Davidson, S. V. Hering and
S. K. Friedlander 6-3
CADMIUM UPTAKE BY CEREALS AND VEGETABLES, A. Dewit,
R. De Jaegere and D. L. Massart 6-4
SOME CONSIDERATIONS ON MONITORING OF TRACE METALS IN ESTUARIES
AND OCEANS, Douglas A. Segar and Adriana Y. Can+illo 6-5
HEAVY METAL CONCENTRATIONS IN MARINE ORGANISMS NEAR AN
INDUSTRIAL WASTE OUTFALL, Raymond Emerson, Dorothy F. Soule,
Mikihiko Oguri, Kenneth Chen and James Lu 6-7
SESSION 7: PESTICIDES - 1
Chairperson: Henry Enos
Equipment and Techniques Division
Office of Monitoring Systems
U.S. Environmental Protection Agency
Washington, OC
ANALYSIS OF FOOD FOR RESIDUES OF PESTICIDES,
Jerry Burke and Bernadette McMahon 7-1
THE DETERMINATION OF PESTICIDE RESIDUES IN AIR, James N. Seiber
and James E. Woodrow 7-2
ANALYSIS OF PESTICIDE RESIDUES IN FIELD SOILS: OPTIMIZING
SOIL SAMPLING AND PESTICIDE EXTRACTION, Joseph H. Caro
and Alan W. Taylor 7-5
CONFIRMATION OF PESTICIDE RESIDUE IDENTITY BY CHEMICAL
DERI VATIZATION, W. P. Cochrane 7-4
APPLICATION OF ANALYTICAL METHODS RESEARCH TO MONITORING
ORGANIC RESIDUES IN FISH, David L. Stalling 7-5
DETERMINATION OF Ml REX AND PCBS IN HERRING GULL FROM
LAKE ONTARIO BY PERCHLORI NAT IONS AND GC MS, Douglas J. Hallett 7-6
PESTICIDES IN DEVELOPING COUNTRIES,. M. T. Farvar 7-7
SESSION 8: DESIGN FOR ENVIRONMENTAL MONITORING SYSTEMS - 2
Chairperson: R. Hal Langford
Office of Water Data Coordination
U.S. Geological Survey
Reston, Virginia
ENVIRONMENTAL MONITORING IN LATIN AMERICA AND THE CARIBBEAN,
Ricardo Haddad and Walter Castagnino 8-1
LARGE SCALE AIR POLLUTION MONITORING IN THE NETHERLANDS,
T. Schneider ............... 8-3 -
viii
-------
A DESCRIPTION OF A TECHNIQUE FOR DEVELOPING AN OPTIMUM AIR
POLLUTION AND METEOROLOGICAL SAMPLING NETWORK IN URBAN REGIONS,
Fred M. Vukovich ' „ .
DESIGN AND REDESIGN OF AN AIR SAMPLING NETWORK, James H. Price,
Gary K. Tannahill, Duane J. Johnson, Andy C. Wheatley
and Roger R. Wall is g
THE REGIONAL AIR MONITORING SYSTEM ST. LOUIS, MISSOURI, U.S.A.
R. Lee Myers and James A. Reagan g 5
ENVIRONMENTAL QUALITY SURVEILLANCE IN INDONESIA,
Sri Suwasti Susanto „
...... 0-7
SESSION 9: THE EVALUATION AND ASSESSMENT OF GROUNDWATER QUALITY - 1
Chairperson: R. Tin!in
General Electric - TEMPO
Santa Barbara, California
LEGAL REQUIREMENTS FOR MONITORING GROUNDWATER QUALITY
Harvey 0. Banks and Leslie G. McMiliion ' g I
ECONOMIC FRAMEWORK FOR GROUNDWATER QUALITY MONITORING
Robert L. Crouch, Ross D. Eckert and Donald D. Rugg. ! g_2
DEVELOPMENT OF A METHODOLOGY FOR MONITORING GROUNDWATER
QUALITY, David K. Todd and Lome G. Everett 9 3
-MONITORING GROUNDWATER POLLUTION, Kenneth D. Schmidt ' '9.4
MANAGEMENT OF GROUNDWATER QUALITY DATA, Norman F. Hampton 9-5
GROUNDWATER POLLUTION MONITORING CASE STUDIES, L. Graham Wilson 9-6
MONITORING OF SUBSURFACE WASTE-DISPOSAL OPERATIONS: A GENERALIZED SYSTEMS
APPROACH WITH EXAMPLES FROM CANADIAN CASE HISTORIES, Frank Simpson 9.7
SESSION 10: THE APPLICATION OF REMOTE SENSING TECHNIQUES FOR MONITORING
AND ASSESSING ENVIRONMENTAL POLLUTION - 1 ™JNiiUKlNb
Chairperson: S. H. Melfi
U.S. Environmental Protection Agency
Las Vegas, Nevada
APPLICATION OF REMOTE MONITORING TECHNIQUES IN AIR ENFORCEMENT
C. B. Ludwig and M. Griggs '
10-1
LASER DOPPLER SYSTEMS IN POLLUTION MONITORING, Christopher R. Miller
Charles M. Sonnenschein, Dr. William F. Herget and R. Milton Huffaker 10-2
DOWNLOOKfNG AIRBORNE LfQAR STUDIES - AUGUST 1974 j A Eckert
J. L. McElroy, D. H. Bundy, J. L. Guagliardo and's. H.'Melfi .' |0-3
VISUALIZATION OF EDDIES IN THE PLANETARY BOUNDARY LAYER BY
MEANS OF LIDAR, K. E. Kunkel, E. W. Eloranta and J. A. Weinman
REMOTE SENSING OF ATMOSPHERIC POLLUTANT GASES USING AN INFRARED
HETERODYNE SPECTROMETER, R. K. Seals, Jr. and B. J. Peyton . l0-5
REMOTE SENSING OF ATMOSPHERIC SO USING THE DIFFERENTIAL ABSORPTION
LIDAR TECHNIQUE, J. M. Hoell, Jr?, W. R. Wade a„d R. T. Thomson, Jr . . ,0-6
TOTAL ATMOSPHERIC COLUMN AND TROPOSPHERIC ABUNDANCE MEASUREMENT
OF NITROGEN DIOXIDE BY ABSORPTION SPECTROSCOPY, Watson R. Henderson. . . ,0-7
SESSION 11: PANEL DISCUSSION. FACILITATING THE MERGER OF
TECHNOLOGY AND GOVERNANCE
Chairperson: B. H. Manheimer
Department of Housing and Urban Development
Washington, DC
Panel Members: J. Primack, S. T. Quigley, A. Daush and M. M. McNamara H_l
ix
-------
SESSION 12: INORGANICS - 1
Chairperson: Elmer Robinson
College of Engineering
Washington State University
Pullman, Washington
EPIDEMIOLOGICAL STUDIES RELATIVE TO EFFECTS OF SO AND
SULFURIC ACID, V. A. Bus+ueva ? 12-1
THE INDOOR ENVIRONMENT — A NEW FRONTIER?, John E. Yocum 12-2
AN EXAMPLE OF THE USE OF MONITORING AND MODELING FOR EVALUATING
AIR POLLUTION CONTROL PLANS, F. L. Ludwig and R. T. H. Co I I i s 12-3
DETAILED CHEMICAL ANALYSIS OF AIRBORNE PARTICLES, D. A. Levaggi,
J. S. Sandberg, R. E. De Mandel and M. Feldstein 12-4
THE CONTINUOUS MONITORING OF SULFUR DIOXIDE AND SUSPENDED
PARTICULATES NEAR NON-FERROUS SMELTERS, K. W. Nelson,/M. 0. Varner,
R. D. Putnam, M. A. Yeager and R. B. Watson 12-5
ON THE DEFINITION OF REQUIREMENTS FOR MONITORING OF THE PHYSICAL
AND CHEMICAL PROPERTIES OF TROPOSPHERIC SULFATE AEROSOLS RELEVANT
TO HEALTH EFFECTS, R. J. Charlson, A. P. Waggoner, D. S. Covert
and N. C. Ahlquist 12-6
THE NEED FOR INTERNATIONAL MODELS OF ENVIRONMENTAL STANDARDS,
Raisaku Kiyoura 12-7
SESSION 13: PESTICIDES - 2
Chairperson: Henry Enos
Equipment and Techniques Division
Office of Monitoring Systems
U.S. Environmental Protection Agency
Washington, DC
MACRO AND MICRO APPROACHES TO THE DETERMINATION OF PESTICIDE
RESIDUES IN HUMAN AND ANIMAL TISSUES, Robert F. Moseman 13-1
CHOLINESTERASE ACTIVITY AS A BIOCHEMICAL INDICATOR FOR MONITORING
EXPOSURE TO CERTAIN PESTICIDES, F. Kaloyanova 13-2
AQUATIC SNAKES AS COMPOSITE SAMPLES FOR ORGANOCHLORINE PESTICIDE
RESIDUES, H. Erie Janssen, Reld Dennis and James R. DeShaw 13-3
A NATIONAL PESTICIDE MONITORING PROGRAM OVERVIEW, Thomas C. Carver, Jr 13-4
F$NEL DISCUSSION. MONITORING PANEL OF THE FEDERAL WORKING GROUP ON
PESTICIDE MANAGEMENT
Chairperson: Carl Bulger
Army Material Command
Washington, DC
Panel Members: Clifford Roan, Dan Donahue, Don White, Herman Feltz,
Harold Trabosh, Tom Carver, Paul Corneliussen, John Wessel,
Ann Carey, Bob Heath, Frederick Kutz and Philip Butler 13-5
SESSION 14: ENVIRONMENTAL MODELING - 1
Chairperson: Alan Eschenroeder
Environmental Research and Technology, Inc.
Santa Barbara, California
THE DEVELOPMENT OF A NUMERICAL MODEL TO PREDICT POLLUTANT
CONCENTRATIONS DURING FUMIGATION CONDITIONS WITH AN
ONSHORE FLOW, H. S. Rosenblum, T. F. Lavery and B. A. Egan 14-1
MULTIPLE-TRACER HIGHWAY DISPERSION STUDY, Walter F. Dabberdt 14-2
** SETT IMG OF A POTENTIAL SOURCE OF CONTAMINATION REGARDING
POPULATION AREAS: A METHOD FOR CALCULATION THE AIR POLLUTION
CONCENTRATION, Srdjan Mitrovic . (4-3
X
-------
A MODEL STUDY OF THE IMPACT OF EMISSION CONTROL STRATEGIES ON
LOS ANGELES AIR QUALITY, S. Hameed, S. A. Lebedeff and R. W. Stewart
POLLUTIONAL EFFECTS OF SUSPENDED, SEDIMENTED AND ERODED PARTICULATE
MATERIAL IN THE AQVEOUS ENVIRONMENT, Hermann H. Hahn and
Rudolf Klu+e . . c
• ••••.... 14-3
AUTOMATED FORECAST PROCEDURES FOR RIVER QUALITY MANAGEMENT,
D. A. Dunsmore and R. J. Boes ' , „ ,
» 14-0
A STRATEGY FOR AQUATIC POLLUTANTS FATE AND TRANSPORT DETERMINATIONS
Walter M. Sanders III '
• • •
SESSION 15: THE EVALUATION AND ASSESSMENT OF GROUNDWATER QUALITY - 2
Chairperson: D. A. Sangrey
Cornell University
Ithaca, New York
REMOTE SENSING OF GROUND AND SURFACE WATER CONTAMINATION BY
LEACHATE FROM LANDFILL, Dwight A. Sangrey, William L. Teng, Warren
R. Phi1ipson and To Liang
LAND DISPOSAL OF WASTES; POTENTIAL FOR GROUNDWATER POLLUTION
Richard A. Carnes, D. R. Bunner, R. E. Landreth and M. H. Roulier 15-2
**DETERMINATION OF TRACE ELEMENTS IN WATER BY THE INDUCTIVELY COUPLED PLASMA
MULTIELEMENT OPTICAL EMISSION SPECTROSCOPIC (ICP-MOES) TECHNIQUE: '
APPLICATION TO SOFT, HARD AND SALINE WATERS, Royce K. Winge, Velmer A. Fassel
and Richard N. Kniseley
ii)"J
A PROPOSED NATIONAL MONITORING SYSTEM FOR ORGAN ICS IN WATER
"J . M. McGuire ' , , „
15-4
CAPABILITIES AND LIMITATIONS IN IDENTIFYING AND MEASURING
AQUATIC POLLUTANTS, William T. Donaldson , 5_5
INVESTIGATION OF EXIT AREAS OF GROUNDWATER RELATED TO ANTHRACITE
DEEP MINES, Carolyn A. Petrus b
**AN APPROACH FOR ASSESSING THE ENVIRONMENTAL QUALITY OF TEXAS
ESTUARIES, Don Rauschuber and Mike Ellis , 15 7
rrrcTQM 16- the application of remote sensing techniques for monitoring
' AND ASSESSING ENVIRONMENTAL POLLUTION - 2
Chairperson: 0. Hinkley
Massachusetts Institute of Technology
Lincoln Laboratories
Cambridge, Massachusetts
ASSESSMENT OF THE BENEFITS OF ENVIRONMENTAL REMOTE SENSING
Alex Hershaft
MONITORING ESTUARINE CIRCULATION AND OCEAN WASTE DISPERSION USING AN
INTEGRATED DROGUE-AIRCRAFT-SATELLITE APPROACH, V. Klemas G. Davis
H. Wang, W. Wbelan and G. Tornatore '
EVALUATION OF WATER SAMPLES COLLECTED DURING LANDSAT-I OVERPASSES
OF THE LOWER CHESAPEAKE BAY AREA, D. E. Bowker and W. G Witte
LONG PATH INFRARED AMBIENT AND LABORATORY MEASUREMENTS. CHEMICAL
REACTIONS AND FREON STUDIES Bruce W. Gay, Jr.. Richlrd C Noonan,
Joseph J. Bufalim and Philip L. Hanst ~ l6 5
REMOTE SENSING OF TRACS CONSTITUENTS FROM ATMOSPHERIC INFRARED
EMISSION AND ABSORPTION SPECTRA, D. B. Barker, D. B. Murcray
J. N. Brooks, A. Goldman, J. J. Kos+ers, F. H. Murcrav
W. J. Williams and J. Van Allen. ...
..J
IN SITU AIRCRAFT MEASUREMENTS OF THE OH Frff p&nir&i . i pecp
INDUCED FLUORESCENCE, D. D. Davis, T. McGee, W. Heaps, A. Mori arty
and R. Schiff 7
xi
-------
SESSION 17: THE EVALUATION AND ASSESSMENT OF PROBLEMS ASSOCIATED
WITH ENERGY EXTRACTION AND UTILIZATION PROCESSES
Chairperson: T. Thoem
U.S. Environmental Protection Agency
Denver, Colorado
A NET ENERGY ANALYSIS OF THE USE OF NORTHERN GREAT PLAINS SURFACE
MINED COAL IN LOAD CENTER POWER PLANTS, Thomas Ballen+ine 17-1
TRACE CONTAMINANTS FROM COAL-FIREO POWER PLANTS,
R. C. Ragaini and J. M. Ondov 17-2
EVALUATION OF EFFECTS OF MULTIPLE POWER PLANTS ON A
RIVER ECOSYSTEM, G. C. Slawson, Jr. and B. C. Marcy 17-3
MONITORING OF ENVIRONMENTAL EFFECTS OF STRIP MINING FROM
SATELLITE IMAGERY, Ronald L. Brooks and Carlos G. Parra 17-4
AERIAL REMOTE SENSING APPLICATIONS IN SUPPORT OF OIL SPILL CLEANUP
CONTROL AND PREVENTION, Donald Jones, Robert Landers and Albert Pressman 17-5
POWER VS. POLLUTION: A NUMERICAL APPROACH, Harold I. Zeliger
and Marsha Funk 17-6
SESSION 18: INORGANICS - 2
Chairperson: Elmer Robinson
College of Engineering
Washington State University
Pullman, Washington
AIR POLLUTION IN THE VENETIAN AREA, Susana Cerquiglini and
E. Bianco 18-1
EVALUATION OF AIR QUALITY IN THE VICINITY OF THE INTERSECTION
OF WISCONSIN AND WESTERN AVE., N.W., Donna M. O'Toole and
Ronald C. Hilfiker 18-2
OZONE IN RURAL AREAS, H. Westberg, K. J. Altwine, R. A. Rasmussen
and E. Robinson 18-3
THE USE OF FISH TO CONTINUOUSLY MONITOR AN INDUSTRIAL EFFLUENT,
Garson F. Westlake, W. H. van der Schalie, John Cairns, Jr.
and Kenneth L. Dickson 18-4
SOURCES AND SAMPLING OF POLLUTANTS FROM GEOTHERMAL STEAM AREAS,
L. A. Cavanagti and R, E. Ruff 18-5
ESTIMATION OF POINT SOURCE EMISSION STRENGTHS WITH AIRCRAFT,
J. E. Cunningham, J. VI. Key, and C. D. Vtolbach 18-6
SESSION 19: THE MEASUREMENT AND ASSESSMENT OF PROBLEMS ASSOCIATED
WITH NUCLEAR FUEL PROCESSING
Chairperson: E. W. Bretthauer
U.S. Environmental Protection Agency
Las Vegas, Nevada
DEVELOPMENT OF A REFERENCE METHOD FOR THE MEASUREMENT OF PLUTONIUM
IN SOIL, E. W. Bretthauer and P. 8. Hahn 19-1
PROBLEMS IN ENVIRONMENTAL MONITORING AROUND NUCLEAR FACILITIES,
B. Robinson, W. H. Westendorf and C. A. Phillips 19-2
THE NUCLEAR REGULATORY COMMISSION CONFIRMATORY MEASUREMENT PROGRAM
FOR ENVIRONMENTAL AND EFFLUENT MEASUREMENTS, Bernard H. Weiss 19-3
INSTRUMENTATION FOR OFF-SITE REACTOR PLUME STUDIES, R. C. Ragaini,
D. E. Jones and G. W. Huckabay 19-4
DEVELOPMENT OF THE NATIONAL BUREAU OF STANDARDS ENVIRONMENTAL
RADIOACTIVITY STANDARD: RIVER SEDIMENT, James R. Noyce,
John M. R. Hutchinson, Wilfrid B. Mann and Patricia A. Mullen 19-5
xii
-------
PRACTICAL PROBLEMS OF MONITORING FOR PLUTONIUM IN THE
ENVIRONMENT, H. C. Woodsum 19-6
SESSION 20: ENVIRONMENTAL MODELING - 2
Chairperson: Alan Eschenroeder
Environmental Research and Technology, Inc.
Santa Barbara, California
DEVELOPMENT OF CRITERIA FOR ESTABLISHING GUIDELINES FOR OPTIMIZATION
OF ENVIRONMENTAL MONITORING NETWORKS: AIR MONITORING NETWORKS,
Joseph V. Behar, Leslie M. Dunn, James L. McElroy, Robert R. Kinnison
and Pong N. Lem 20-1
A COMPUTER PROGRAM FOR GENERATING THE DIURNAL VARIATION OF PHOTOLYTIC
RATE CONSTANTS FOR ATMOSPHERIC POLLUTANTS, Kenneth L. Demerjian and
K. L. Schere 20-2
MODELING SOLUTE TRANSPORT IN GROUNDWATER, Leonard F. Konikow 20-3
THE APPLICATION OF STEINBUCH'S "LERNMATRIX" AS A NEW MATHEMATICAL
APPROACH IN THE ASSESSMENT OF AIR POLLUTION EFFECTS, B. Prinz
and J. Hower 20-4
DEVELOPMENT OF AN IMPROVED MODEL FOR STATISTICAL ANALYSIS OF ENVIRONMENTAL
DATA; SIMULATION AND VERIFICATION, David T. Mage and Wayne R. Ott 20-5
DESIGN OF POLLUTANT ORIENTED INTEGRATED MONITOR IMG SYSTEMS,
E.. A. Schuck and G. B. Morgan 20-6
SESSION.21: MARINE POLLUTION - MEASUREMENT AND PROBLEM ASSESSMENT
Chairperson: C. S. Giam
Texas A 8 M University
College Station, Texas
PROBLEMS IN ANALYSES OF ORGANIC POLLUTANTS IN OPEN-OCEAN SAMPLES,
C. S. Giam, H. S. Chan and G. S. Neff 21-1
**ADVANCED OCEAN MEASUREMENT TECHNOLOGY: ITS ROLE IN THE INTERNATIONAL DECADE
OF OCEAN EXPLORATION (I.D.O.E.), Deane E. Holt 21-2
**PR0BLEMS OF DETERMINING LEAD CONCENTRATIONS IN MATERIALS TAKEN FROM
THE MARINE ENVIRONMENT, C. C. Patterson and D. M. Settle 21-3
MARINE AIR CHEMISTRY MEASUREMENT CRITERIA, R. A. Rasmussen and
J . Al Iwine 21-4
**PETROLEUM IN THE SEA: RECOGNITION AND BIOLOGICAL EFFECTS, P. L. Parker 21-5
THE MEANING AND MEASUREMENT OF TURBIDITY, B. S. Pijanowski 21-6
SESSION 22: EVALUATION AND ASSESSMENT OF PROBLEMS ASSOCIATED WITH
TRANSPORTATION PROCESSES
Chairperson: F. F. Marmo
Transportation Systems Center
Dept. of Transportation
Cambridge, Massachusetts
CRITERIA FOR THE OBJECTIVE EVALUATION OF NO/NO /Q AIR
MONITORING DATA, Donald Stedman 2. f 22-2
EFFICIENT COLLECTION AND ANALYSIS OF HAZARDOUS ORGANIC COMPOUNDS
FROM COMBUSTION EFFLUENTS, P. E. Strup, P. W. Jones, R. D. Giammar
and T. 0. Stanford 22-3
AIR QUALITY MONITORING AND ANALYSIS PROGRAM FOR THE J. F. KENNEDY
MEMORIAL LIBRARY ENVIRONMENTAL IMPACT STATEMENT, R. D. Siege I,
P. H. Guldberg and R. P. Hebert 22-4
MEASUREMENTS OF AIR POLLUTANTS OVER A LOS ANGELES FREEWAY WITH
A BI STATiC LASER SYSTEM, Robert T. Menzies and Michael S. Shumate 22-5
xiii
-------
EVALUATION OF CHEMILUMINLSCENT OXIDES OF NITROGEN
MONITORS, W. J. F i ndlay, G. Dowd and N. Q.uickert
22-6
SESSION 23: ATMOSPHERIC SULFUR - MEASUREMENT AND EVALUATION
Chairpersons: fi. Hidy and P. Mueller
Environmental Research & Technology
Westlake Village, California
OX I DAT I ON OF SULFUR DIOXIDE IN THE ATMOSPHERE: A REVIEW,
Ha I stead Harrison, T. V. Larson and Pe+er V. Hobbs 23-1
SULFUR SPECIES AND HEAVY METALS IN PARTICULATES FROM
A COPPER SMELTER, Lee D. Hansen, Delbert J. Eatough,
Nolan F. Mangelson, Trescott E. Jensen and
Douglas Cannon ......... 23-2
A COMPARATIVE STUDY OF METHODS FOR SULFATE ANALYSIS IN
ATMOSPHERIC PARTICLES, B. R. Appel, £. L. Ko+bny, E. M. Hotter,
J. J. Wesolowski and R. D. Giauque 23-3
THE CONTINUOUS MONITORING OF PARTICULATE SULFATE BY FLAME
PHOTOMETRY, James J. Hun+zicker, Lome Isabella and J. C. Watson 23-4
ABSOLUTE RAMAN SCATTERING CROSS SECTIONS OF SULFATE AND 01 SULFATE
AMD THEIR APPLICATION TO AQUEOUS AEROSOL MONITORING, Richard G. Stafford
and Richard K. Chang . . 23-5
THE DISTRIBUTION AND CHARACTER I ST ICS OF PARTICULATE SULFATES IN
THE SOUTHWEST DESERT ATMOSPHERE, R. G. Keesee, S. B. Hopf and
J, L. Moyers 23-6
**THE ATMOSPHERIC CHEMISTRY OF REGIONAL - SCALE SULFATE AND OZONE,
James P. Friend and Douglas D. Oavis - 23-7
SESSION 24: ORSANICS - 1
Chairperson; James N. Pitts, Jr.
Statewide Air Pollution Research Center
University of California
Riverside, California
MEASUREMENTS, MECHANISMS AND MODELS: CRITERIA FOR THEIR COST EFFECTIVE
APPLICATION TO AIR POLLUTION CONTROL, James N. Pitts, Jr 24-1
THE APPLICATION OF CHEMI LUMINESCENCE TO THE MEASUREMENT OF REACTIVE
HYDROCARBONS IN AMBIENT AIR, J. C. Hilborn, W. J. Find I ay and
N. Quickert 24-2
SOURCE RECONCILIATION OF ATMOSPHERfC HYDROCARBONS, 1974,
Henry Mayrsohn, James H. Crabiree and M. Kuramoto. 24-3
FOURIER-TRANSFORM SPECTROSCOPIC STUDIES OF ORGANIC COMPOUNDS
PARTICIPATING IN PHOTOCHEMICAL SMOG FORMATION, H. Miki, P. Maker,
C. Savage and L. Brei+enbach 24-4
ORGAN ICS IN ATMOSPHERIC AEROSOLS IN BELGIUM, D. Rondia,
F. Dewiest and H. Oeila Fiorentina 24-5
THE DESIGN OF A SOURCE AND AMBIENT AIR MONITORING SYSTEM FOR A
MAJOR URBAN INDUSTRIAL COMPLEX IN PORTUGAL, G. C. Martins and
L. Canaies 24-6
SESSION 25: DATA MANAGEMENT AND ASSESSMENT - 1
Chairperson: Dai 1 W. Brown
National Oceanic and Atmospheric Administration
Office of Associate Administrator for Marine Resources
Office of Nonliving Resources
RockviTle, Maryland
DATA MANAGEMENT: KEY TO ENVIRONMENTAL QUALITY ASSESSMENT,
Thomas S. Austin
xiv
-------
THE ENVIRONMENTAL INFORMATION SYSTEM (EIS) IN SWEDEN
AND MONITORING OF THE ENVIRONMENT, Ingar Palmlund and
Ingvar Andersson 25-2
EPA'S AEROMETRIC AND EMISSIONS REPORTING SYSTEM (AEROS):
AN APPROACH TO AIR QUALITY DATA MANAGEMENT, John C. Bosch, Jr 25-3
THE CONCEPT OF THE "MAXIMUM I MM ISSI ON VALUE" IN THE FGR AS
A PRELIMINARY STAGE FOR AMBIENT AIR QUALITY STANDARDS,
B. Prinz and H. S+ra+mann . 25-4
A PRACTICAL DATA MANAGEMENT SYSTEM FOR MONITORING AND
MANAGING LARGE DISTRIBUTED ECOSYSTEMS, P. David Fisher
and J . R. Gru i ft 25-5
THE ROLE OF ENVIRONMENTAL DATA BANKS IN ENERGY RESOURCE
DEVELOPMENT, J. E. Jones and G. E. Smith 25-6
SESSION 26: MONITORING AND ASSESSMENT OF GLOBAL POLLUTANT TRANSPORT
Chairperson: J. C. Gil 1e
National Center for Atmospheric Research
Boulder, Colorado
*
GLOBAL TRANSPORT OF POLLUTANTS AND TRACE GASES
J. C. Gi.lle '
I
THE USE OF GLOBAL CIRCULATION MODELS FOR CLIMATE SIMULATION
.AND TRANSPORTS OF POLLUTANTS, Warren M. Washington 26-2
BACKGROUND MONITORING (HAWAII, ALASKA, SAMOA, ANTARCTICA) -
The LAST TWO YEARS, O. H. Pack 26_3
AN AUTOMATED AIR SAMPLING SYSTEM OPERATING ON 74? AIRLINERS
Porter J. Perkins and UIf R. C. Gustafsson ' „
* * * • • • « « ~ ^.d-4
A REVIEW OF LOWER STRATOSPHERIC-UPPER TROPOSPHERIC TRACF CAS
CONSTITUENTS TO BE MONITORED BY THE GLOBAL AIR SAMPLING
PROGRAM (GASP), Phillip D. Falconer and John M. Miller 26-5
TEMPORAL MD SPATIAL CHANGES IN OZONE FROM THE NIMBUS IV
BUV EXPERIMENT, Ronald M. Nagatani, Alvin J. Miller and
Donald F. Heath
ANTARCTIC SULFATE AEROSOLS: STRATOSPHERIC OR TROPOSPHERIC
ORIGIN?, W. H. Zoller, E. J. Mroz and W. Maenhaut. ... 26 7
SESSION 27: TRANSPORTATION - 2: AIRCRAFT EMISSIONS
Chairperson: W. M. Roquemore
Wright-Patterson Air Force Base, Ohio
rnlmtlFDp!ir'i^FltG^T,ANS<:T HTDR0CARB0N EMISSIONS FROM AIRCRAFT GAS
TURBINE ENGINES, S. A. Stumpf and W. S. Blazowski 27-1
ORGANIC COMPOUNDS IN TURBINE COMBUSTOR EXHAUST, James P Conkie
William W. Lackey, Charles L. Martin and Richard L. Miller ! 27-2
ANALYSIS OF THE HYDROCARBON FRACTION OF JET ENGINE EXHAUST RY
SUBTRACTIVE GAS CHROMATOGRAPHY, Marilyn S. Black, William R Rehq
and Robert E. Sievers a
27-3
FLUORESCENCE OF HYDROCARBONS IN JET ENGINE EXHAUST
W. M. Roquemore and F. N. Hodgson ' •- 27 4
PARTICULATE SAMPLING FROM GAS TURBINE ENGINES,
A. F. Klarman and J. E. Horlinq. , .
a 27-5
EVALUATION OF PROBE SAMPLING VERSUS AN IN SITU OPTICAL TECHNIOUE FOR
NITRIC OXIDE CONCENTRATION MEASUREMENT IN COMBUST ION GASSTREAMS
J. 0. Few, R. J. Bryson, W. K. McGregor and M. G. Davis ' 27-6
SESSION 28: THE ASSESSMENT OF AMBIENT ENVIRONMENTAL AIR QUALITY
Chairperson: H. Kramer
pan American Health Organization
Mexico City, Mexico
xv
-------
ENVIRONMENTAL EDUCATION AND TRAINING IN A DEVELOPING
COUNTRY, H. Kramer 28-1
PLANNING CONTROL STRATEGIES TO REDUCE CARBON MONOXIDE
CONCENTRATIONS - MINNESOTA'S EXPERIENCE, Ingrid Ritchie
and John Seltz 28-2
USING AMBIENT MONITORING TO IDENTIFY AIR POLLUTION SOURCES
WITH APPLICATION TO OXIDANT TRANSPORT, Ralph C. Sklarew
and John C. Wilson 28-3
A NEW METHOD OF SAMPLING FOR SUSPENDED PARTICULATES,
Edward Hyne and Paul Norton 28-4
OZONE TRENDS IN THE EASTERN LOS ANGELES BASIN CORRECTED
FOR METEOROLOGICAL VARIATIONS, Melvin D. ZeIdin
and D. M. Thomas 28-5
AMBIENT AIR MONITORING USING LONG PATH TECHNIQUES,
Willi am A. McClenny 28-6
CHEMI LUMINESCENT DETERMINATION OF NO AND NO ,
George S. Turner and R. M. Neti 28-7
SESSION 29: FINE PARTICLES - 1
Chairperson: Arthur M. Langer
Mount Sinai School of Medicine
New York, New York
ATMOSPHERIC AEROSOLS - MEASUREMENT AND CHARACTERISTICS,
K. T. Whitby and B. Cantrell 29-1
**AIRBORNE PARTICULATE TRACE ELEMENTS - INTERRELATIONS OF TEMPORAL VARIATIONS,
HUMAN EXPOSURES, AND SAMPLING APPROACHES, Theo. J. Kneip,
Micheal Kleinman, David Bernstein and Merril Eisenbud 29-2
AIR POLLUTION MEASUREMENTS IN A EUROPEAN EPIDEMIOLOGICAL SURVEY
OF RESPIRATORY AILMENTS IN CHILDREN, A. Berlin, E. Di Ferrante,
Ph. Bourdeau and J. Smeets 29-3
^ENVIRONMENTAL ASSAY: PARTICLES IN HUMAN TISSUES,
A. M. Langer, M. Wolff and I. B. Rubin 29-4
YALE ART AND ARCHITECTURE BUILDING ASBESTOS CONTAMINATION:
PAST, PRESENT, AND FUTURE, Robert N. Sawyer 29-5
ASBESTOS CONTAMINATION OF BUILDING AIR SUPPLY SYSTEMS,
William J. Nicholson, Arthur N. RohI, Irving Weisman 29-6
SESSION 30: ORGANICS - 2
Chairperson: James N. Pitts, Jr.
Statewide Air Pollution Research Center
University of California
Riverside, California
ANALYSIS FOR ORGANIC POLLUTANTS IN DRINKING WATER,
I. H. M. Suffet and J. V. Radziul 30-1
FLUORESCENCE MEASUREMENTS OF CARCINOGENIC AND POLYCYCLIC
AROMATIC HYDROCARBONS IN WATER, Frederick P. Schwarz
and Stanley P. Wasik 30-2
MONITORING FOR ORGANIC AIR POLLUTANTS IN GREAT BRITAIN,
Anthony Verdin 30-3
THE ENVIRONMENTAL CYCLE AND BALANCE OF POLYCYCLIC
AROMATIC HYDROCARBONS, Michael J. Suess 30-4
YIELD AND HEAVY METAL CONTENT OF FOOD CROPS GROWN ON SOILS AMENDED
WITH SEWAGE SLUDGE, F. T. Binqham and G. A. Mitchell 30-6
N-NITROSOAMINES IN THE ENVIRONMENT, D. H. Fine,
D. P. Rounbehler, Nancy M. Belcher and S. Epstein 30-7
xvi
-------
SESSION 31: DATA MANAGEMENT AND ASSESSMENT - 2
Chairperson: Dail VI. Brown
National Oceanic and Atmospheric Administration
Office of Associate Administrator for Marine Resources
Office of Nonliving Resources
Rockville, Maryland
ENVIRONMENTAL INDICES AND MONITORING, H. Inhaber 3,.,
SOME PRACTICAL PROBLEMS IN DEVELOPING AND PRESENTING
ENVIRONMENTAL QUALITY INDICES TO THE PUBLIC, L. Edwin Coate
and Anthony K. Mason
SAMPLE TRACKING DATA MANAGEMENT SYSTEM, Robert N. Snellinq
Mark J. Madsen and George C. Allison .... 31 3
HAZEL: A HAZARDOUS WASTE DATA BASE, G. L. Ochs 3, 4
SWIRS: THE EVOLUTION OF A SOLID WASTE INFORMATION
RETRIEVAL SYSTEM, John A. Connolly
TOWARDS A COMPUTER-BASED INFORMATION MONITORING SYSTEM FOR
GROUNDWATER DATA IN NEW ZEALAND, Richard 8. McCammon 3I_5
THE URDP/ISRAEL PROJECT FOR ESTABLISHING AIR DUALITY
ASSESSMENT PROGRAM IN ISRAEL, K. H. Jones and Uri Marinov
SESSION 32: THE INTERACTION OF CLIMATE AND POLLUTION
Chairperson: H. E. Landsberg
University of Maryland
College Park, Maryland
ACTUAL AND POTENTIAL EFFECTS OF POLLUTANTS ON CLIMATE
H. E. Landsberg '
32-1
TRENDS IN ATMOSPHERIC PROPERTIES, L. Machta 32_
CLOUD CONDENSATION NUCLEI FROM A PAPER MILL —THEIR EFFECTS
ON CLOUDS AND PRECIPITATION, Edward E. Hindman II and
Peter V. Hobbs '
32-3
INTERACTIONS BETWEEN AIR POLLUTION AND SOLAR RADIATION
J. T. Peterson and E. C. Flowers . , '
32-4
UNCERTAINTIES IN THE APPLICATION OF THEORETICAL MODELS TO
STRATOSPHERIC POLLUTION, Julius S. Chang 32 5
ORGANIZATION OF LONG RANGE AIR POLLUTION MONITORING IN EUROPE
Brynjulf Ottar '
32-6
OZONE IN HIGH SIERRA VALLEYS, Thomas Y. Palmer L R. Smith
and J. Neirinck '
32-7
SESSION 33: THE ASSESSMENT AND EVALUATION OF PROBLEMS ASSOCIATFD
WITH VARIOUS INDUSTRIAL PROCESSES
Chairperson: B. Linsky
Department of Civil Engineering
West Virginia University
Morgantown, West Virginia
#*
A NEWLY DEVELOPED CONCEPTUAL FRAMEWORK ABOUT POLLUTING PROCESSES
THAT HAVE INADEQUATE EMISSION SENSING OR ASSESSMENT, Benjamin Linsky 33-,
MEASUREMENT OF PARTICLE SIZE AND OTHER FACTORS INFLUENCING
PLUME OPACITY, A. Weir, Jr., Dale G. Jones, Larry T. Papay
Seyrr.our Calvert and S. Yunq. ...
33-2
A COMPREHENSIVE ENVIRONMENTAL MONITORING SYSTEM AS A TOOL FOR
IMPACT ASSESSMENT AND OPERATIONAL COMPLIANCE, G. R. Goldgraben
and W. B. Montano
33-3
**
xvii
-------
I DENT IFI CATiON OF EMISSION SOURCES BY INDUSTRIAL FINGERPRINT
James D. Joseph, Michael F. Terraso and W. A. Quebedeaux, Jr
33-4
COOLING TOWER PLUME MEASUREMENTS, L. Winiarski, W. Frick
and B. Tichenor
AN APPROACH TO ASSESSING THE SIGNIFICANCE OF FUGITIVE
PARTICULATE EMISSIONS ESCAPING FROM A COKESIDE SHED,
Louis R. Pa ley and Richard PowaIs
THE DEVELOPMENT OF A METEOROLOGICAL FORECASTING SYSTEM
DESIGNED TO PREDICT SULFUR DIOXIDE CONCENTRATIONS IN AMBIENT
AIR, Robert J. Groves
33-7
33-5
33-6
SESSION 34: FINE PARTICLES - 2
Chairperson: Arthur M. Langer
Mount Sinai School of Medicine
New York, New York
X-RAY DIFFRACTION AND ELECTRON BEAM ANALYSIS OF ASBESTI FORM
MINERALS IN LAKE SUPERIOR WATERS, Philip M. Cook,
Ivan B. Rubin, Carl J. Maggiore and William J. Nicholson 34-1
FINE PARTICLES AND WATER QUALITY IN THE COASTAL MARINE
ENVIRONMENT, J. R. Schubel 34-2
THE BLACK HILLS AS A POSSIBLE SINK FOR AITKEN AND NEAR-CCN
PARTICULATES, Briant L. Davi's, Donald N. Bla'ir,
L. R. Johnson and Steven J. Haggard 34-3
AMMONIUM SULPHATE AEROSOLS, Cyril I Brosset 34-4
EFFECTS OF CHRONIC, CONTINUOUS EXPOSURE TO SIMULATED URBAN AIR
POLLUTION ON LABORATORY ANIMALS WITH CARDIOVASCULAR AND
RESPIRATORY DISEASES, Phyllis M. Hartroft, Charles C. Kuhn III,
Sandra K. Vocelka, Chusak Tansuwan, Robert 0. Gregory and
Richard A. Gardner 34-5
PULMONARY VS NASAL DEPOSITION OF WATER SOLUBLE FINE PARTICULATE,
John J. Godleski and J. Peter Bercz 34-6
Chairperson: M. Bortner
General Electric Company
Philadelphia, Pennsylvania
THE MEASUREMENT OF CARBON MONOXIDE AND METHANE IN THE NATIONAL
CAPITAL AIR QUALITY CONTROL REGION I. MEASUREMENT SYSTEMS,
Peter J. LeBel, Robert A. Lamontagne and Harold W. Goldstein 35-1
THE MEASUREMENT OF CARBON MONOXIDE AND METHANE IN THE NATIONAL
CAPITAL AIR QUALITY CONTROL REGION I I. METEOROLOGICAL CONDITIONS
AND CHROMATOGRAPHIC AND SPECTROMETRIC RESULTS, Robert A. Lamontagne,
John W. Swinnerton, Peter E. Wilkniss, David J. Bressan,
Peter J. LeBel and Harold W. Goldstein 35-2
THE MEASUREMENT OF CARBON MONOXIDE AND METHANE IN THE NATIONAL
CAPITAL AIR QUALITY CONTROL REGION III. CORRELATION INTERFEROMETER
RESULTS, Harold W. Goldstein, M. H. Bortner, Robert N. Grenda,
Robert Dick, Robert A. Lamontagne and Peter J. LeBel 35-3
THE CO CYCLE IN THE ATMOSPHERE, W. Seiler 35-4
HIGH SENSITIVITY AMBIENT AIR CO MONITOR, J. F. M. van Dijk
and R. A. Falkenburg 35-5
CH4, CO MIXING RATIOS IN THE TROPOSPHERE AND LOWER STRATOSPHERE,
R. A. Lamontagne, D. J. Bressan, J. W. Swinnerton and
P. E. WiIkniss 35-6 ,
SESSION 35: THE EVALUATION AND MEASUREMENT OF ATMOSPHERIC CARBON
MONOXIDE AND METHANE
xviii
-------
HUMAN EXPOSURE TO EXCESSIVE LEVELS OF CARBON MONOXIDE DUE
TO TRAFFIC CONGESTION IN THE CITY OF TEHRAN, IRAN,
B. Samimi and H. Nanbakhsh ' „ _
35-7
SESSION 36: QUALITY ASSURANCE
Chairperson: William H. Kirchhoff
National Bureau of Standards
Washington, DC
QUALITY CONTROL AND HARMONIZATION PROGRAMS FOR THE MEASUREMENT
OF ENVIRONMENTAL POLLUTION IN THE EUROPEAN COMMUNITIES
P. Recht, J. Smeets, R. Amavis and A. Berlin . ^
QUALITY ASSURANCE OF STATIONARY SOURCE EMISSION MONITORING DATA
R. E. James and C. D. Wolbach. . . '
. . 36-2
APPLICATIONS OF QUALITY ASSURANCE IN MAJOR AIR POLLUTION
MONITORING STUDIES - CHAMP AND RAMS, Darryl J. von Lehrnden
Raymond C. Rhodes and Seymour Hochheiser ' ^ ^
EPA QUALITY ASSURANCE PERFORMANCE SURVEYS, S. M Brombera
G. G. Akland, B. I. Bennett and T. A. Clark. . . . ! 36_4
VALIDATION OF ENVIRONMENTAL MEASUREMENT METHODOLOGY
John A. Wi nter '
36-5
IMPROVED ANALYTICAL QUALITY ASSURANCE FROM LABORATORY
AUTOMATION, William L. Budde and John M. Teuschler . 6
INTERCOMPARI SON PROGRAM OF SULPHUR DIOXIDE ANALYSES
-Alexandre Berlin, A. M. Price and Steven M. Bromberg 36_7
SESSION 37: THE MEASUREMENT AND ASSESSMENT OF PROBLEMS ASSOCIATED
WITH AGRICULTURAL PROCESSES AbMJUAltD
Chairperson: D. J. Ward
Department of Agriculture
Washington, DC
INCREASED FOOD PRODUCTION AS RELATED TO ENVIRONMENTAL QUALITY
David J . Ward '
OUTPUTS AND DEVELOPMENTS OF A NATIONAL FOOD CONTAMINATION
MONITORING PROGRAMME, A. W. Hubbard and D. G. Lindsay. 37_2
ENVIRONMENTAL IMPACTS OF AGRICULTURE AND THE FUTURE WORLD FOOD
SITUATION: FIRST-ORDER EFFORTS, Lawrence H. De Bivort . ... 37.3
DUST STORMS DUE TO THE DESICCATION OF OWENS LAKE, Roger F. Reinking!
Larry A. Mathews and Pierre St. -Amand ^ ^
CONCENTRATION OF MSMA IN SOILS AND ADJACENT WATERWAYS
J. Walter Mason, Andrew J. Englande, Ann C. Anderson '
A. Assaf Abdelghani and Peter M. Smith ' $7 5
A FIELD STUDY TO DETERMINE THE EFFECTS OF CALCIUM AMENDMENT OF SERPENTINE
SOILS AND ITS INFLUENCE ON PONDEROSA PINE AND DOUGLAS-FIR SEEDLINGS
Ken Lanspa ' 37 6
METHYLATION OF MERCURY IN A TERRESTRIAL ENVIRONMENT
R. D. Rogers ' 37_7
SESSION 38: THE MEASUREMENT AND EVALUATION OF PERSONAL
EXPOSURE MONITORS
Chairperson: M. G. Morgan
Brookhaven National Laboratories
Upton, New York
and Carnegie Mellon University
Pittsburgh, Pennsylvania
INDIVIDUAL AIR POLLUTION MONITORS: AN ASSESSMENT OF RESEARCH
NEEDS, M. Granger Morgan and Samuel C. Morris IO ,
JO- I
xix
-------
MONITORING PERSONAL EXPOSURE: PRESENT AND FUTURE,
Andrew E. O'Keeffe 38-2
EXPOSURE MONITORING AND COMMUNITY HEALTH STUDIES,
J. G. French, G. C. Or+man, P. E. Brubaker and
Ferris Benson 38-3
SEQUENTIAL SAMPLING PLANS AND DECISION THEORY FOR EMPLOYER
MONITORING OF EMPLOYEE EXPOSURE TO INDUSTRIAL ATMOSPHERES,
D. H. Budenaers, Y. Bar-Shalom, K. A. Busch and N. Leidel 38-4
ASSESSMENT OF TOTAL EXPOSURE TO AN AIR POLLUTANT,
Mirka Fugas 38-5
A MINIATURE CONTINUOUS MONITOR FOR THE DETERMINATION
OF BREATHING ZONE AND AMBIENT AIR CONCENTRATIONS OF
TOXIC SUBSTANCES, Byron A. Denenberg, R. S. Kriesel
and R. W- Mi I ler 38-6
PERSONAL MONITORING OF MERCURY VAPOR EXPOSURES, D. L. Braun ¦ 38-7
PLENARY SESSION II LEGAL ASPECTS OF MONITORING
Chairperson: Judge Thomas B. Yost
Administrative Law Judge
U.S. Environmental Protection Agency
Atlanta, Georgia
STATEMENT OF JUDGE YOST . PI 1-1
STATEMENT OF THOMAS H. TRUITT PI I -1
STATEMENT OF JAMES W. JEANS PI 1-3
STATEMENT OF JOHN P. HILLS PI 1-5
DISCUSSION WITH AUDIENCE PARTICIPATION PI 1-7
CONCLUDING STATEMENT OF JUDGE YOST Pll-ll
PLENARY SESSION III FUTURE OF ENVIRONMENTAL QUALITY ASSESSMENT
Chairperson: Frank Clarke
Senior Scientist
Office of the Director
U.S. Geological Survey
Washington, D.C.
NATURAL SYSTEMS ANO ASSESSING FUTURE ENVIRONMENTAL QUALITY
FOR MAN, M. GORDON WOLMAN PI I I-1
FUTURE ENERGY DEVELOPMENT AND RELATED ENVIRONMENTAL
MONITORING, STEPHEN J. GAGE Pill-.5
FUTURE ENVIRONMENTAL MONITORING AS IT RELATES TO
HUMAN HEALTH, VAUN NEW ILL Pill-II
xx
-------
ABBREVIATED SESSION TITLE INDEX
TITLE SESSION NO.
Agriculture 37
Ambient Monitoring 28
Carbon Monoxide 35
Data Management I 25
Data Management II 31
Design of Monitoring Systems I 3
Design of Monitoring Systems II 8
Energy Extraction 17
Env i ronmentaI Mode Ii ng I 14
EnvlronmentaI Modeling II 20
Global Monitoring 26
Groundwater 1 9
Groundwater II 15
Halogenated Organics 2
Heavy Metals I I
Heavy Meta I s I I 6
Industrial Processes 33
Inorganics I 12
Inorganics II 18
Marine Environment 21
Nuclear Fuel Processing 19
Organ Ics I 24
Organ ics II 30
Particulate Matter I . . . 29
Particulate Matter II 34
Particulate Monitoring 5
Personal Exposure Monitoring 38
Pesticides I ,7
Pesticides II 13
Pol lutlon and CI imate 32
Quality Assurance 36
Remote Sens i ng I 10
Remote Sensing II 16
Sulfates * 23
Technology and Government II
Transportation I, General 22
Transportation II, Aircraft 27
Waste Disposal 4
xxi
-------
AUTHOR INDEX
(37 - 5) = Session and Paper Number
(PI Monday) = Plenary Session - Monday a.m.
(PII Friday] § (PIII Friday) = Plenary
(PA - 1) = Panel Member § Session Number
(C - 1 - 5) = Chairperson and Session
* = Principal Author
** = Manuscript not Received in Time for
Publication
Abdelghani, A. A. (37-5)
Ahlquist, N. C. (12-6)
Ahn, S. H. (4-5)
Akland, G. G. (36-4)
Allison, G. C. (31-3)
Allwine, K. J. (18-3) (21-4)
Altpeter, L. L. * (5-7)
Amavis, R. (36-1)
Anderson, A. C, (37-5)
Andersson, I. (25-2)
**Appel, B. R. * (23-3)
Austin, Thomas S. * (25-1)
' Ballentine, T. * (17-1)
Banks, H. 0. * (9-1)
Bar-Shalom, Y. (38-4)
Barker, D. B. * (16-6)
Barth, D. S. (PI)
Behar, J. V. * (20-1)
Belcher, N. M. (30-7)
Bennett, B. I. (36-4)
Benson, F. (38-3)
Bercz, J. Peter (34-6)
Berlin, A. * (29-3)* (36-7) (36-1)
**Bernstein, David (29-2)
Bianco, E. (18-1)
Bingham, F. * (30-6)
Black, M. S. * (27-3)
Blair, D. N. (34-3)
Blazowski, W. S. (27-1)
Blejer, H. P. (1-6)
Boes, R. J. (14-6)
Bourdeau, P. (29-3)
Bortner, M. H. (35-3) (C-35)
Bosch, John * (2S-3)
Bowker, D. E. * (16-3)
Braun, D. L. * (38-7)
Breitenbach, L. (24-4)
Bressan, D. J. (35-2)(35-6)
Bretthauer, E. W. * (19-1) (C-19)
Bridboard, K. * (1-6)
Bromberg, S. M. * (36-4) (36-7)
Brooks, J. J. (27-3)
Brooks, J. N. (16-6)
Brooks, R. L. * (17-4)
Brosset, Cyrill * (34-4)
Brown, Dail W. (3-5) (C-25-31)
Brubaker, P. E. (38-3)
Bryson, R. J. (27-6)
Budde, W. L. * (36-6)
Budenaers, D. H. * (38-4)
Bufalini, Joseph J. (16-S)
Bulger, Carl (PA-13)
Bundy, D. H. (10-3)
Bunner, D. R. (15-2)
Burke, Jerry A. * (7-1)
Busch, K. A. (38-4)
Bustueva, V. A. * (12-1)
Butler, Philip (PA-13)
Cairns, J., Jr. (18-4)
Calvert, Seymour (33-2)
Canales, L. (24-6)
Cannon, D. (2 3-2)
Cantillo, A. Y. (6-5)
Cantrell, B. (29-1)
Carey, Ann (PA-13)
Carnes, Richard A. * (15-2)
Caro, Joseph H. * (7-3)
Carver, Thomas C., Jr. * (13-4) (PA-13)
Castagnino, Walter (8-1)
Cavanagh, L. A. * (18-5)
Cerquiglini, Susana * (18-1)
Chan, H. S. (21-1)
**Chang, J. S. * (32-5)
Chang, R. K. (23-5)
Charlson, R. J. * (12-6)
Chen, Kenneth (6-7)
Citron, Robert * (3-6)
Clark, T. A. (36-4)
Clarke, Frank (C-P III Friday)
Clarkson, Thomas * (1-S)
Coate, L. Edwin * (31-2)
Cochrane, W. P. * (7-4)
Collis, R. T. H. (12-3)
Conkle, J. P. * (27-2)
Connolly, John A. * (31-5)
Cook, Philip M. * (34-1)
Corneliussen, Paul (PA-13)
Covert, D. S. (12-6)
Cox, G. (C-4)
Cox, W. G. * (4-6)
Crabtree, James H. (24-3)
Crouch, R. L. * (9-2)
Cruitt, J. R. (25-5)
Cunningham, J. E. * (18-6)
Dabberdt, W. F. * (14-2)
**Daush, A. (PA-11)
Dave, J. M. * (3-1)
Davidson, C, I. * (6-3)
Davis, B. L. * (34-3)
Davis, D. D. * (16-7) (23-7)
Davis, G. (16-2)
Davis, M. G. (27-6)
De Jaegere, R. (6-4)
DeBivort, L. H. * (37-3)
DeCarlo, Vincent (2-3)
DeMandel, R. E. (12-4)
Demerjian, Ken * (20-2)
Denenberg, B. A. * (38-6)
Dennis, Reid (13-3)
DeShaw, James R. (13-3)
Dewiest, F. (24-5)
Dewit, A.. * (6-4)
Di Ferrante, E. (29-3)
Dick, R. (35-3)
Dickson, K. L. (18-4)
Dieterick, B. N. (PI)
Donahue, Dan (PA-13)
Donaldson, W. T. * (15-5)
Dorko, W. D. (2-5)
Dowd, G. (2 2-6)
Dunn, L. M. (20-1)
Dunsmore, D. A. * (14-6)
Eatough, D. J. (23-2)
Eckert, J. A. * (10-3)
Eckert, R. D. (9-2)
Echeverria, Eduardo (PI)
Egan, B. A. (14-1)
**Eisenbud, M. (29-2)
Elfers, Lawrence A. * (2-2)
**Ellis, Mike (15-7)
Eloranta, E. W. (10-4)
Emerson, Raymond * (6-7)
Englande, A. J. (37-5)
Enos, Henry (C-7-13)
Epstein, S. (30-7)
Eschenroeder, Alan (C-14-20)
Everett, L. G. (9-3)
Falconer, P. D. * (26-5)
xxii
-------
Falkenberg, R. A. (35-5)
Farvar, M. T. * (7-7)
*Fassel, V. A. (15-3)
Feldstein, M. (12-4)
Feltz, H&jman (PA-13)
Few, J. D. * (27-6)
Ficke, J. F. (3-3)
Findlay, W. J. * (22-6) (24-2)
Fine, D. H. * (30-7)
Fiorentina, H. Delia (24-5)
Fisher, P. D. * (25-S)
Flowers, E. C. (32-4)
French, J. G. * (38-3)
Freund, S. M. (Z-5)
Friberg, Lars (C-l-6) * (1-0)
Frick, W. (33-5)
Friedlander, S. K. (6-3)
*FrIend, J. P. * (23-7)
Fugas, M. * (38-5)
Funk, M. (17-6)
Gaarenstroom, P. (5-4)
Gage, Stephen J. (P III Friday)
Gardiner, Richard A. (34-5)
Gay, Bruce W. * (16-5)
Gay, D. D. * (1-7)
Giam, C. S. * (21-1) (C-21)
Giammar, R. D. (22-3)
*Giauque, R. D. (23-3)
*Gille, J. C..* (26-1) (C-26)
Godleski, John J- * (34-6)
*Gillespie, R. E. (2-7)
Goldgraben, G. R. * (33-3)
Goldman, A. (16-6)
Goldstein, H. W. * (35-3) (35-1) (35-2)
Green, Fitzhugh (PI)
Gregory, Robert 0. (34-5)
Grenda, R. N. (35-3)
Griggs, M. (10-1)
Groth, David H. * (6-1)
Groves, Robert J. * (33-7)
Guagliardo, J. L. (10-3)
Guldberg, P. H. (22-4)
Gustafsson, U. R. C. (26-4)
Haddad, R. * (8-1)
Haggard, S. (34-3)
Hahn, Hermann, H. * (14-5)
Hahn, P. B. (19-1)
Hall, J. S. (5-5)
Hallett, D. * (7-6)
Hameed, S. * (14-4)
Hampton, N. F. * (9-5)
Hansen, L. D. * (23-2)
Hanst, Philip L. (16-5)
Harris, Daniel J. * (4-7)
Harrison, H. * (23-1)
Hartroft, Phyllis M. * (34-5)
Heaps, W. (16-7)
Heath, Bob (PA-13)
Heath, D. F. (26-6)
Heesen, T. C. (4-1)
Henderson, W. R. * (10-7)
Herbert, R. P. (22-4)
Herget, W. F. (10-2)
Hering, S. V. (6-3)
Hershaft, Alex * (16-1)
Herter, Christian A. Jr. (PI)
Hidy, G. (C-23)
Hilborn, J. C. * (24-2)
Hilfiker, Ronald, C. (18-2)
Hills, John P. (P II Friday)
Hintfman, E. E. * (32-3)
Hinkley, D. (C-16)
Hobbs, P. V. (32-3) (23-1)
Hochheiser, S. (36-3)
Hodgson, F. N. (27-4)
Hoell, J. M., Jr.* (10-6)
**Hoffer, E. M. (23-3)
**Holt, D. E. * (21-2)
Hopf, S. B. (23-6)
Horling, J. E. (27-5)
Hower, J. (20-4)
Hubbard, A. W. * (37-2)
Huckabay, G. W. (19-4)
Huffaker, R. M. (10-2)
Hughes, Ernest * (2-5)
Huntzicker, J. J. * (23-4)
Husar, J. D. (5-2)
Husar, R. B. (5-2) ,
Hutchinson, J. M. R. (19-5)
Hyne, E. * (28-4)
Inhaber, Herbert * (31-1)
Isabelle, L. (23-4)
James, R. E. * (36-2)
Janssen, H. Erie * (13-3)
Jeans, James W. (P II Friday)
Jenkins, Dale * (1-1)
Jensen, Clayton E. * (3-5)
Jensen, T. E. (23-2)
Johnson, J. D. (8-5)
Johnson, L. R. (34-3)
Jones, D. * (17-5)
Jones, D. E. (19-4)
Jones, Dale G. (33-2)
Jones, K. H. * (31-7)
Jones, J. E. * (25-6)
Jones, P. W. (22-3)
Joseph, J. D. * (33-4)
Kaloyanova, F. * (13-2)
Keesee, R. G. * (23-6)
Keffer, William J. (4-7)
Key, J. W. (18-6)
Kinnison, R. R. (20-1)
Kirchhoff, W. H. (C-36)
Kiyoura, Raisaku * (12-7)
Klarman, A. F. * (27-5)
**Kleinman, M. (29-2)
Klemas, V. * (16-2)
Klute, R. (14-5)
**Kneip, T. J. * (29-2)
**Kniseley, R. N. (15-3)
Konikow, Leonard S. * (20-3)
Kosters, J. J. (16-6)
**Kothny, E. L. (23-3)
Kramer, H. * (28-1) (C-28)
Kreikebaum, G. * (5-6)
Kriesel, R. S. (38-6)
Kuhn, Charles C., III (34-5)
Kunkel, K. E. * (10-4)
Kuranoto, M. (2 4-3)
Kutz, Frederick (PA-13)
Lackey, W. W. (27-2)
Lamontagne, R. A. * (35-6)* (35-2)
(35-1) (35-3)
Landers, R. (17-5)
Landreth, R. E. (15-2)
Landrigan, P. J. (1-6.0
Landsberg, H. E. * (32-1) (C-32)
':<*Langer, A. M. * (29-4) (C-29-34)
Langford, R. Hal (C-3-8)
Lanspa, K. * (37-6)
Lao, R. C. * (2-6)
Larson, T. V. (23-1)
Lavery, T. F. (14-1)
Law, S. L. (4-3)
Lebedeff, S. A. (14-4)
LeBel, Peter J. * (35-1) (35-2) (35-
Leidel, N. (38-4)
Lem, P. N. (20-1)
Lemen, R. A. (1-6)
Levaggi, D. A. * (12-4)
Liang, T. (15-1)
**Lindelin, K. R. * (2-7)
xxiii
-------
Lindsay, D. G. (37-2)
*Linsky, B. * (33-1) (C-33)
Lodge, James (C-5)
Lu, James (6-7)
Ludwig, C. B. * (10-1)
Ludwig, F. L. * (12-3)
Lysyj , I. * (4-2)
Machta, L. * (32-2)
Macias, E. S. * (5-2)
Madsen, M. J. (31-3)
^Maenhaut, W. (26-7)
Mage, David T. * (20-5)
Maggiore, Carl J. (34-1)
Maker, P. (24-4)
Malm, W. C. * (5-5)
Mangelson, N. F. (23-2)
Manheimer, B. (C-ll)
Mann, W. B. (19-5)
Marcy, B. C. (17-3)
Marinov, Uri. (31-7)
Marino, F. F. (C-22)
Marr, H. E. * (4-3)
Martin, C. L. (27-2)
Martins, G. C. * (24-6)
Mason, Anthony K. (31-2)
Mason, J. W. * (37-5)
Massart, D. L. (6-4)
Mathews, L. A. (37-4)
Matt son, D. (2-7)
Mayrsohn, Henry * (24-3)
McCammon, Richard B. * (31-6)
McClenny, W. A. * (28-6)
McCrone, W. C. * (5-1)
McDermott, D. J. (4-1)
McElroy, J. L. (10-3) (20-1)
McGee, T. (16-7)
McGregor, W. K. (27-6)
McGuire, J. M. * (15-4)
McMahon, Bernadette (7-1)
McMillion, L. G. (9-1)
McNamara, M. M. (PA-11)
McNesby, James (C-2)
Melfi, S. H. (10-3) (C-10)
Menzies, R. T. * (22-5)
Miller, A. J. (26-6)
Miller, C. R. * (10-2)
Miller, J. M. (26-5)
Miller, R. L. (27-2)
Miller, R. W. (38-6)
Mitchell, G. A. (30-6)
'¦'Mitrovic, Srdjan * (14-3)
Monig, F. J. (5-3)
Monkman, J. L. (2-6)
Montano, W. B. (33-3)
Morgan, G. B. (20-6)
Morgan, James * (1-2)
Morgan. M. G. * (38-1) (C-38)
Moriarty, A. (16-7)
Morris, S. C. (38-1)
Moseman, Robert F. * (13-1)
Moyers, J. L. (23-6) (5-4)
*Mroz, G. (26-7)
Mueller, P. (C-23)
Mullen, P. A. (19-5)
Murcray, D. G. (16-6)
Murcray, F. H. (16-6)
Myers, R. L. * (8-6)
Nagatani, R. M. * (26-6)
Nanbakhsh, H. (35-7)
Neff, G. S. (21-1)
Neirinck, J. (32-7)
Nelson, K. W. * (12-5)
*Neti, R. M. (28-7)
Newill, Vaun (P III Friday)
Neylan, D. L. (4-3)
Nicholson, William J. * <"29-6) (34-1)
Niki, Hiromi * (24-4)
Noonan, Richard C. (16-5)
Norton, P. (28-4)
Noyce, J. R. * (19-5)
0'Dell, K. (5-5)
O'Keeffe, A. * (38-2)
O'Toole, Donna M. * (18-2)
Ochs, G. L. * (31-4)
Oguri, Mikihiko (6-7)
Ondov, J. M. (17-2)
Ortman, G. C. (38-3)
Ott, Wayne R. (20-5)
Ottar, B. * (32-6)
Ozolins, G. * (3-4)
Pack, D. H. * (26-3)
Paley, L. R. * (33-6)
Palmer, T. Y. * (32-7)
Palmlund, I. * (25-2)
Papay, Larry T. (33-2)
**Parker, P. L. * (21-5)
Parra, C. G. (17-4)
Pavlov, A. S. X(PI)
**Patterson, C. C. * (21-3)
Perkins, P. J. * (26-4)
Perone, S. P. * (5-4)
Peterson, J. T. * (32-4)
Petrus, C. A. * (15-6)
Peyton, B. J. (10-5)
Philipson, W. R. (15-1)
Phillips, C. A. (19-2)
Pichler, M. (5-4)
Pickering, R. J. * (3-3)
Pijanowski, B. S. * (21-6)
Pipes, W. 0. * (4-5)
Pitts, J. N., Jr. * (24-1) (C-24-
Powals, R. (33-6)
Pressman, A. (17-5)
Price, A. M. (36-7)
Price., J. H. * (8-5)
Primack, J. (PA-11)
Prinz, B. * (20-4) *(25-4)
Putnam, R. D. (12-5)
Quebedeaux, W. A., Jr. (33-4)
Quickert, N. (22-6) (24-2)
Quigley, S. T. (PA-11)
Radziul, J. V. (30-1)
Ragaini, R. C. * (17-2) *(19-4)
Rasmussen, R. * (21-4) (18-3)
**Rauschuber, Don * (15-7)
Reagan, James A. (8-6)
Recht, P. * (36-1)
Rehg, W. R. (27-3)
Reinking, R. F. * (37-4)
Rhodes, R. C. (36-3)
Richter, Harold G. (2-2)
Ritchie, I. * (28-2)
Roan, Clifford (PA-13)
Robinson, B. * (19-2)
Robinson, E. (18-3) (C-12-18)
Rogers, R. D. * (37-7)
Rohl, A. N. (29-6)
Rondia, D, * (24-5)
Roquemore, W. M. * (27-4) (C-27)
Rosen, R. H. * (3-2)
Rosenblum, H. S. * (14-1)
Roulier, M. H. (15-2)
Rounbehler, D. P. (30-7)
Rubin, Ivan B. (34-1) (29-4)
Ruff, R. E. (18-5)
Rugg, D. D. (9-2)
Ruopp, Doris J. * (2-3)
Samimi, Behzad * (35-7)
Sandberg, J. S. (12-4)
Sanders, W. M. * (14-7)
Sangrey, D. A. * (15-1) (C-15)
Savage, C. (24-4)
xxiv
-------
Sawyer, R. * (29-5)
Schere, K. L. (20-2)
**Schiff, R. (16-7)
Schmidt, X. D. * (9-4)
Schneider,-T. * (8-3)
Schubel, J. R. * (34-2)
Schuck, B. A. * (20-6)
Schwarz, F. P. * (30-2)
Seals, R. K., Jr. * (10-5)
Segar, D. A. * (6-5)
Seiber, James N. * (7-2)
Seiler, W. * (3S-4)
Selikoff, I. (C-29-34)
Seltz, J. (28-2)
**Settle, D. M. (21-3)
Shofner, F. M. (5-6)
Shumate, M. S. (22-S)
Siegel, R. D. * (22-4)
Sievers, R. E. (27-3)
Simpson, F. * (9-7)
Sklarew, R. C. * (28-3)
Slawson., G. C. Jr., * (17-3)
Smeets, J. (29-3) (36-1)
Smith, G. E. (25-6)
Smith, L. R. (32-7)
Smith, P. M. (37-5)
Snelling, R. N. * (31-3)
Soesanto, Sri Soewasti * (8-7)
Sonnenschein.'C. M. (10-2)
Soule, Dorothy F. (6-7)
St. Amg.nd, P. (37-4)
Stafford, R. G. * (23-5)
Stalling, David L.*(7-5)
Stanford, T. B. (22-3)
Stedman, Donald * (22-2)
Stewart, R. W. (14-4)
Stober, W. * (5-3)
Stratmann, K. (25-4)
Strap, P. E. * (22-3)
Stumpf, S. A. * (27-1)
Suess, M. J. * (30-4)
Suffett, I. H. * (30-1)
Swirmerton, J. W. (35-2) (35-6)
Tannahill, G. K. (8-5)
Tansuwan, Chusak (34-5)
Taylor, Alan W. (7-3)
Taylor, C. E. * (4-4)
Teng, W. L. (15-1)
Terraso, M. F. (33-4)
Teuschler, J. M. (36-6)
Thomas, D. M. (28-5)
Thomas, R. S. (2-6)
Thompson, R. T,, Jr. (10-6)
Tichenor, B. (33-5)
Tinlin, Richard (C-9)
Todd, 0. K. *(9-3)
Tornatore, G. (16-2)
Trabosh, Harold (PA-13)
Train, Russell, E. (PI)
Truitt, Thomas H, (P II Friday)
**Turner, G. S. * (28-7)
Tyler, Germund * (1-3)
Van Allen, J. (16-6)
van Dijk, J. F. M. * (35-5)
van der Schalie, W. H. (18-4)
VanDerwalker, J. (C-17)
Varner, M. 0. (12-5)
Verdin, A * (30-3)
Vocelka, Sandra X. (34-5)
von^Lehmden, D. J. * (36-3)
Vukovich, F. M. * (8"4)
Wade, W. R. (10-6)
Waggoner, A. P. (12-6)
Wagoner, J. K. (1-6)
Wallen, I. E. * (2-1)
Wallis, R. R. (8-5)
Wang, H. (16-2)
Ward, D. J, * (37-1) (C-37)
Ward, P. (4-5)
Washington, W. M. * (26-2)
Wasik, S. P. (30-2)
Watson, J. C. (23-4)
Watson, R. B. (12-5)
Weinman, J. A. (10-4)
Weir, A., Jr. * (33-2)
Weisman, I. (29-6)
Weiss, B. * (19-3)
**Wesolows)ci, J. J. (23-3)
Wessel, John (PA-13)
Westberg, H. * (18-3)
Westendorf, W. (19-2)
Vfestlake, G. F. * (18-4)
Wheatley, A. C. (8-5)
Whelan, W. (16-2)
Whitby, K. T. * (29-1)
White, Don (PA-13)
Wilkniss, P. E. (35-2) (35-6)
Williams, W. J. (16-6)
Wilson, J. C. (28-3)
Wilson, L. G. * (9-6)
R. X, * (15-3)
Winiarski, L. * (33-5)
Winter, John * (36-5)
Witte, W. G. (16-3)
Wolbach, C. D. (18-6) (36-2)
**Wolf£, M. S. (29-4)
Wolraan, M, Gordon (P III Friday)
Woodrow, J. E. (7-2)
Woodsum, H. C. * (19-6)
Yeager, M, A. (12-5)
Yocum, John E. * (12-2)
Yost, Thomas B. (C-P II Friday)
Young, D, R. * (4-1)
Yung, S. (33-2)
2app, John A. * (2-4)
Zeldin, M. D. * (28-5)
Zeliger, H. I. * (17-6)
**ZolIet, W. *(26-7)
XXV
-------
Accu-Labs Research, Inc.
i1485 W. 48th Ave.
Wheat Ridge, Colorado 80033
AeroVironment, Inc.
145 Vista Ave.
Pasadena, California 91101
AID, Analytical Inst. Devel., Inc.
Rt. 41 & Newark Rd.
Avondale, Pennsylvania 19311
American Chemical Society
i155 16th St., N.W.
Washington, O.C. 20036
Analog Technology Corporation
3410 East foothill Blvd.
Pasadena, California 91107
Analytical Biochemistry Labs, Inc.
P. O. Box 1097
Columbia, Missouri 65201
Bausch & Lomb, Inc.
Box 543
Rochester, New York (4602
&
441 Cienaga Dr.
Fullerton, California 92635
Bechman Instruments, Inc.
2300 Harbor Blvd.
Fullerton, California 92634
Bench'x Corporation
Process Instruments Division
P. 0. Drawer 831
Lewisburg, West Virginia 24901
Contel Corporation
1779 VaI tec Lane
P. 0. Box 3465
Boulder, Colorado 80303
Controls for Environ. Pollution, Inc.
1925 Rostra St.
P. 0. Box 5351
Santa Fe, New Mexico 87501
Dasibi Environmental Corp.
616 E. Colorado
Glendale, California 91205
Environmental Measurements.Inc.
215 Leidesdorff St.
San Francisco, California 941 I I
Environmental Res. STech., Inc./
NERA, Inc.
696 Virginia Road
Concord, Massachusetts 01742
EXHIBITORS
Flanders Filters, Inc.
845 Hinckley Rd.
Burlingame, California 94010
Hewlett Packard
Rt. 41
Avondale, Pennsylvania 19311
Horiba Instruments, Inc.
1021 Duryea Avenue
Santa Ana, California 92705
Hydro!ab Corporation
P. 0. Box 9406
Austin, Texas 78766
The Inst, of Electrical &
Electronics Engineers, Inc.
345 E. 47 Street
New York, New York 10017
International Plasma Corp.
31 159 San Benito St.
Hayward, California 94544
Kahl Scientific Instrument Corp.
P. 0. Box I 166
El Cajon, California 92022
Lear Siegler, Inc.
Environmental Technology Div.
74 Inverness Orive E.
Englewood, Colorado 80110
Mast Development Co.
2212 E. 12th St.
Davenport, Iowa 52803
McMillan Electronics Corp.
I 1950 Jo I Iyv iIle Rd.
Austin, Texas 78759
MDA Scientific, Inc.
808 Busse Highway
Park Ridge, Illinois 60068
Meloy Laboratories, Inc.
671 5 Electronic Or.
Springfield, Virginia 22151
Meteorology Research, Inc.
464 W. Woodbury Rd.
Altadena, California 91001
Monitor Labs, Incorporated
4202 Sorrento Valley Blvd.
San Diego, California 92121
Philips Electronic Instruments, Inc.
750 South Fulton Ave.
Mount Vernon, New York 10550
Radian Corp.
8500 Shoal Creek Blvd.
P. O. Box 9948
Austin, Texas 78766
Source Gas Analyzers, Inc. (SGA)
725 I -C Garden Grove Blvd.
Garden Grove, California 92641
Spectrex Corporation
3594 Haven Avenue
Redwood City, California 94063
Swagelok Distributors
2229 Via Guadalana
Palos Verdes Estates, Cal. 90274
Teledyne Gurley
514 Fulton St.
Troy, New York 12181
Teledyne Energy Systems
110 W. T I mon i urn Road
Timonium, Maryland 2)093
Thermo Electron Corp.
101 First Ave.
Valtham, Massachusetts 02154
Anti-Pollution Technology
Division of Thermotron Corp.
Holland, Michigan 49423
U.S. Geological Survey
12201 Sunrise Valley Dr.
Reston, Virginia 22092
WeatherMeasure Corp.
P. 0. Box 41257
Sacramento, California 95841
Wilks Scientific Corporation
140 Water Street
South Norwalk, Conn. 06856
Center for Int'l. Environ. Inform.
345 East 46th St.
New York, New York 10017
Xonics, Inc.
6862 Hayvenhurst Avenue
Van Nuys, California 91406
E0C0M Corporation
19722 Jamboree Blvd.
Irvine, California 92664
Fed. Working Grp. on Pest. Mgmt.
Rm. E43I, (WH-566)
401 M Street, S.W.
Washington, D.C. 20460
Rockwell International
Air Monitoring Center
242IA West Hi I Icrest Drive
Newbury Park, California 91320
xxvi
-------
PLENARY SESSION I. INTRODUCTION AND KEYNOTE
September 15, 1975
D. S. Barth
Director
Environmental Monitoring and Support Laboratory-Las Vegas
U.S. Environmental Protection Agency-
Las Vegas, NV
I do not believe it is possible to over-emphazise
the importance of environmental monitoring to any envi-
ronmental or health protection action plan for protec-
tion from adverse effects of environmental pollutants.
Accurate monitoring data are essential for elucidation
of the following aspects of the general environmental
or health protection problem:
1. Detection, identification, and measurement of
environmental pollutants likely to be causing unaccept-
able adverse effects;
2. Quantification of the exposure-effect relation-
ships for the pollutants in populations of receptors
to be protected;
3. Development of linkages between important source
emissions of the pollutants and ambient levels;
4. Development and implementation of cost-effective
source emission control plans to reduce ambient levels
of the objectionable pollutants to acceptable levels;
5. Development of adequate compliance monitoring
systems to insure that control plans can be properly
enforced; and
6. Development of adequate cost-effective monitoring
networks to verify that desired control is being
achieved.
During the course of this Conference all of the
topics briefly outlined here will be at least touched
by the many papers to be presented. It is the hope of
the organizers that the program is sufficiently rich to
be of some value to each of the participants.
As you know, environmental pollutants do not
respect national or international boundaries. Similar-
ly the development of adequate environmental monitoring
systems for these pollutants should not be limited to
single national jurisdictions. It is imperative that
common systems for quality assurance of monitoring data
be adopted by all data collectors. When this has been
accomplished, environmental monitoring data which are
accurate and comparable may move across national or
international boundaries as freely as the environmental
pollutants themselves now do.
The accomplishment of this ultimate goal will
enable all nations to share, in a common scientific
language, their knowledge of environmental pollutants
to include their source and ambient concentrations,
their health and environmental effects in relationship
to the ambient concentrations, and their control in
local and regional situations as well as national ones.
It is only a short step from this sharing of knowledge
to the development of coordinated environmental pollu-
tion control plans which will ultimately improve the
quality of life for all of us.
Hopefully this Conference may contribute at least
one step toward achieving a common environmental moni-
toring data base which may be shared by all nations.
1
PI
-------
*
REMARKS BY THE HONORABLE RUSSELL E, TRAIN
September 15, 1975
Administrator
Environmental Protection Agency
Washington, DC
It Is an honor and a pleasure to be here today at
this distinguished, international gathering of environ-
mental scientists and administrators.
The fact that nearly forty countries are repre-
sented at this gathering demonstrates clearly the grow-
ing international concern for protecting and improving
our earth's fragile remaining resources.
More and more, we are awakening to the fact that,
in the complex web of the world's ecosystem, no nation
can subscribe to the notion that it is an island unto
itself, isolated, insulated and safe.
You may have read recently that the United States
and the Soviet Union have jointly presented to the
Geneva Disarmament Conference a draft treaty that would
ban catastrophic means of waging environmental warfare.
It is a modest beginning in the right direction
and reflects an underlying recognition that what af-
fects one, affects all, I believe this is an impor-
tant step.
But we have begun to understand that the inadver-
tent, mindless act of a peaceful party can have conse-
quences that may turn out to be just as disastrous as
the overt, lethal act of a warring party.
As you know, a committee of prominent scientists,
under the auspices of the National Academy of Sciences,
recently reported the results of a two-year study into
changing trends in the earth's climate. They conclud-
ed, among other things, that we really do not know what
makes the climate change or what effects our activities
have upon the climate, but that we had better hurry up
and find out before it is too late. The report sug-
gested that—in the words of a New York Times story—
"global use of land for agriculture, water for irriga-
tion and drinking, and air and watersheds for waste
disposal was approaching the limits." In short, the
report said: "Our vulnerability to climatic change is
seen to be all the more serious when we recognize that
our present climate is in fact highly abnormal, and
that we may already be producing climatic changes as a
result of our own activities." In view of the south-
ward march of the Sahara, the current drought in parts
of Asia and in central Africa, and the fact that even
slight climatic variations can upset the already ex-
tremely precarious balance between food needs and food
supply, the need to find out what our influences are
upon the climate is a matter of utmost urgency.
Again on a more global level, scientists have
recently warned of the possibility that flurocarbons
from aerosol cans may be destroying the primary shield
we have against the ultraviolet rays of the sun—the
band of ozone that envelopes the earth. It would, I sup-
pose, he an appropriate irony if, after all the concern
over nuclear holocaust, we should finally be undone by
the sprays we use to hold our hair in place and hide our
underarm odors. W.H. Auden once wrote: "Our world will
be a safer and healthier place when we can admit that to
make an atomic bomb is morally to corrupt a host of in-
nocent electrons below the age of consent." I'm sorry
•k
Delivered by Dr. Wilson K. Talley, Assistant Adminis-
trator for Research and Development, EPA.
he isn't here to tell us what he thinks of the aerosol
can.
With each passing year the need to control the
increasing quantity of toxic substances in our environ-
ment intensifies. An estimated 500 to 700 new chemicals
enter commerce in significant quantities every year.
Substances once considered safe for widespread use are
suddenly suspect and pulled off the market. In too
many cases, the public and the environment continue to
serve as testing grounds for such products. The more
we learn about the health effects of pollutants, the
worse things look. Researchers at the National Cancer
Institute are reported to have estimated that 60 to 90
percent of all human cancers are caused by a broad
array of environmental factors—including our diet,
personal habits such as smoking, and products and pro-
cesses such as ultraviolet rays, plastics and pesti-
cides. And while progress has been made in treating
this disease, it is obvious that the most sensible
course lies in prevention, in controlling carcinogens
before they enter man's environment.
Finally, in this loose litany of unfortunate human
interventions, there is the matter of the bees. You
may have read the press accounts that appeared several
months ago concerning the fact that, both in this coun-
try and across the globe, the steady decline in bee
populations—principally as a result of the widespread
use of pesticides—has begun to pose a very real threat
to food supply.
What these assorted items—and I could cite count-
less others—add up to is this: the fundamental fact
of life is that birds, bees and other species, that the
biosphere with its vast and varied life, cannot be
plundered at our pleasure except at our peril. Their
life is linked to ours in intimate and intricate ways
that we have barely begun to understand. And we must,
for survival's sake—if nothing else—find ways of
living and growing that do not render the natural world
which sustains and supports us increasingly inhospita-
ble, if not uninhabitable.
Indeed, compared with our skill and sophistication
in creating pollution, our ability and instruments for
comprehending and controlling it must rank somewhere at
the level of the Dark Ages.
We must make enormous strides in the months and
years ahead to bring our environmental skills and tech-
nology out of the dark and into the post-industrial,
scientific age. To this end, we need far-sighted sci-
entists and creative administrators to develop inte-
grated pollution monitoring and control systems for the
future.
We meet here today together to draw up an agenda
for reaching this goal, focusing as we do, on the twin
themes of monitoring toxic pollutants which affect
human health, and the application of science and tech-
nology to monitoring pollutants which change or affect
our environment, both on a regional and global basis.
The importance of this conference is reflected both by
the distinguished audience and also by the diversity of
conference sponsors. These include, in addition to the
United States Environmental Protection Agency, three
other United States agencies, five professional
PI
-------
societies, the University of Nevada, located here in
Las Vegas, and the World Health Organization. The
participation of the World Health Organization under-
scores the worldwide concern for environmental
protection.
Here, let me suggest some ingredients which should
be a part of any global agenda for addressing the
future of scientific advancement in the environmental
field.
We will need, first of all, to identify all poten-
tially dangerous pollutants. Once this is done, and
that is by no means a simple task, a detailed assess-
ment of the health and welfare effects of these pol-
lutants must be established using scientific techniques
and principles. When these adverse effects have been
documented, a complete inventory and accurate measure-
ments of their effluent concentrations must be estab-
lished. Having done this, we will have a more complete
knowledge of the total emissions of specific pollutants
into the environment and an estimate of their poten-
tial environmental impact.
However, this is only the beginning. We must
further identify the pathways, the transport mechanisms
and chemical transformations which occur as these pol-
lutants move and interact throughout the ecosystem.
The process of identifying the pathways provides us
with the ability to make an accurate estimate of the
pollutant dose to the most susceptible element of the
ecosystem- We will then be in a position to scientifi-
cally determine whether there is a need for control and
where the control action could be most effectively
applied.
Such an evaluation is a continuing process. Obvi-
ously, in order to perform this analysis we will need
to have reliable monitoring systems in all areas. We
can no longer depend upon outdated monitoring systems
to meet tomorrow's needs. In fact, I don't believe it
is unduly alarmist to say that we must design optimum
monitoring systems and networks to provide the data
we need before it is too late, both for our world and
for our civilization.
Both the assessment of health-related aspects of
pollution and the monitoring of pollutants are major
areas to be discussed during the next four days. The
complexity of this task requires not only continued
emphasis on improving methods with which most of us are
familiar, but also on developing innovative ways of
looking at old problems.
In the design of these optimum monitoring systems
we need an integrated approach which looks at the whole
ecosystem. We cannot clean up the atmosphere at the
expense of polluting the waters and the land. We need
to design monitoring systems and networks in a system-
atic, scientific manner which take into consideration
the emission of pollutants, the movement of pollutants
through the ecosystem, the assessment of the exposure
of humans to the pollutant so that the information
generated will enable us ultimately to make intelligent
decisions. And we must be consistently alert to the
quality of such a network. This means quality assur-
ance throughout the system, in terms of the initial
instrumentation, assurance that the data are scientifi-
cally valid, that the data are collected in a reliable
and representative fashion, that the information is
assimilated so that it is of maximum use to environ-
mental quality managers, and that the information is
made quickly and' easily available to other investiga-
tors. At the same time, we must be aware that environ-
mental samples are fragile and require proper handling
until they are analyzed and that the results must stand
the test of independent analysis.
Because of the expanding global nature of pollu-
tant problems, we need techniques that can rapidly
assess environmental contamination over broad geographi-
cal areas. The aerial monitoring and remote sensing
systems now being employed in this country and abroad
provide the potential for meeting this monitoring need.
Even though these advanced techniques appear most
attractive to developed countries at the present, these
techniques will, no doubt, benefit all in the future.
Within this country, a number of significant
national programs are currently underway designed to
both improve our knowledge of environmental processes
and increase our knowledge of health effects from pol-
lutants. The National Eutrophication Survey (NES),
started over three years ago, was designed to assess
the state of our lakes and streams. The field portion
of this program, which has included the measurement of
some 800 lakes and their tributaries, is drawing to a
close this year. We expect as a result of this pro-
gram, to obtain a better understanding of the status of
our fresh water resources.
Other examples of national programs include the
National Air Surveillance Network (NASN), which is a
series of monitoring stations designed to provide envi-
ronmental data which can be used to assess the impact
on urban populations of health-related environmental
pollutants; the National Pesticide Monitoring Program
(NPMP), which is designed to measure the level and type
of pesticide contamination on a wide range of environ-
mental components; and, the Regional Air Pollution
Study (RAPS), which is a program, conducted in
St. Louis, designed to obtain data for a model of the
atmospheric distribution, diffusion and transport of
pollutants over an urban area. I might mention, too,
that St. Louis was selected for the Soviet-American
exchange of information on urban pollutants with
Leningrad being the counterpart Soviet city. This
cooperation represents only one of the many areas where
we have encouraged international approaches to the
problem of environmental pollution control.
Around the globe, one country after another has
established new governmental mechanisms for more effec-
tive environmental management. Departments and minis-
tries of environment are now commonplace. These have
been important not only in terms of furthering internal
environmental improvement in the particular countries
but in providing focal points for more effective envi-
ronmental cooperation internationally.
Over the past three years, we have seen the con-
clusion and continuing implementation of the Great
Lakes Water Quality Agreement with Canada (1972), the
U.N. Conference on the Human Environment at Stockholm
(1972), the Agreement at Moscow on a U.S.-U.S.S.R.
Comprehensive Joint Program on Environmental Cooperation
(1972), the Agreement at London on an Ocean Dumping
Convention (1972), the Agreement at Paris on The World
Heritage Trust Convention to Prevent Pollution of the
Seas by Vessels (1973), and the Bi-lateral Agreements
for Environmental Cooperation with the Federal Republic
of Germany, Poland (1974), and, most recently, arrange-
ments reached with Japan.
The U.N. Environmental Program has been established
at Nairobi. Recently, EPA was designated as this
Nation's focal point, or clearinghouse, in an inter-
national data exchange to be operated by this United
Nations program. In addition, there is continuing
activity within the environment committee of The Orga-
nization for Economic Cooperation and Development
(OECD) at Paris, the Committee on the Challenges of a
Modern Society (CCMS) of NATO, the Law of the Sea Meet-
ings, other international organizations such as WHO,
3
PI
-------
WMO, FAO and UNESCO, as well as bilateral cooperation
with Mexico and other nations.
I have touched upon only a few of the relevant
issues confronting us today in the area of environ-
mental sensing and assessment. Many of these issues
will be discussed at length in the technical sessions
during the next four days. It is obvious that we can-
not solve all these important environmental problems
within the next few days, but it will certainly be an
opportunity to continue at a more rapid pace toward
solutions. The problems we face are difficult, but
they must be faced, and progress toward solutions will
be well worth the effort.
These agreements between nations, like the efforts
within nations, can be no stronger or more successful
than the scientific evidence and understanding that
supports them. We have, in recent centuries, acquired
a truly awesome capacity to alter our environment in
ways we do not intend or fathom. We have only barely
begun to build the capacity we need to really know what
we are doing to our environment. This conference is an
important step in the right direction.
Even though we still have far to go in developing
common approaches to health and pollution problems,
and even farther to go toward developing anything
resembling a cooperative international policy that
addresses the deeper issues of the earth's ecological
balance, we are, I think, making real progress in both
directions.
Nations around the world are resolving to protect
their environmental heritage and are adopting the
necessary legal authorities to abate pollution—to con-
trol the city smog and solid wastes, the sewage and
silt-leaden streams, and the noise, congestion, and
general disfigurement and disruption of the urban and
rural landscape.
New legislation has begun to place the incentive
for cleanup on the polluter. In deference to the
principle that the polluter should pay, industries have
begun to incorporate the price of pollution control in
the cost of production and operation.
The city dweller in the nations of the world has
begun to rej ect the idea that environmental degrada-
tion, bodily discomfort and disease are the price he
must pay for increased personal opportunity. He has
begun to identify his basic needs for health and wel-
fare with an improved environment.
In response to rising public expectations at the
local level, enlightened city officials in many
nations have begun to implement many improvements to
better fulfill basic human desires. They have opened
malls to people while closing them to cars, as they
find increasingly with the American city planner
Lewis Mumford that, "Cities exist not for the constant
passage of motor cars, but for the care and culture
of man."
If one thing has become clear in recent years it
is the overriding fact that to effectively abate pol-
lution we must act together. The execution on a
smaller scale of agreements between and among coun-
tries to limit pollution loads and implement abatement
measures sets precedents for the execution of conser-
vation measures on a global scale. Formal instruments
for cooperation of the countries on the great river
systems of Western Europe, the Rhine and the Danube,
and the Colorado, Rio Grande and Great Lakes systems
in the Americas have been accomplished; and inter-
national agreements to reduce ocean pollution and sup-
port the free exchange of pertinent scientific and
technical information and methods are being implement-
ed. The success of these early pollution control
measures will affect the degree and efficacy of wider
international cooperation for resource conservation.
International agreements that are faithfully served
will strengthen the degree to which one nation identi-
fies its continued and improved well being with that
of its neighbor and will heighten official and public
sensitivity to the interdependence of our systems and
societies.
4 PI
-------
REMABKS BY DR. B. H. DIETERICH
Director
Division of Environmental Health
World Health Organization
Switzerland
Ladies and Gentlemen: I am happy to be here and
to be one of the Chairman of the symposium on Environ-
mental Monitoring. The main reason being that in my
own mind there is no doubt that the information obtain-
ed from monitoring is central to our efforts to under-
stand and control pollution and other environmental
hazards that may affect the health of people.
It is also a very special pleasure for us to join
the EPA and the University of Nevada in sponsoring this
international symposium. The history of our collabo-
ration with EPA is long and encouraging. Both of us
feel a deep responsibility for protecting the environ-
ment while being aware that the insults on our environ-
ment are caused by technological developments which—
if adequately managed and controlled—will be benefi-
cial to man. But we have learned that this management
and control is dependent on a deep consciousness, both
scientific and political and that very easily the pres-
sures of today's problems and expectations make us
forget that health hazards of tomorrow require continu-
ous alertness and a deep sense of responsibility. The
collaboration between WHO, the international agency
within the UN system responsible for health, and EPA, a
national agency concerned with environmental protection
has been of mutual benefit.
The field of monitoring, as we all know, is a very
fast developing one both in terms of data needs for an
ever increasing number of pollutants as well as the
technological developments to facilitate the accurate
measurement of these pollutants in the various media.
We must, therefore, from time to time, bring together
the various experts to 4iscuss and transmit current
information on measurement methodologies and associated
monitoring aspects.
This symposium is an excellent example of such an
event. Looking at the program one finds 17 sessions
dealing with a variety of topics which are all of great
current interest. Also a number of papers will be pre-
sented by experts coming from all parts of the world,
I feel very strongly that this is a considerable asset
for our symposium—one that should be exploited to
maximum advantage through discussion and exchange of
informatipn.
We, at WHO, look at this symposium not only as a
Means of promoting exchange of information among experts
in the field of environmental monitoring, but also as a
means of fostering the adaption, not blind adoption, of
modern techniques by countries that are still developing
their environmental control programs. In this task we
have been assisted by the United Nations Environment
Programme which has made it possible for a number of
scientists from developing nations to attend this sympo-
sium, We are extremely grateful for this generous
support which I feel is also very timely.
Today we have with us, I am pleased to say,
Dr. Pavlov to address you on behalf of WHO. Dr. Pavlov
has served as Assistant Director-General of the World
Health brganization since July 1971. He is responsible,
among other things, for the Divisions of Environmental
Health and Health Statistics, as well as the Cancer
Program of WHO.
Prior to coming to Geneva, Dr. Pavlov was the
Director of the Herzen Cancer Research Institute in
Moscow and Member of the USSR Academy of Medical
Sciences.
Dr. Pavlov graduated in Medicine in 1941 in Moscow,
worked as a general practitioner and later trained in
lung surgery. From 1949 onwards, he specialized in
radiology and oncology.
Dr. Pavlov was Chief of the Department of Clinical
Radiology at the Central Institute of Postgraduate
Training in Moscow before being appointed Director of
the Herzen Cancer Research Institute. He was responsi-
ble for organizing the radiology services of the USSR
and is the author of numerous papers and monographs on
important aspects of the radiotherapy of cancer. Dr.
Pavlov has been a full member of the USSR Academy of
Medical Sciences since 1971.
During his long and distinguished career Dr. Pavlov
has developed a deep understanding of environmental
problems and he is here with us today to speak on this
subject. It is my pleasure to present to You:
Dr. Pavlov.
5
PI
-------
OPENING STATEMENT BY DR. A. S. PAVLOV
Assistant Director General
World Health Organization
Geneva, Switzerland
Ladies and Gentlemen: It is indeed a great plea-
sure to address, on behalf of Dr. Mahler, Director-
General of the World Health Organization, this distin-
guished gathering of scientists who are meeting here
this week to discuss one of the most important aspects
of the protection of man's health against the effects
of environmental contamination—the monitoring of our
environment.
At this time, which is beset with new problems
caused by the energy crisis and the resulting economic
stresses, I feel this Symposium will make a significant
contribution towards safeguarding the quality of the
environment in which we live and work.
The World Health Organization, since its inception
in 1947, has been actively concerned with environmental
factors and their effects upon human health. This con-
cern has been largely based upon the fact that poor
sanitary conditions, and the accompanying diseases, are
among the greatest causes of morbidity and mortality in
the developing nations. Although substantial progress
has been made in many parts of the -world, the problem
of providing basic sanitary measures for a large part
of the world's population is still with us. This situ-
ation is continually aggravated by high population
growth rates, and has resulted in increasing demands
for water supply and accelerated generation of waste
and waste water. In addition, the rapid rate of urban-
ization and industrialization—not only in the eco-
nomically advanced countries, but also in parts of the
developing world—have given rise to other environ-
mental hazards to human health which often exert their
effects more subtly than the communicable diseases.
They include the physical and chemical factors which,
together with psycho-social influences, seem to affect
man's health in a multitude of ways.
WHO's concern with the environment is, in fact,
part of its constitutional obligations, as stipulated
by the International Health Conference held in New York
in 1946. It is stated that one of the functions of the
Organization is "to promote, in cooperation with other
specialized agencies where necessary, the improvement
of nutrition, housing, sanitation, recreation, economic
or working conditions and other aspects of environ-
mental hygiene" and "to promote cooperation among
scientific and professional groups which contribute to
the advancement of health."
The subject of this Symposium is one of the most
important aspects of environmental protection. The
information produced from environmental monitoring is
one of the fundamental supports on which governments
should base their policies and decisions related to the
prevention and control of environmental contamination.
As the world population grows and industry expands
to make more, and increasingly diverse, products, the
discharge of pollutants will inevitably increase. In-
creased discharges have already led to environmental
conditions that were associated with dramatic increases
in morbidity and mortality. For example, the air pol-
lution episode in London In 1952 caused an estimated
excess of 4,000 deaths.
As the Director-General recently reported to the
World Health Assembly, health-oriented environmental
control programmes differ substantially from the
traditional medical and public health programmes. The
relevant policy formulation is a new challenge because
it involves not only sciences and technology, legisla-
tion and administrative enforcement, but also economic
policy as related to energy production and its use,
industrial structure, land use, national and interna-
tional trade, consumption and conservation of natural
resources, and finally the allocation of costs and
fiscal problems of taxes and charges.
Environmental pollution has adverse effects on
health and is of primary importance to health authori-
ties. In view of its health implications, health
authorities should always be actively engaged in its
control, as already Exists in some countries, thereby
requiring the development of new relationships between
health authorities and environmental pollution control
authorities, both at national and local levels.
The formulation of sound environmental control
policies depends on the availability of adequate infor-
mation on which to base decisions. WHO has given con-
siderable attention to this subject through the
development of its environmental health criteria pro-
gramme. The main objective of this programme, which is
financially supported by the United Nations Environment
Programme, is to assemble and assess existing informa-
tion on the relationships between exposure to environ-
mental pollutants and the effects on man's health. The
work is carried out by panels of eminent scientists from
all over the world. Their task is to prepare critical
reviews of the work carried out by WHO Member States on
the research into the health effects of specific
pollutants.
Some half a million chemicals are currently used,
and about ten thousand are produced annually in amounts
of between five hundred and one million kilograms. In
addition, a variety of physical hazards have to be con-
sidered. A practicable programme for the preparation
of environmental health criteria should be based on
clearly-defined priorities. A list of priorities has
been established by a scientific group taking-into con-
sideration, among other things, toxicity, persistence,
and abundance of the various substances.
These documents should contain accurate information
about what is known and where more information is
needed. It is also highly desirable for them to include
risk information, expressed in terms of the probability
of adverse effects. Such risk data will be of help to
policy makers when taking decisions based on informa-
tion available.
The expertise needed for the evaluation of risk is
different from that needed for benefit evaluation. On
the risk side, concern is focused on adverse health
effects on man, damage to the environment, and misuse of
resources. On the benefit side, emphasis is laid on
value to the consumer and the country; that is, improve-
ment in the health of the population, and the avail-
ability of cheaper and better products, particularly
those that man considers indispenable.
In practice, however, it is usually difficult to
obtain quantitative information on exposure and re-
sponse to man, particularly for chronic effects caused
by low-level exposure. Qualitative information can be
obtained more readily. Dose-response relationships can
6
PI
-------
be obtained from animal experiments with reasonable
accuracy, especially in tests at fairly high levels of
exposure, but it is not an easy task to obtain
exposure-response information with satisfactory accu-
racy at low levels of exposure. Retrospective studies
in man have shown information on exposure is usually
seriously limited and the time periods and levels of
exposure pose difficult problems. It should be
stressed that for epidemiological studies to be mean-
ingful, there should be full monitoring of the envi-
ronmental media to assess exposure and measurement of
the effects on the population.
The World Health Organization is very interested
in promoting the development of adequate environmental
monitoring programmes, including monitoring activities
which are directed towards the measurement of human
exposure and the ensuing effects on human health. Our
efforts in this regard have been to provide technical
guidance and assistance to Member States in the estab-
lishment and strengthening of monitoring programmes
which are needed for public health purposes at the
national level. We are also interested in furthering
the collection and synthesis of information on an
international and global scale, where such information
is required. The full development of the Global Envi-
ronmental Monitoring System of UNEP should provide the
basic framework for such data compilations and data
synthesis.
From, an international perspective, and also from
a national point of view, the development and use of
comparable .measurement methodology is extremely impor-
tant. Although different governments may not employ
quite the same monitoring methods, all may benefit
from using methods of sampling and analysis which can
be related to an internationally accepted standard and
from expressing the results uniformally.
I have no doubt that this week's deliberations
will make a most valuable contribution to some of the
points which I have raised, and thus provide a better
scientific basis for those who have to make decisions
on measures to be taken to prevent or control environ-
mental pollution.
May I conclude by wishing you success in your
work.
7
PI
-------
ADDRESS BY DR. EDIJARDO ECHEVERRIA ALVAREZ
Presidente del Consejo Tecnico de la
Subsecretaria de Mejoramiento del Ambiente
Mexico City, Mexico
Mr. Chairperson, Members of the Presidium, Ladies
and Gentlemen: Progress creates noble occupations for
mankind. Although until now it has been directed
toward the discovery, creative integration and exploi-
tation of heretofore unknown areas and dimensions, in
our times there is also a need, not to be postponed,
to preserve what has been accomplished. We have ob-
tained this patrimony through centuries of struggle
and efforts, and we are coramited to leave it to future
generations enriched with our daily contribution, and
not impoverished by misuse and the growing waste char-
acteristic of our shortsighted activities.
The constant and grave deterioration of the envi-
ronment in which man moves and develops has compelled
him to reconsider the bases on which true progress is
founded. In our large urban centers we can see in all
its fearful magnitude the spectacle of what our daily
surroundings may become if we do not apply suitable
and timely corrective measures. The trends observed
in population movements indicate that in the near
future the metropolitan complex will prevail over all
other forms of human conglomeration.
The growth of cities is accompanied by deteriora-
tion of numerous urban features and by the degradation
of the living conditions of a large portion of the
population.
We may ask ourselves if those who conceived and
slowly shaped our cities ever foresaw such inordinate
growth and its allienating consequences. By their own
inertia, instruments and institutions have obscured
man's vision of the future. Indeed, it seems as if
technological progress leads to a spiral of uncontrol-
led transformations that are increasingly difficult to
put into context, to foresee and prepare for.
Because of the nature and scope of his works, man
is the principal transformer of the environment. Per-
fection of the techniques that made it possible to
utilize natural resources and to benefit from them, is
doubtless one of the greatest achievements of culture.
Nevertheless, at the same time it constitutes the
essence of the conflict of human beings in their rela-
tion to nature. If development cannot be stopped or
held back, the problem then takes on special contours.
In one aspect, the necessity of solving the problems
of a constantly multiplying population, forces us to
have recourse to industrial development as the most
effective means of satisfying the growing demand for
food, housing, goods and, generally speaking, all
kinds of services. At the other extreme, the rationale
of industrial development as conceived at certain times
and places has now been frequently questioned.
In its deepest sense, development of a country
should be understood to mean progressive transformation
of the environment. Urban growth, the exploitation of
natural resources, the improvement of farming tech-
niques, industrialization, all are but alterations of
the natural order designed to ensure the existence of
man. This is how we should consider the two aspects
that serve simultaneously as the point of departure
and the objective of this process—that which implies
destruction of former harmony, and that which aims at
humanization of nature.
Preference is given to increasing rates of econom-
ic development, but in exchange a high price is paid in
social welfare. "The ultimate goal of science is the
survival of man," says Professor Marshall Walker, "Some
occurrences make it possible to make very accurate pre-
dictions."
This is the status of the problem on the environ-
ment .
In this conference we have listened with great
attention to the very learned opinions expressed by the
three previous speakers. There is no need to reiterate
in detail what is widely known—the ever increasing
poisoning and deterioration of soil, water, air and the
seas. Suffice it to say that the process of mass sui-
cide we are witnessing was well understood by U. Thant
when he pointed out the short time remaining for us to
initiate action in an attempt to save threatened
mankind.
Pollution of the Earth is not an isolated danger.,
on the geopolitical map. Rather, it is a conflict that
concerns all nations—developed and underdeveloped—one
which affects children, young people, adults and old
people alike.
What is more, social phenomena are also included
within the scope of ecology. Thus, noise, speed, over-
crowding and deplenishment of nonrenewable resources
are all noxious elements of a worldwide effect whose
solution cannot be deferred. A primary cause of this
crisis is the spontaneous transformation of an agricul-
tural society in harmony with nature into economic
structures of an industrial type. To put it briefly,
man has created a world of things, of machines, work-
shops, factories, automobiles, airplanes, and so forth,
which in their daily functioning produce chemical pol-
lutants of all sorts. The word "smog" could symbolise
all of them. As a matter of fact, smog is perhaps the
phenomenon that has galvanized mankind to corrective
action.
From the beginnings of civilization man has been
altering environmental processes in his use of natural
resources for the extraction of energy.
This energy is found in the ensemble of natural
conditions within our earth—rocks, minerals, soil,
air, water, vegetation and wildlife—the environment
itself.
Nevertheless, natural resources did not constitute
resources until man was both present and able to make
use of them.
Skill in identifying, obtaining and using natural
resources has been a continuous human mental process.
We must also have a clear understanding of the differ-
ent degrees of exploitation, brough about by changes in
human conditions. The population, the demand for food,
industrialization, pollution and consequently th^ use
and exploitation of natural resources are increasing.
Annual increases are plotted as exponential growth.
Almost all of mankind's common activities, from the use
of fertilizers to the expansion of cities, may be
represented by exponential curves.
8
PI
-------
Environmental problems crop up In all societies
to some degree, and they are made up of interacting
technical, social, and political elements. But pre-
cisely herein lies the grave problem. Despite his
considerable knowledge and' tools, man does not under-
stand the origins, the future, the significance and
the interrelations of the numerous components of this
problem and has therefore been unable so far, to pro-
pose effective answers of political, social or econom-
ic nature. This incapacity might be largely ascribed
to the fact that we keep studying isolated aspects of
the situation without understanding that the whole is
greater than the parts, that change in one element
signifies change in another, and that benefit to one
part may be detrimental to another.
The application of rational environmental tech-
nology, orienting human skills to be management of the
natural resources of a given area can multiply that
area's capacity for sustaining more population and at
the same time raise the level and quality of man's
life.
What, then, should be the goal of a policy for
the management of natural resources? Should it be
optimization of the economic product? Or perhaps of
the social or political product? Should it be the
search for mental, social and political structures
which will make it possible for us to not alter the
environment? Or should it be to maintain a balance
between the exploitation of nonrenewable and renewable
resources"so as to make them adequately available to
future generations?
An environment preservation policy should be
founded on the quality of life—not only that of to-
day, but that of today and tomorrow; it should not be
aimed at an increase in population; it must recognize
that the nonliving environment can only interact with
living creatures within certain limits and therefore
the aim should be focused to attain true improvement
of the quality of life.
It is not a question here of preservation but
rather one of conservation, inasmuch as preservation
is the absence of exploitation. Similarly, if con-
servation is practiced very rigidly, a society may
reduce its own standard of living by not using that
which can be used to increase its own well-being.
Nevertheless, the contrary has usually been the case,
that is, over-exploitation of resources, either
through lack of care or—and this is even more tragic—
because the population of an area is so dense that it
can only survive by using those resources that right-
fully belong to the population of tomorrow.
There is certainly no room for complacency in a
world in which there is so much hunger and poverty, in
which there are so many areas whose rate of economic
growth does not keep pace with population increases.
People in these areas can at best maintain a level of
meager survival. This is true ecological imbalance,
which in economic terms is called underdevelopment.
Obviously, and this we know through experience in
Mexico, the complexity of ecological phenomena is
absolutely alarming. The President of my country,
Luis Echeverria, has been a witness to the blood cur-
dling images of poverty and hunger in his recent con-
tact with the Third World.
\
Umberto Melotti, a sociologist of hunger, has
coined this bit of truth: "Hunger kills more than
war." From all this Mexico considers it to be a good
strategy to maintain an unflagging determination to
diminish the social injustice that still tolerates
sharp inequalities between very rich and very poor
nations.
The philosophy of the Charter of Economic Rights
and Duties of States, which Mexico proposed at
UNCTAD III and which was approved last December at the
United Nations by more than 100 countries, is founded
on a most resolute return to the principles of human-
ism. Two different connotations may be discerned in
this humanistic concept. For scholars it signifies
harking back to the study of classical languages and
Greek and Roman civilization. But for human suffering,
humanism is awareness of the solidarity among men as
well as among nations, a solidarity based on mutual re-
spect and on unyielding consideration for human digni-
ty. This philosophy embodied in the Charter, stems
from the impeccable ethic of one of Benito Juarez best
known maxims:. "Respect for the rights of others is
peace."
Thoreau, the solitary philosopher, whose utopic
view was that environmental deterioration could be
halted by systematic destruction of man-made things and
our return, like prodigal sons, to the irreplaceable
joy of nature. But Thoreau would not have found the
means to do so since the course of history cannot be
stopped. Through subsequent evolution and revolution
primitive man uncovered the use of those scientific
and technical means that we possess to fight the battle
for survival.
It is obvious that peaceful cleanup operations
which will serve to create new humanistic values demand
the expenditure of resources and the utilization of
highly-trained specialists. This is a responsibility
to be shared by all human beings; however, one which is
most particularly the responsibility of those countries
wealthy both in money and technology. The assistance
provided by the highly developed nations to impover-
ished peoples is essentially the spirit of behaviour
compatible with humanistic morality.
At the beginning of this century it was still com-
mon to juxtapose two apparently irreconcilable worlds,
the East and the West. This concept has been super-
seded. The East must preserve its spiritual values,
transformed, however, by the creative impetus of con-
stant action. We, on the other hand, cannot forsake
our vital dynamism, on the condition that we encourage
the realm of the spirit. Indira Gandhi, in her speech
at the Plenary Session of the United Nations Conference
on the Human Environment in Stockholm in 1972, said:
"One cannot be truly human and civilized unless one
looks upon, not only all fellow-men but all creation,
with the eyes of a friend."
Ladies and Gentlemen: We may say here, that each
of the countries of the world, by adopting appropriate
measures, is attaining practical, positive results,
particularly with regard to-monitoring of the environ-
ment. We must quantify and qualify the environment of
today and what it will be in the future. We are al-
ready aware of the mechanisms producing and spreading
pollution. We have reached the point of putting tech-
nical accomplishments to the service of mankind.
Notwithstanding our positive achievements to date,
I am still concerned by a disturbing question: When
will the curative measures we have employed give rise
to preventive measures?
In any event, we have been fortunate to live in an
era of intense creative potentiality.
9
PI
-------
We have preferred to employ a mass education sys-
tem rather than implement a program of coercion and
compulsion for those who disrupt environmental balance.
We want man to recover his capacity for reasoning, to
assume complete responsibility, and to live the kind
of life that will contribute to both his own happiness
and that of his fellow-men.
This meeting has been entrusted with taking a
new and firm step in a profoundly humanistic and vital-
ly important task. I express my sincerest hopes that
the work to be carried out here will result in clear-
cut, specific action, and that it will be translated
into an open and effective desire for cooperation.
The creative genius of man has been manifested in
his scientific and artistic works. Sustained effort,
true diligence and a desire for innovation are irre-
placeable premises on the road to discovery.
Let us permit the freedom of the spirit and free
imagination, imbued with deep and genuine human feel-
ing, to lead us on to new possibilities. This same
freedom and imagination must conjure away negative
tendencies and from now on, lay the foundations for
a dignified promising future for mankind.
In conclusion, I would like to quote these poetic
lines of the great Walt Whitman:
"Victory, union, faith, identity, time,
The indissoluble compacts, riches, mystery,
Eternal progress, the cosmos, and the
modern reports,
This then is life . . ."
Thank you.
10
PT
-------
INTERDEPENDENCE EQUALS ENVIRONMENTAL IMPROVEMENT
Fitzhugh Green
Associate Administrator
U.S. Environmental Protection Agency
Washington, DC
After World War II the U.S. began a long period of
technical assistance not only to revitalize the war-
ravaged countries of Europe but also to strengthen
fledgling nations emerging from dismantled empires. We
were still operating this large but disminishing pro-
gram when mankind officially entered the age of envi-
ronmental concern at Stockholm in 1972. So you might
logically have expected the U.S. would continue to
extend its technicians and technology to help stop
this new danger. Well, we did, but much more modestly
and deferentially than we had been doing with other
types of foreign need. Why?
To begin with, this is a domestic need, and chari-
ty begins at home!
Secondly, the U.S. has not yet found all the tech-
nological or scientific answers to the ecological ques-
tions being raised at home and abroad. Although they
are not specifically addressed in this great Conference,
I mean such items as the response of the atmosphere, or
bodies of water to pollutants being added, one at a
time, or in multiple numbers. We don't always know
what happens even when we are able to reduce loads of
pollutants into the air or water.
For instance, now that we have begun to decrease
the dumping of phosphorus into the Great Lakes, our
scientists are not sure what the results will be.
They can't tell us the fate of that foot-thick mat of
phosphorus-laden material reported to be lying along
a large portion of the floor of Lake Erie. Will it be
absorbed as the Lake flushes itself out toward the
St. Lawrence Seaway every 12 years, or will it remain
stable, or even increase due to chemical actions we
cannot predict?
We must also admit to vast areas of ignorance in
the health and ecology effects of toxics. How do you
determine dose response to one or more substances for
an urban population when virtually every human being
has different tolerance levels; and moving about the
city each receives different combinations and amounts
of pollutants, singly or in multiples? We sought light
on these issues from other nations in two conclaves we
held with Common Market experts: In Amsterdam in 1973,
and in Paris in 197A.
The third reason we are not offering broad tech-
nical assistance is that so many other countries, re-
presented richly right here in this room, have
knowledge, techniques and instruments as good, or bet-
ter than our own. Indeed, we are frequent beneficiar-
ies of other nations. This Conference vividly
illustrates that fact. I'm sure we Americans here will
be wiser at the end of this week thanks to what our
visitors can give us.
Interdependence, which Secretary Kissinger calls
the keystone of international relations today, perme-
ates ^environmental relations as strongly as it does
the political and economic realities of food, mineral
and energy resources.
Because interdependence dominates our policy, our
Government is busy collaborating with individual
nations, and groups of nations like the UN and its
specialized agencies, the Organization for Economic Co-
operation and Development (OECD), the European Common
Market, the Conference on Security and Cooperation in
Europe (CSCE), NATO, and more.
EPA is the chief U.S. action agency in these ef-
forts. We participated in scores of symposia, seminars
and conferences at which this meeting, after Stockholm,
may well be the most important so far. We benefit from
our formal bilateral agreements with the Soviet Union,
Japan, Germany and Canada, plus informal arrangements
with dozens of other governments.
You might be surprised to know what our EPA
personnel think of all this activity—every year it
totals many man trips to foreign shores and briefing up
to 2,000 foreigners at our Washington and Regional
Headquarters and Laboratories. One of our bilaterals
alone involves 150 EPA men and women. Since EPA is a
domestic agency you might think our people would com-
plain at spending so much time and energy with for-
eigners. But this isn't so. In fact, we have had to
impose vigorous traffic rules to keep enough of our
personnel at home to do their job here!
In order to learn what's going on overseas without
excess travel, we have initiated an information and
document exchange with over 50 countries and multi-
national groups. By this method we discover month by
month what other countries are doing to save the bio-
sphere. I can tell you they are binding themselves
into a steadily tighter strait-jacket of law and regu-
lation. In short, they are making a commitment to
environmental sanity that is becoming harder and harder
to ignore. Eventually we hope to extend this microfiche
network to all UN members.
Meanwhile we have worked with UNEP to establish a
vital part of EARTHWATCH, namely—the Information
Referral Service (IRS). IRS will offer a handy clear-
inghouse to sources of environmental information
wherever they exist. We plan to inaugurate the Ameri-
can National Focal Point of IRS—operated by EPA—next
month.
Will you allow me to make a special plea? Eco-
logical interdependence means we must in all countries
train more non-governmental scientists and technicians,
and to press into service more private commercial and
academic institutes like those who share sponsorship of
this meeting.
Why is this necessary? Because there aren't
enough people or organizations—in or out of govern-
ment—able to do the research and build the control
hardware. This process is beginning and we believers
in free enterprise are pleased to note already a tiny
but accelerating international trade in measuring
instruments and technology transfer.
Now that we've discovered interdependence you must
forgive us if we're too enthusiastic about its merits.
I realize we may be like that famous lady convert to
Catholicism who explained her naw religion to the Pope
with so much vigor that he finally had to tell her "but
you see madam, I am already a Catholic!"
11
PI
-------
And with so many converts to clean water, air and
earth assembled here this morning, we are really de-
lighted. So much practical progress has happened so
fast—it's been almost a flash-fire response to all
that tear-dripping rhetoric at the start of this
decade.
In the short life-span of our own EPA—less than
five years—this planet has seen its nation-states go
from no EPAs anywhere to a total of over 50 and still
growing. In the years ahead we Americans will doubt-
less continue to exchange ideas and know-how with our
opposite numbers everywhere on the globe.
But so far man has not yet risen above the bucket-
brigade stage of pollution protection. Too often we
still must put out fires with relatively primitive and
flawed devices, like the catalytic converter, because
they are all we have to do the job.
The ultimate solution will be most likely to re-
design fuel-burning engines, power plants and manu-
facturing processes so they will operate with minimal
waste byproducts, or at least with recyclable ones.
Then high environmental quality standards will be
attainable by all nations. And violators will be
easily detected by monitoring expertise you are per-
fecting here in Las Vegas and in your own localities.
As nations collaborate ever more interdependently,
new industrialization may be preceded with global envi-
ronmental impact statements to avoid eco-damage in
advance, instead of having to correct it as we older
industrialized societies must do now at such high cost.
When man has accomplished such steps, and we have
to be encouraged by forward motion already made, our
life support system will be secure, and protected by
EPAs in all countries as routinely effective as fire
departments are today. If not, man almost surely will
be consumed by the conflagration of filth that has
started already to threaten his survival.
So I wish you well in your labors this week. I
for one recognize you as elite fighters for the health
and lives of this planet's citizens—both today's and
their descendants.
12
PI
-------
LUNCHEON ADDRESS BY CHRISTIAN A. HERTER, JR.
Deputy Assistant Secretary of State
for Environmental and Population Affairs
State Department
Washington, DC
I am particularly pleased and flattered to be
included on your program at this very excellent Con-
ference on Environmental Sensing and Assessment, spon-
ored by the International Symposium on Environmental
Monitoring, the Third Joint Conference on Sensing of
Environmental Pollutants, and their constitutent orga-
nizations. As I have looked through your program, it
seems to me there is virtually no subject in the area
of environmental sensing and assessment that you have
not covered in your daily sessions, conducted by ex-
perts in the field. I am afraid there is very little
that I can add from a substantive point of view.
Under these circumstances, you might well ask why
I am participating in this Conference or why a member
of the Department of State should in any case have a
contribution to make. It is a good question, but per-
haps there is a persuasive answer,
I hope you will permit me a few minutes to de-
scribe where the Department of State fits into your
deliberations and why it is particularly interested
in the areas with which you are concerned. I am
Deputy Assistant Secretary of State in the Bureau of
Oceans and "International Environmental and Scientific
Affairs, and my particular responsibility, in the
international context, is environment, population, the
scientific aspects of food production, particularly in
the developing countries, climate, weather modifica-
tion, and renewable natural resources.
From professional point of view, it seems to me
I have no alternative but to be here! Beyond, that,
however, there are more serious reasons.
In the first place, this is an international ^on-
ference and amongst you* numbers are a considerable
number of participants from foreign countries who have
an invaluable contribution to make. Even though many
of you here work for domestic U.S. agencies or with a
variety of domestic private concerns, what you are
doing not only applies to the domestic scene, but has
considerable importance in our international work. It
enables us to make valuable contributions to inter-
national efforts, and, in many instances, provides
information for other countries to use in their
related programs.
Secondly, although in excess, perhaps, of ninety
per cent of the environmental problems, particularly
of industrialized countries, have to be dealt with on
a national or local basis, there are obviously many
urgent concerns—those affecting the oceans, the atmo-
sphere, terrestrial ecosystems, monitoring, and the
control of toxic substances, for example—that can
only be dealt with on a global or international basis.
And to deal with them effectively requires interna-
tional cooperation—with international organizations
and foreign countries; it requires preparations for
and the negotiation of multilateral conventions deal-
ing <
-------
including representatives of 113 nations, the questions
were constantly asked: what do we really know about
man's impact on the global environment? How can we
find out the scientific information necessary to tell
use what is happening? And only if we can find out
over a period of time, what then has to be done and in
what order of priority? This isn't to say there were
not then, nor are there now, immediate environmental
problems of an urgent nature which require immediate
attention. You are considering many of them in your
sessions here. But we still do not know the big pic-
ture, even today. We can only speculate. This is why
your conference on environmental sensing and assessment
and the work that you are doing at home in this field
is so relevant.
One of the most significant achievements of the
Stockholm Conference was an international agreement on
the need to establish a worldwide capability for con-
tinuous assessment of the status and trends of the
earth's environmental quality as a prerequisite to ef-
fective environmental management. In UNEP's program,
this total concept has been given the name EARTHWATCH,
a four-pronged initiative involving monitoring, re-
search, evaluation and information exchange.
UNEP has been very slow in getting started on
this effort; now, however, there are the beginnings of
progress in monitoring, an information exchange system,
and some aspects of research. The.monitoring program
is called GEMS (Global Environmental Monitoring System)
and seven program goals exist for GEMS. These are:
- An expanded human health warning system;
- An assessment of global atmospheric pollution
and its impact on climate;
- An assessment of the extent and distribution
of contaminants in biological systems, par-
ticularly food chains;
- An improved international disaster warning
system;
- An assessment of the state of ocean pollution
and its impact on marine ecosystems;
- An assessment of the response of terrestrial
ecosystems to environmental pressures; and
- An assessment of critical problems arising
from agricultural and land use practices.
The GEMS program will be built upon a variety of
national and international monitoring activities al-
ready in existence, the latter including the Integrated
Global Ocean Station System (IGOSS), the World Weather
Watch, the Global Atmospheric Research Program, and the
monitoring network in the process of being established
around the Mediterranean. Of course a number of highly
sophisticated national systems are already in existence,
such as those in the U.S., the Soviet Union, and parts
of Europe, and they will form a major part of the basic
foundation for the GEMS program.
The GEMS secretariat in UNEP has recently held
discussions with a number of United Nations Specialized
Agencies which have a capability in monitoring—e.g.
WMO, WHO, FAO, etc.—, and it is contemplated that over
the next year or two there will be a series of meetings
of highly expert country representatives to design
pieces of the system in the seven program areas that I
have already mentioned.
The first such meeting, on ocean monitoring, will
take place early next year, to be followed, we hope
shortly thereafter, by a meeting on global atmospheric
pollution and its impact on climate.
Here in the U.S. Government, an interagency task
force has prepared a series of documents on the design
philosophy for a global environmental monitoring sys-
tem, and these documents will be reviewed in detail in
a three-day symposium this October in Washington, spon-
sored by the National Academy of Sciences and the De-
partment of State. The symposium will include about
thirty experts from the private scientific community
and thirty experts from Government. After necessary
modification, the basic documents will then be used in
developing U.S. participation in GEMS.
No doubt it will take two or three years for the
international community to complete the basic design
for these various monitoring systems and to make appro-
priate recommendations as to how they should be imple-
mented. Much of th^ detail of U.S. thinking in this
area was set forth by Dr. Clayton Jensen, formerly of
NOAA, during Session III on Monday afternoon.
In the view of many countries, including the U.S.,
EARTHWATCH is a keystone of UNEP's international ef-
fort, and we hope the Congress will pass a concurrent
resolution indicating its continuing support of the
program.
The next significant portion of EARTHWATCH is
information exchange, and this has produced a new pro-
gram, which is now well underway, called the Information
Referral System or IRS, designed to facilitate world-
wide access to information on the cause, effects, and
control of environmental problems. XRS, which is
headquartered in Nairobi, will not of itself be a
source of information on environmental activities;
rather it is a directory on a worldwide computerized
basis of national inventories of environmental informa-
tion. Its purpose is to provide a link or switchboard
between those countries or governments needing informa-
tion to solve environmental problems with prospective
sources of that information. UNEP has asked each mem-
ber country to designate a lead agency as the IRS
national focal point for sources of information in that
particular country. All organizations, governmental or
private, which can provide environmental information,
will be invited to register as sources. In the case of
the U.S., the Environmental Protection Agency has been
designated as the U.S. focal point and in early October
it will have a formal opening of its facilities to pro-
vide this service. In sum, if X country wants to know
more about sulphur oxide controls or advanced municipal
waste disposal systems, for example, through IRS it
will be referred to whatever sources of information in
various countries exist on these subjects and it will
then establish its own direct communication with these
sources.
Similar in concept, but more specialized, is the
International Registry of Potentially Toxic Chemicals
(IRPTC), now being established by UNEP and presumably
to be located in Europe. As the name suggests, the
IRPTC will emphasize the assembling of information—
hazard assessment reports, criteria documents, etc.—
on selected chemicals for use in exchanges of regula-
tory interest. A part of this effort would include
gathering data on actual regulation of chemical sub-
stances along with the justification for such regula-
tions. The information will be available to all UN
members, although a computer capability probably won't
exist for two or three years.
14
Pt
-------
Continuing research is of course another major
element of the EARTHWATCH program, and UNEP itself has
done a small amount of contracting for specific forms
of research. As an example, broadly under the umbrella
of UNEP and EARTHWATCH, is the so-called Man and the
Biosphere Program (MAB) located in Paris and operating
under the aegis of UNESCO. Basically MAB operates
through a network of national committees—in the United
States, the MAB Committee of the National Commission
for UNESCO—and these national committees have links
with both governmental and non-governmental institu-
tions. Two or more countries cooperate internation-
ally in areas of environmental research and training
where there is a common interest. The results of the
research will be made available to everyone. There is
an International Coordinating Council for this total
effort headquartered in Paris, and some fourteen areas
in which cooperative research is possible have already
been designated. For example, MAB has committees on
tropical and sub-tropical forest ecosystems, grazing
lands, inland water systems, coastal zones, deltas
and estuaries, island ecosystems, and a more recent
committee on "biosphere reserves." Vie have high hopes
that productive research of use to us and other coun-
tries will emerge from the international cooperative
program that is now taking place. It is, of course,
additional to the extensive environmental research that
is now taking place at national levels.
The final function of EARTHWATCH is to provide a
periodic evaluation and review of global environmental
conditions and trends as the basis for improved envi-
ronmental management and decision making.
This is, of course, the "bottom line" of the
entire EARTHWATCH program. How does one evaluate the
sum total of the information obtained from the moni-
toring systems and national and international research,
and how does one bridge the gap between what the system
tells us about urgent or longer-term threats and what
action must be taken by decision-makers at the national
level? International institutions, at least thus far,
cannot make decisions binding on governments; they can
only recommend. In the environmental field, there is
no international enforcement machinery.
Partly, this is a question of credibility. Is the
Information on which governments are expected to act of
such unimpeachable authority—the product of such
really first rate international scientific expertise—
and the evaluation of the information so credible that
decision-makers have really no choice but to act? And
this may not be a question of a clear and present dan-
ger, as for example, PCBs, DDT, mercury, or, as the
•Japanese are now telling us, hexa valent chromium. It
may involve a serious, professional judgment as to
changes that are taking place in the earth's albedo,
in the upper atmosphere, in the global chemical inter-
change between carbon dioxide and oxygen, the implica-
tions for human health, food and forest production, of
increased energy production and man's impact on his
gene pool as well as the world's climate. In sum, are
the totality of man's activities in one way or another
degrading or endangering his life support systurns, and
if so, to what extent?
The answers to these questions must be found out
and the answers must be believable to political leaders
and their constituencies. EARTHWATCH, if successful,
can provide basic information—from monitoring and
research—but it cannot provide 1 believable judgment
and evaluation to the international community.
Who can?
Ideally, UNEP could become the great intellectual
force in environmental matters, utilizing and taking
advantage of the best scientific brains in the inter-
national community. It could become the world's re-
pository of environmental expertise and gradually
achieve a very high order of credibility. This is, of
course, a monumental undertaking, far beyond UNEP's
present capabilities. Yet is must face this challenge
if it is to become a real force in global environmental
affairs.
Or one can conceive of some sort of international
environmental entity, essentially private but funded by
governments, able to reach out to the vast expertise
that exists and to itself develop the kind of world
overview of the "state of the environment" that mankind
now needs but will need even more urgently as the years
go by.
I suspect that parts of the scientific community
would not be enthusiastic about such an entity on the
grounds that once a problem is identified, the scien-
tific response is rapid, further research is immediate-
ly undertaken, and, as in the case of potential threats
to the ozone layer, the government itself comes forth
with substantial financial support for research. In
the view of some scientists, adequate work is being done
on threats that are known and systematic efforts to
seek out threats that are not known are apt to be un-
productive, In their opinion, what is needed is a bet-
ter system of communications between scientists, not an
international institutional structure.
I'm not professionally capable of making an intel-
ligent evaluation of these differing views. I do,
however, feel that there is already a vast amount of
information and research in the environmental field, to
which you yourselves are making an invaluable contribu-
tion, but which is not generally known; that a great
deal more has to be learned, particularly on a "macro"
basis; that scientific and governmental responses are
'ad hoc," usually after a threat has surfaced; and
that we need a framework of knowledge on the global
situation against which new and old threats can be
judged. Above all, we need an international voice, not
just a U.S. voice, that is so authoritative and so
persuasive that decision-makers in all countries cannot
fail to respond.
15
PI
-------
INTRODUCTION
Lars Friberg, M.D.
Professor and Chairman
Department of Environmental Hygiene of
The Karolinska Institute and of
The National Environment Protection Board
S-104 01 Stockholm 60
Sweden
I am glad to welcome you to this session on
metals and I hope that you will find the pro-
gram to be informative. Today, metal toxicol-
ogy holds a great global interest. With toxi-
cology I mean here not only direct toxic ef-
fects upon humans but also adverse effects
upon the entire ecological system. Granted,
public attention toward a substance or a group
of substances may not always adequately re-
flect the real risks; it nonetheless seems
fair to say that ample evidence shows that
certain metals and metal-like substances may
be extremely toxic, not only after exposure
to high concentrations but also after exposure
to minute amounts if the exposure is repeated
over long time periods. One reason is that
some metals tend to accumulate in the animal
and human organism. Methyl mercury is a good
example, showing a biological half-time in
humans of about 70 days on the average. With
biological half-time is meant the time the
organism requires to excrete half the accumu-
lated amount. In fish and seals the biological
half-time is still longer, reaching values of
between 500-1,000 days. A still more accumu-
lative metal is cadmium. In humans the bio-
logical half-time has been estimated to be
20-40 years, steady state may thus not be
reached during the life-span of human beings,
and should exposure be continuous, accumula-
tion will take place until death.
Of equal importance is the tendency for sever-
al metals to accumulate in soil and water,
accentuated by an often pronounced tendency
to be taken up by basic foodstuffs such as
grain and rice. The property of metals neither
to disintegrate nor to metabolize into non-
toxic substances can mean that a polluted en-
vironment remains so for a very long time. We
still find high concentrations of mercury in
the bottom sediment of waters where the pollu-
tion with mercury occurred several years ago.
For the time being, a suitable way of ridding
the soil of cadmium has yet to be found.
Until about two decades ago, toxic metals were
considered a problem almost exclusively for
exposed workers. Reports of severe toxic ef-
fects within industries are numerous, but un-
fortunately very few useful epidemiological
data, i.e. reliable data on exposure as well
as response, were reported. To date, the situ-
ation has not changed dramatically.
As far as the general environment is concerned
it would seem that at least cadmium, mercury
and lead have long been so well known for
their toxic effects that more care should have
been taken to limit their discharge into air,
water and soil. This has not been the case.
World-wide interest in metal toxicology has
gained momentum, though slowly at first cor
example, considerable time has elapsed since the mass
outbreaks of methyl mercury poisonina and cadmium
poisoning resulting from contaminated fish and rice'
in Japan. It is from such incidences that most
data on relations between exposure and re-
sponse have been collected. One important dis-
covery in the wake of the Japanese and other
mercury problems was that of the dangerous
methylation whioh mercury undergoes in the
environment. In spite of the increased knowl-
edge and scientific efforts, toxic effects
continue to emerge. The recent mass outbreak
of methyl mercury poisoning due to contami-
nated grain in Iraq is an example of this.
Mercury and cadmium are not the only sub-
stances which have given rise to well recog-
nized epidemics. In the 1950's about 10,000
Japanese children were poisoned due to expo-
sure to dry milk contaminated with arsenic.
Lead in air and in pica remains a problem in
several areas of the world. More recently,
attention has been drawn to cancerogenic
properties of metals, particularly some forms
of arsenic, nickel and chromium.
It may be mentioned that within the WHO pro-
gram to put forth environmental health cri-
teria for different substances, a high prior-
ity has been given to metals. Final evalua-
tions have been made concerning cadmium,
lead and mercury. Of the other international
organizations involved in metal toxicology,
I wish particularly to refer to the Subcom-
mittee on the Toxicology of Metals under the
Permanent Commission and International Associ-
ation on Occupational Health. For some years,
this Subcommittee has been active in organiz-
ing workshops where general principles con-
cerning metal accumulation and toxicity have
been put forth. Thus far, these meetings have
culminated in three published reports.
The program for this session consists of two
parts. The first part, primarily during this
afternoon, is devoted to broad reviews. The
presentations will focus on important aspects
of metal toxicology, including the needs for
future research as well as some pitfalls in
connection with carrying out studies on met-
als . The rest of the program is devoted to
presentations of original studies concerning
environmental aspects of metals. Apart from
giving new data, these presentations will ex-
emplify metal studies under way. It is with
great pleasure that I invite the speakers of
this session to start presenting their
papers.
1-0
-------
FLOW OF TOXIC METALS IN THE ENVIRONMENT
Dale W. Jenkins
Pan American Health Organization
Mexico City, Mexico
Environmental flow of 12 toxic metals is pre-
sented with quantitation of sources, flow in the
atmosphere, terrestrial ecosystems, freshwater and
the ocean. Emissions from man-induced and natural
sources are compared. New biological accumulation
and concentration data are presented based on a
world-wide literature survey. Enrichment and
concentration factors, and development of biological
monitoring systems are discussed. Historical trend
data on the metals is reviewed.
Introduction
ecosystems with man and his technology as the
dominant component. The biotic flux in relation to
these metals will be discussed below.
The flow diagram shows that each of the major
environmental components have metals flowing in and
out except the ocean. The ocean receives large
amounts of the toxic metals but only a small amount
leaves it in the form of marine products, particu-
larly marine life.
Figure 1
Of the 105 known elements, 77 are metals of
which 52 are of economic importance, and about one-
fourth of these are toxic to biological organisms.
Twelve toxic metals have been selected on the basis
of biological toxicity, relative abundance, and
present and potential hazard to man, and the world
biota. These include antimony, arsenic, cadmium,
chromium, cobalt, copper, lead, mercury, nickel,
tin, vanadium, and zinc. These metals occur in
elemental form or in various chemical compounds
that vary widely in toxicity, absorption, and
retention.
Copper, mercury, tin, cobalt and nickel are
toxic to plants below 1 ppm. Orally arsenic is
highly toxic to animals, and cadmium, copper,
mercury, lead, antimony, and vanadium are moder-
ately toxic. When injected, cadmium, chronium,
mercury and vanadium are toxic. Some of the toxic
metals are retained in the body for long periods of
time, thus functioning as cumulative poisons in-
cluding arsenic, cadmium, lead and mercury.
Evidence is accumulating that beryllium, chromium
and nickel cause cancer. Cadmium and lead, at the
level to which many people are exposed, cause life-
shortening and frequent neurological disruption in
experimental rats.
There are many conflicting claims based on
limited scientific facts about the levels of these
metals in the environment as each new discovery
occurs. It is now necessary to put all of the
available data into perspective to find out how badly
we are polluting our environment and how the metals
ultimately flow to man.
Environmental Flow
Flow of these metals through the environment is
mediated by both natural and man-induced sources.
Man has greatly increased the mobilization of many
of the toxic metals and has increased the flow of
many of these metals and thus increased their hazards.
A model flow diagram (Figure 1) shows the various
pathways and flows of these metals through air,
terrestrial ecosystems, freshwater, and the ocean.
Most of the sources come from the terrestrial
Tain,
fallout
direct use
biota
flux
fraah watar
& marina
producta
watar uaa,
Irrigation
biota
flux
AIR
TERRESTRIAL ECOSYSTEMS
(man, lnduatry)
FRESH HATER
dapoaltlon (uplift ?)
OCEAN
ENVIRONMENTAL PIXM UP TOXIC METALS IN THE WORLD
Sources and Emissions
Man has extracted these metals from the earth
for many years and the amounts show increases
nearly every year. The 1972 world production^® of
these metals is shown in Table 1. Copper, zinc,
lead and chromium are mined in greatest quantities,
but some with smaller production are highly toxic.
After extracting these metals, man reintroduces
them into the environment directly in elemental form
or in a wide variety of compounds some of which are
more toxic than the elemental form.
The major emission sources of the toxic metals
are shown in Table 2. These emission sources
include smelting and metallurgy, combustion of coal
and oil, industrial effluent discharge, incineration,
scrap recovery, combustion of gasoline, direct use
1
1-1
-------
of metal products and compounds, and use of pesti-
cides, herbicides and fertilizers.
Table 1
TOXIC METAL POLLUTION OP THE ATMOSPHERE
AMOUNTS Of METALS jl
1000 TONS
C0NC.IH
AIB #ig/a3
19?2
World
Production
Saelting
4 Inci-
neratioa
Coal &
oil
coabustauQ
Total nan
caused
pollution
Urban
Bural
Antiaony
62.9
0.2
0.001
Arsenic
59.0
0.7
0.02
CadniuB
14.0
3.0
lu.O
u.002
Chroaiua
2,770.0
1.5
20.0
0.015
Cobalt
18.0
0.7
<0.0005
Copper
Lead
6,780.0
3,410.0
(a
300.0
2.1
3-6
200.0
303.6
0.09
l.il
0.06
0.0005
Mercury
8.5
10.0
5-5
36.0
0.00$
0.0007
Nickel
614.3
3-7
69.0
u.034
0.002
Tin
196.5
u.28
30.0
0.02
Vanadiua
11.8
12.0
20.0
0.05
Zinc
5,550.0
7.0
0.67
(a. gasoline combustion (TEL)
Table 2
ailSSIOH 8QURCES OF TOXIC WEEAL&
S«#lting
&
Metallurgy
Coabustion
coal
oil
Industrial
effluent
(.water)
Inciner-
ation &
scrap
recovery
Gasoline
whiiiMnn
Direct
use
Barbette
Sl
ferti-
lizer
Antimony
X
Arsenic
XX
X
X
X
X
Cadnxua
XX
X
X
X
X
Chroaiua
X
X
X
X
Cobaic
A
X
Copper
XX
X
X
X
i/eau
aX
XX
X
X
tX
*
X
Mercury
X
XX
X
X
X
X
nickel
xx
X
X
X
X
Tin.
X
Vanadius
XX
XX
Zinc
XX
XX
X
X
X
Atmosphere and Ocean
An attempt has been made to quantitate the
emissions which cause pollution of the atmosphere,
(Table 1). This is based on data from many sources
cited in the references, and includes smelting and
incineration, and combustion of gasoline, coal and
oil. Estimated total man-caused pollution is shown
and urban and rural air pollution data for comparison.
Twenty two metals have been found in polluted
air in the United States. Four of these constitute real
or potential public health hazards--lead, cadmium,
nickel, and mercury.
The ocean is the object of both natural and man-
induced pollution, and Table 3 shows the concen-
tration of the metals in ppb as well as the total
content in millions of tons. The metal mobilizations
for both natural geological and man-induced sources
are presented for comparison.. Also the amounts of
metals added by the rivers and by atmospheric
washout are shown.
It is interesting to compare the total atmospheric
and ocean pollution caused by both natural and man-
induced with the amount mined by man each year.
For mercury and vanadium several times the amount
mined is added, for arsenic and cadmium it is about
the same general level, for cobalt about one third,
and for the rest much less is added than mined.
Table 1
TOXIC METAL FOLLUTIOH ar THE nrmw
X 1000 TOS8/YH.
Oceaj
3
Metal aobilization
Added to
ocean/yr.
Cone.
PPB
Total
content
x million
tone
Geological
Man induced
fly rivers
By ata.
vashout
Antiaony
0.45
617
1.3
t>3
10
jLTuenic
2.0
2,740
59
72
Cadaiua
0.02
27\
14
0.5
10
uhroiDJ.ua
u.Ot
55
2,770
36
20
Cobalt
0.5
685
18
7
Copper
1.0
1,3?0
375
6,760
250
200
Leaa
0.04
55
180
3 ,410
110
300
Mercury
0.1
137
>
8.5
3.8
80
flicxei
2.0
2,740
300
614
11
30
Tin
3.0
4,110
1 .5
197
2
3°
Vanfc'lina
2.0
2,740
li
20
Z.Lt
i c.e
i;,?0C
y;o
(-LC,'t".. "j
" , %<-
JN
^eertsooK
(1974)
720
Bertini
&
uuidoog
(1971;
Environmental Studies
Environmental flow of toxic metals is becoming
much better understood and quantitative data are
becoming available as the result of a number of
detailed environmental studies. The movement,
pathways, and much quantitative information was
obtained by the AEC and other groups from many
experimental studies using radionuclides and from
nuclear explosions. The intensive studies of the
International Biological Program (IBP) and
especially the large scale studies of the International
Decade for Oceanographic Exploration (IDOE) have
contributed much. The RANN program of the
National Science Foundation (NSF) have supported
many studies on heavy and toxic metals including
IDOE, and such studies on lead and mercury and
other metals in the lead belt in Missouri and at the
University of Illinois. The Illinois study on the
quantitative flow of lead from source through many
pathways including biological organisms to runoff is
of great value. Various other research programs
and many detailed monitoring programs of toxic
metals in the air, freshwater, the ocean, soil, and
biological organisms provide much valuable
environmental flow information when assembled and
evaluated.
Mathematical models of toxic metal flow into
the various compartments of the environment from
source through various pathways to final destination
have been developed by several groups. Of especial
value is the development of integrated monitoring
systems which take into account critical sources, and
critical pathways to the critical target (usually an
exposed population of human beings). This concept
2 1-1
-------
permits development of a highly effective monitoring
system with relevant and useful specific data per-
mitting intelligent action to be taken on emission
sources.
Biological Organisms
Biological organisms are of great importance in
accumulating, transporting, and changing the form
of toxic metals.
A comprehensive world-wide bibliography,
literature review, and evaluation of the accumulation
and concentration of these toxic metals in biological
organisms is being completed under a contract with
the Environmental Protection Agency. This includes
the species (common and scientific name) measured,
the geographic location, special conditions (as near
a smelter, outfall, or roadside), organ or tissue
measured, levels of metal in ppm, and how measured
(wet, dry, or ash), and the authority. AU of the
reported measurements have been converted to ppm
except measurements of radioactivity. The tables
are organized for each metal, in phylogenetic
sequence of the various organisms.
This compilation allows an evaluation to be made
of the relative importance of biological OTganismB in
the environmental flow of these toxic metals. The
data are also of value for selecting the most valuable
biological Organisms for use in biological monitoring.
Maximum concentrations of toxic metals are shown
for various taxonomic groups of biological organisms
in Table 4. The data are given in wet weight or dry
weight if preceded by a letter D.
organisms. Concentration factor has frequently been
used in the literature comparing the level of the
metal in the organism with the level in its aquatic
environment. Calculations were discarded which
used dry or ash weight which would give errors from
about 3 to 25 times too high. Many calculations were
based on old or dubious measuring methods which
were also deleted. Algae and other aquatic plants
would have the factors for enrichment and concen-
tration the same. The maximum enrichment factors
were found in birds and mammals and molluscs. No
data were used from known heavily polluted areas,
but pollution data were frequently not given so that
some of the figures are probably too high.
mirrmw raBICffllBrT TiOVltB Of KAMI UK aiGAMISKS
EKBXCHHKHT
flCTOfi a 1.
000
HOfO
hBO BO
o <
s ?
Alga®
(a
zoo plankton
(Ave.)
Inverte-
brate*
incl.
Crustacea
Mollusci
JIB*
Marine
—ah
k
birda
Antiaony
0,45
0.2
0.2
0.1
32.0
0.1
Arsenic
2,0
7-5
36.0
40.0
10.0
1,0
Oadaiiim
0,02
100.0
30.0
70,0
390.0
1,25.0
780.0
Chroaiu*
0,04
17.0
25.0
25-0
320,0
2-5
Cobalt
0,5
0.8
3-2
6.0
6.0
2,0
Copper
1,0
10.0
4.9
20.0
350.0
7.5
1,2
Lead
0*04
500.0
147.0
375.0
1,000.0
75-0
300.0
MarcAx?
0,1
0.8
6.4
10.0
31.0
1,730.0
Kickel
2,0
3-0
6-5
2.5
7.9
0.3
•1'i.u
u.43
1.5
0.5
Vanadiua
2.0
1.0
0.5
103-0
4.5
4.0
Zinc
10.0
13.6
6.0
57.0
500,0
160,0
30
(a Ave, (Sot aaxisua)
Ikoovs heavy polluted ar«a* excluded)
Table 4
w.aTWIM QQHGEgTBATIOHS OF TOXIC METALS !¦ BIOLOGICAL QBOAMlBHfl
(21 PFK VET WEIGHT)
Biological Monitoring
Antimony
Arsenic
Cadniua
Cobalt
Copper
Leao
Mercury
Hickel
Tic
Vanadi.UK
Zinc
The term "enrichment factor" is used to show
the relationship of the metal content of an organism
or organ (wet or live weight) to the level of metal in
the ocean or freshwater. Enrichment factors have
been newly calculated for large numbers of marine
organisms using recent measurements of metal
content in the ocean. Maximum enrichment factor
levels were compiled (Table 5), for use in selecting
organisms that may be of potential value in biological
monitoring.
The term "concentration factor" is used only to
show the accumulation or concentration of a metal by
a species over the level of metal in itB food
3
The detailed tables on biological accumulation and
concentration of toxic metals are being used to design
biological monitoring systems and networks. New or
improved biological monitoring concepts have been
developed. It is necessary to measure both high and
low trophic levels in food chains and networks.
However, highest trophic level predators such as
eagles, hawks, and owls cannot be killed and
measured. A new biologic monitoring concept is
recommended using carrion feeding scavengers such
as vultures, crows, arid ravens which have been
shown to have high levels for some of the toxic
metals. These scavengers may feed on organisms
which have become weak or died because of heavy
toxic metals. They may also feed on dead high
trophic level predators. Crabs and crayfish are good
scavengers, but highest trophic level predatory fish
can be monitored as well as aquatic concentrator
organisms such as molluscs.
Another concept is to monitor the concentrator
organs or tissues with highest concentrations such as
hair, feathers and nails for mercury and arsenic,
bone and liver for lead, and kidney for cadmium.
Another biological monitoring concept is to conduct
monitoring on organisms closely associated with man
and impact areas where man lives and works, and
organisms which share his food. The Norway or
Brown rat and the pigeon are obvious candidates.
1-1
Algae
Higher
piante
Invertebrates
incl.
Crustacea
Molluscs
fish
Birds fc
ftawale
D 0.09
D 0.0b
i; 0.24
0 90
D u.2
0 0.35
o 90
51
174
119
26
5.6
D 20
D 50
1.1
0 2000
23
330
D 1£
1
D 14.7
U
10
1.8
0.4
D 1.1
2
3.4
1.1
15
300
i) ?.B
50
3,300
40
0 80
D 1,200
0 35,000
li 1,200
148
D 50
297
D 14
59
D 100
30
j6
VOO
D 58
D 23
13
15-8
» 33
» 6
D 3.8
B 2.3
D 32
D 15T
5-4
B <0.16
0 16
D 1.6
0 620
0 44
10
D 0.05
138
D 160
D 566
8,582
1»597
D 120
-------
Monitoring these organisms would help determine
biological levels of the pollutants in the environment
of man and relate these to adverse effects on man.
Historical Trends of Toxic Metals
The historical record of toxic metals provides
data to show the relative amounts of the metals in the
environment in the past compared with the present.
This record can be obtained by analyzing certain
metals in permanent snow and ice fields, in caves,
in levels of the ocean, in tree rings, and perhaps in
bog and varve layers. Museum collections provide
specimens which have been used to show trends such
as hair, bones, feathers, mosses and other plant
specimens. All specimens must be used with care
particularly those in liquid preservatives which may
have been contaminated during collection or storage.
The historical trend of lead in the environment
is shown in Figure 2. This shows® the change in
levels from 800 B. C. to I960 A. D. The lead level
in mdss is shown in ppm, elm tree rings in parts
per 100 million, Poland glacier in ppb, and Green-
land glacier in parts per trillion. An increase is
shown after 1750 correlated with smelters, and in
the late 1880 period with coal heating of houses, and
a dramatic increase about 1940 correlated with lead
in gasoline.
There have been very great increases of mercury
shown in bird leathers correlated with use of mercury
in the wood pulp industry and for seed coating to
prevent mold. Some of the high levels have
decreased recently with curtailed use of mercury.
Arsenic showed a five-fold increase in tobacco from
10 ppm in 1918 to 52 ppm in. 1952, but dropped to
8 ppm shortly after 1952 when lead arsenate was
banned and DDT was substituted. Copper, zinc,
chromium, and nickel have shown increases in
mosses from 1880 to 1970, and increase of cadmium
in elm tree rings from 1880 to I960. When more
data on historical trends are available it should be
possible to predict trends into the future if emission
sources are not controlled or decreased. This is a
fascinating field but it is difficult to interpret and
to decipher the records.
Pathways to Man
The environmental flow of toxic metals to man
may be direct or extremely complex and circuitous.
It may be as direct as man breathing mercury fumes
in a mine or smelter, or lead from gasoline exhaust,
or nearly direct when a small child gets lead dust
on his hands or food and into his mouth. On the other
hand the route may be very indirect going through the
atmosphere, to freshwater, to the ocean, and through
a long and complex food chain or network ending up
in a high trophic level predator such as a swordfish
and finally getting to man. The metal may go through
several types of compounds and physical forms before
reaching man.
In many cases these complex pathways are not
known. A metal may be a hazard because of multiple
sources and pathways all of which contribute to
increasing the dose. Little is known of the multiple
effects of several toxic metals acting together or
synergistically.
Since man is ever increasing the amounts and
uses of toxic metals each year, and is polluting
the environment more (with some exceptions such
as mercury) it would appear that these metals
will become of greater importance as health
hazards unless they are controlled wherever
possible.
References
1. Qertine, K. K. andE. D. Goldberg, 1971.
Science. 173:233-235.
2. Billings, C. E. and W. R. Matson. 1972.
Science. 176:1232-1233.
3. Bowen, H. S . M. 1966. Trace Elements in
Biochemistry. New York. Academic Press.
4. Council Env, iQual. 1971. Toxic Substances.
Apr. 1971 :pp 25.
5. Goldberg, E. D. 1970. The Chemical Invasion
of the Ocean by Man. McGraw Hill Yearbook
Science Technol.
6. Golds chmidt, V, M.
Oxford Univ. Press.
7. Hazards of Mercury.
4:1-69.
8. Jenkins, D. W. 1972.
9. Jenkins, D. W. 1975.
for Environmental Pollutants.
Contract 68-03-0443. pp 317.
10. Joensuu, O. I. 1971. Science, 172:1027-1028.
11. Klein, D, H. and E. D. Goldberg. 1970. Environ.
Sci. Technol. 4:(9)765-768.
12. Livingston, D. A. 1963. U. S. Geol. Surv. Prof.
Pap. 440 G.
13. Mason, R. 1958. Principles of Geochemistry.
New York. John Wiley & Sons, Inc.
14. Mercury in Waters of the United States (1970-71).
1972. U.S. Dept. Interior.
15. Mineral Yearbook. 1974. U. S. Dept. Interior.
U. S. Govt. Print. Off.
16. Morgan, G. B., G. Ozolins and E. C, Tabor. 1970.
Science 170:289-296.
17. Nat. Acad. Sci. 1972. Lead. Airborne Lead in
Perspective. Nat. Acad. Sci. pp 330.
18. Patterson, C. C. 1965. Arch. Environ. Health.
11:344-363.
19. Schroeder, H. A. 1971. Environment 13:18.
20. United Nations Yearbook (1974). 1975.
21. Weiss, H. V., M. Koide and E. D. Goldberg.
1971, Science. 174:692.
22. WlUiston, S. H. 1968. J. Geophys. Res.
73:7051-7055.
1954. Geochemistry,
London pp 730,
1971. Environ. Res.
Smithsonian 3:{l):62-69.
Biological Monitoring
Rept. EPA
4
l-l
-------
Figure 2
Trends of lead m glaciers and plants
,enlandj9*acie^
TTTOSS"
800B.6. 1750A.D. 1860 1880 1900 1920 1940
5
1-1
-------
SAMPLING AND ANALYSIS OF METALS IN AIR, WATER AND WASTE PRODUCTS
James J. Morgan
W. M. Keck Laboratories of Environmental Engineering Science
California Institute of Technology
Pasadena, California 91125
Summary
Monitoring, sampling and analysis for different
environmental compartments are discussed from the point
of view of source-flow-receptor and chemical element
balance concepts. Needs for vapor-particulate fractions
in the atmosphere and for corresponding dissolved-
particulate fractions in the hydrosphere are examined.
The metals of interest in monitoring in air and water
are compared, and typical levels and ranges for metals
are reviewed. Standard, tentative, and approved
methods of analysis in the U.S. are summarized for both
water and air. Effects of sampling, storage, processing,
and pre-concentration steps in overall analysis are dis-
cussed. Various procedures now in use for collection,
pre-concentration, and sample preparation are reviewed,
with relation to accuracy of methods. A range of
analysis methods used in monitoring or potentially use-
ful in monitoring is examined. These include: INAA,
PES, ASV, SSMS, XRF, and NFAA. Needed developments
with respect to speciation and particle sizing are
discussed.
Introduction
The subject of environmental sampling and analysis
of metals is enormous, and only a small part of it can
be addressed in this paper. I have several objectives
in mind in discussing monitoring of metals. One objec-
tive is to discuss some of the general characteristics
of the monitoring of metals in air and water environ-
ments and in sources and to Indicate cross-corrections
of the two media. A second objective is to draw
attention to what is now known and has become more or
less standard practice in the United States (while
recognizing that change is constantly taking place). A
third objective is to point out needs for new work on
metals, in both the areas of application and research.
Over the past decade a framework for thinking
about metal pollution (as well as pollution by syn-
thetic chemicals) has been developed which can be
characterized as global and regional. This framework
is exemplified by the work of Bertine and Goldberg on
element mobilization via fossil fuel combustion, the
discussion by Dyrssen, et ai.2 on inorganic chemicals
in the marine environment, and by the comprehensive
treatment of certain pollutants in the Southern Cali-
fornia region by the Caltech group. 3The key
concepts in the regional assessment of pollutant impact
are: 1) the flow of pollutants from sources to various
environmental compartments (land, groundwater, atmo-
spheric receptor sites, coastal waters, rivers, lakes)
within the region, and removal from the region by wind
or water transport;^ 2) the chemical element balance
approach relating sources to environmental concentra-
tions, developed by Friedlander.5 Gordon, et al.^ have
examined regional source-receptor relationships via an
enrichment factor approach; Gladney, et_ al^ have noted
that joint elemental composition and air particulate
size distributions are so distinctive in some urban
areas that they can be used to identify sources of
pollutants found at different sites.
Regional studies of the flow of metal pollutants
led to the identification of important flows between
the air and water environments. Significant examples
of recent progress in this aspect of monitoring are
provided by the studies of Bruland, et^ al.® on metal
accumulation rates in Southern California marine
sediments, the work of Huntzicker, Friedlander, and
Davidson3 on flows of Pb, Zn, Cd, and Ni in the Los
Angeles basin, and the mass-balance model for Pb in the
Southern California Bight of Patterson and Settle.3
For the coastal waters of Southern California dry
aerosol deposition, rain, and storm runoff are impor-
tant inputs of Pb on an annual basis.
The discharge of primary or secondary effluents to
fresh and marine waters can add large quantities of
metals in both particulate and dissolved forms. Monitor-
ing of effluents in the Los Angeles areas^.lO shows that
some metals exist predominantly in particulate form
(e.g., Pb, Cu, Zn) in primary effluent, while other
metals exist predominantly in solution form (e.g., Ni).
In secondary effluents, the fractionation can be com-
pletely different. Chen1® has attempted to describe
the metal content of wastewaters with respect to parti-
cle size classes; Morel, et al.^ have examined the
settling velocity distributions of metals in various
particle size fractions. We have made attempts to under-
stand the basis for observed dissolved-particulate
fractionations in terms of models incorporating chemical
reactions and adsorptive equilibria. H>12 The metals
considered were: Fe, Mn, Zn, Cr, Cu, Pb, Ni, Cd, Ag,
Co, and Hg (in addition to the principal cations). For
primary (oxygen-free, high sulfide) effluents, the
chemical models predict predominantly solid phases for
all but Ni and Mn. Mixtures of effluent with seawater
(dilutions of 1:10 to 1:1000) have been modeled and the
models predict extensive chemical change and consequent
alterations in the dissolved-particulate fractionation.
General patterns of chemical speciation in seawater have
been worked out by Stumm and Brauner}3
The foregoing ideas and observations have some
important implications for monitoring, sampling, and
analysis. A recent National Academy of Sciences report1^
has described monitoring as "the process of following a
specified chemical through the environment," with the
operational meaning of measurements made over time. The
report suggests that monitoring and development of
critical-point exposure level models should go together.
Thus, choices of when and where to sample should be
guided by monitoring-flow model developments incorpora-
ting current understanding of source distributions,
meteorology, and hydrology (prevailing air movements for
atmospheric transport and current-mixing patterns for
water transport). The ultimate monitoring need with
respect to evaluating and controlling exposure of human
populations is chemical, not elemental. It will also be
necessary to know the sizes in which chemical entities
occur (e.g., respiratory effects in humans, further
atmospheric transport; sedimentation, selective feeding
and transport in aquatic ecosystems). Thus, in the air
environment and in the water environment, the monitoring
information needed ultimately is chemical entity and
size distribution. It is my impression that rather
little direct chemical information about metals in air
and water has been obtained up to now. Some important
exceptions are the forms of mercury in the atmosphere,15
organic mercury forms in water and sediments^and
the identification of PbSO^ in soils by Olson and
Skogerboe.17 Of course, important inferences about
chemical forms in atmospheric particulates can be drawn
from knowledge of the sources (e.g., lead halides in
auto exhaust) and from chemical element balances for an
aerosol. In aquatic environments, we can often make at
least rough calculations of the abundance of various
1
1-2
-------
complexes of a metal in solution when we know the total
concentrations of elements in the solution. Selective
leaching reagents may be of use in indicating fractions
of different reactivity in a sample of sediment
(exchangeable, oxidizable; reducible, etc.). My
emphasis on tfie desirability of data on chemical enti-
ties is not a criticism of present methods and
standards for elements. Rather, I think it is clear
that we will move toward a more and more chemical
approach in standards and monitoring as we gain experi-
ence.
Which Metals'?
A difficulty encountered at the outset of an
attempt to discuss air and water analysis jointly is
the great variety of lists of elements which seem to
be of interest in different analytical contexts. I
suspect that this might be a minor disadvantage to
analysts joining the environmental field. Which are
the elements of greatest interest? (It's interesting
to note in passing that of the 104 elements, 21 are
non-metals. These include As and Se,atomic numbers
33 and 34, which I frequently find on lists of "toxic
metals" and occasionally on lists of "heavy metals."
The elements characterized as "heavy metals" almost
always include Be, Ti, V, and Cr. Perhaps we should
agree cheerfully to accept As and Se into the "metal"
family and forget, the "heavy" burden? Or we might let
it go at elements, without further qualification.)
I consulted a number of sources to become
acquainted with the elements of interest with respect
to environmental standards and monitoring for both
air and water environments: Water Quality Criteria
(the NAS-NAE document, 1973), the U. S. EPA "Interim
Primary Drinking Water Standards" (1975)"Guide-
lines Establishing Test Procedures for the Analysis of
Pollutants" (water, 1973),20 "Methods for Chemical
Analysis of Water and Wastes";2-*- U.S. National Air
Surveillance Network,22 G. B. Morgan, et al., "Atmo-
spheric Surveillance" (1972)," the Intersociety
Committee's "Methods of Air Sampling and Analysis"
(1972),and Lee's useful volume, "Metallic Contami-
nants and Human Health."25 As might be anticipated,
lists of varying length are encountered in such a
search. For water, the 1975 interim primary drinking
water standards set maximum levels for these 8 metals:
As, Ba, Cd, Cr, Pb, Hg, Se, and Ag (this strange
ordering is the result of my mismatch of "chemical"
alphabetics and the English language names of ele-
ments). The 1972 Water Quality Criteria book
discusses the following 22 metals with respect to
either drinking water, fresh waters, or marine waters:
Be, Na, Al, V, Cr, Mn, Mo, Fe, Ni, Cu, Zn, As, Se, Ag,
Cd, Sb, Ba. Hg, Pb, Bi, U, and Tl. The 1973 "Guide-
lines..." lists EPA approved test procedures for the
following 27 metals: Al, Sb, As, Ba, Be, B (!), Cd,
Ca, Cr (total and hexavalent), Co, Cu, Fe, Pb, Mg, Mn,
Hg, Mo, Ni, K, Se, Ag, Na, Tl, Sn, Ti, V, and Zn.
G. B. Morgan's atmospheric surveillance list^ is
the lengthiest of those dealing with air monitoring.
The significant difference between this list and the
water guidelines list is the inclusion of Al, Ag, and
Tl on the latter. The trace metals considered in Lee's
volume2-^ are: As, Be, Cd, Cr, Pb, Mn, Hg, Ni, Se, and
V.
Although no firm analytical priorities are to be
derived from a cursory examination of this sort, it
would seem that the elements of likely general inter-
est for monitoring programs including air, water, and
waste products affecting both media are Be, V, Cr, Ni,
Cu, Zn, As, Se, Ag, Cd, Ba, Hg, and Pb. In addition,
of interest because of their potential for regulating
behavior of other metals in the environment are Fe and
Mn. Finally, Dyrssen, et_ al_. have listed Ti, Al, Sb,
and Bi as potential pollutants of the marine environ-
ment, on the basis of world production figures.
What Metal Levels are of Interest?
It is essential that the analyst have reasonably
correct notions about metal levels of interest in air,
water, or waste products in order to begin to evaluate
potential methods of analysis, assess the potential for
contamination effects, and select pre-concentration
methods where necessary. Levels of metals in atmos-
pheric particulates range from on the order of 1 -
10 v% m (Pb, Fe, Al) to 0.001 n/m3 (Be, Cd). The U.S.
NASN" can be consulted for recent data; Israel and
Israel26 summarize earlier U.S. data. Metal concentra-
tions in NBS standard reference fly ash have been
determined recently by Ondov, et_ aj^.2^ Rhodes, et al?^
has reported extensive data for 17 elements in air
particulates in Texas. Metal vapor levels in the atmos-
phere are less well known, but values of perhaps 0.1 to
1 ng/m3 might be assumed.
Trace element concentrations (17 metals); for U.S.
fresh waters for the period 1962-67 (Kopp and Kroner,
FWPCA, 1967) have been summarized in concise form in
"Instrumentation for Environmental Monitoring" by the
Lawrence Berkeley Laboratory.2^ Burrell's book on
atomic spectrometry for metal pollutants^O contains a
useful summary of recent data on metals in natural and
polluted waters. Metal concentrations in seawater are
under intensive investigation in many quarters, e.g.,
the GEOSECS program. There is a noticeable tendency
for published values for several of the trace metals of
seawater to decrease with the years ("The longer we
study it, the lower it gets"), almost certainly a
reflection of increased recognition of contamination
problems and improved techniques in collecting,storage,
and processing samples. The range of trace metal
levels in open seawater is from about 1 ng/1 (Be) to a
few ug/1 (Fe, Ni, Zn). Coastal water values will tend
to be higher. Our state of knowledge about dissolved
vs particulate fractions of trace metals in seawater is
still very far from satisfactory, as evidenced by the
difficulties in characterizing "soluble" Fe.31 The
range of published values for 11 trace metals (through
1972) in surface seawater was summarized by SCCWRP.9
The same group has recently (1975) reported their own
findings for dissolved and particulate levels of Cd,Cr,
Cu, and Ni in the Southern California Bight.32
Metal contents of coastal marine sediments have
been found to be greatly altered by pollutant discharges,
both from sewage and from atmospheric transport and
fallout and rainout of particulates.9'^ For example,
background sediment levels of metals near Southern
California range from about 0.05 ug/g (Hg) to about
50 ug/g (Zn) to about 25,000 ug/g (?e). Recent surface
sediments near wastewater and sludge outfalls show
enrichments of Hg and Zn by factors of about 50 to 10,
respectively, while Fe is little altered (slightly
depleted).9 Trace metal levels in wastewater particu-
lates and sludges can be quite high, and expectedly
variable. For example, Los Angeles 3ewage siudge (1971)
had Pb levels of ca. 1000 vig/g and Cr levels of ca.
3500 ug/g.9
The ranges of environmental metal concentrations to
be dealt with might be illustrated by considering lead.
The following values are of interest:
Pb
19
U.S. drinking water standard 50 yg/1
13
Typical surface seawater '*0-80 ng/1
2
1-2
-------
Los Angeles sewage effluents
9
Los Angeles sewage sludge
Southern Calif, marine aerosol
or
Southern Calif, urban aerosol
or
Southern Calif, marine sediment
34
,34
Pb
50-250 ug/1
1000 mg/kg
3
0.1 ug/m air
4000 yg/g aerosol
3
3 ug/m air
16,000 yg/g aerosol
10 ug/g
Metal
As
Ba
Cd
Cr
Pb
Hg
Se
Ag
Level, vr/1
50
1000
10
*
50
50
2
10
50
(1962 Standard
specifies hexa-
valent Cr)
The standards also specify sampling and analysis
requirements. An AAS method is specified for all but
Hg, for which the NFAA analysis is specified. For
details of the specified methods "Standard Methods"-''
or "Methods of Chemical Analysis of Water and Wastes"^®
should be consulted.
ASTM"^ provides Standards for determination of the
following metals in water: Al, As, Cd, Cr, Co, Cu, Fe,
Pb, Mn, Hg, Ni, Se, Th, U, and Zn (AAS methods are
described in general terms for Cd, Cr, Co, Cu, Fe, Pb,
Mn, Ni, and Zn.).
The U.S. EPA "Guidelines..." for water pollution
test procedures offers a list of approved test proce-
dures for 26 trace metals
20
Basically, the list is a
Information of this sort should be useful for prelimi-
nary evaluation of suitable analysis methods. Hope-
fully, early consideration of expected levels of metals
in different media can sharply reduce the number of
"less than detection limit" entries so often found in
surveillance data reports.22,35
Standard, Tentative, and Approved Methods
To my knowledge there are no air quality standards
for metals in the general environment. According to
the 1973 ASTM Book of Standards^ there is a tentative
method for atmospheric lead, both in particulate and
vapor form. The Intersociety Committee^ describes
tentative methods for Sb, As, Be, Fe, Pb, Mn, Mo, and
Se (1972). (Methods become "standard" after satisfac-
tory completion of cooperative test procedures using
tentative methods.) The Intersociety Committee's (1972)
methods are all colorimetric, fluorimetric, or volu-
metric. I assume that instrumental methods for metals
in atmospheric samples are now undergoing test pro-
cedures which may lead to tentative or standard
procedures. There has been an impressive output in
the past three years of papers dealing with the environ-
mental application of such techniques as atomic
absorption spectrometry (AAS), non-flame AAS (NFAA),
instrumental neutron activation analysis (INAA), and
x-ray fluorescence analysis(XRF) and ion induced x-ray
emission analysis (IIXE). Some of their applications
to metal monitoring in the atmosphere will be refer-
enced later in the paper.
"Standard Methods for the Examination of Water and
Wastewater,"^7 describes tentative and standard methods
for these trace metals: Al, As, Ba, Be, Cd, Cr, Cu,
Fe, Pb, Li, Mn, Ni, K, Se, Ag, Na, Sr, V, and Zn.
Flame AAS methods are given in detail for Cu, Zn, Fe,
and Ba (direct flame) and for Cr, Mn, Ag, Cd, Pb, Al,
and Be (complexation and extraction prior to flame).
The U.S. Interim Primary Drinking Water Standards
(1975)1® set out the following maximum contaminant
metal levels:
combination of the "Standard Methods,"5' ASTM, and
OO * 7
EPAJO methods. It iB interesting to note that some
form of atomic absorption spectrometry is either the
sole or alternate method in these approved test proce-
dures. Total metal is determined by not using a prior
membrane filtration step and by subjecting the whole
sample to vigorous digestion. In discussing the method
selection criteria the EPA methods report^® identifies
four points: 1) sufficient precision and accuracy in
the presence of normal interferences; 2) needed equip-
ment and skills should be available in the typical water
laboratory; 3) the methods should be in use in many
labs and sufficiently tested to assure their validity;
4) methods should be sufficiently rapid for routine
analysis of large numbers of samples.
It should be fairly evident that as of 1975 the
practical determination of trace metals In water and
wastewaters is dominated for most metals by AAS (with
or without prior complexation as required) and by cold
flameless atomic absorption methods for mercury and
selenium. Spectrophotometric methods continue to be
used for a number of metals, but to a far less degree
than five years ago. Non-flame atomization appears not
to have been widely used yet in order to carry out
routine surveillance and compliance monitoring; this
technique appears certain to see wider use in the
future.
Sampling, Storage, Concentration, and Analysis
In his useful and wide-ranging review (182 refer-
ences) of developments in analysis of toxic elements,
Lisk observes that "in methods such as anodic strip-
ping voltammetry (ASV) and NFAA it may be background
contamination from reagent impurities and surroundings
rather than Instrument sensitivity which controls the
limit of detection." He further suggests that future
improvements may well be in ultrapurification of
reagents and In "clean-room techniques." Altshuller^
in his remarks on analytical problems in air pollution
control, speaking of high-volume air sampling procedures
in use for air monitoring, said: "It should be apparent
that...sampling problems cause greater obstacles to
overall analytical accuracy than those caused by limita-
tions in the analytical techniques themselves." Tolg's
(1975) article on "Elemental Analysis with Minute
Samples"^ is recommended reading for those seeking a
general introduction to the positive errors caused by
impurities of reagents, impurities in laboratory air,
desorption and ion exchange from container walls, and
the negative errors brought about by adsorption effects
on surfaces and volatility or other losses of elements.
Impurities (blank values) can seriously influence the
reproducibility and limit of identification of the
analysis method. Materials introduced by filters as
well as those retained (when not expected to) need to be
accounted for. Tolg stresses the fact that neither
ultrapure nor completely inert stock materials are known;
only vessels made of quartz, TFE Teflon, and to a
limited degree polyethylene are suitable for long-term
storage of solutions following purification.
Patterson and Settle^ discuss procedures to be
followed in lead analysis of fresh waters, snow, sea-
water, and biological materials in order to control
industrial lead contamination during sample collection
3
1-2
-------
and analysis. A recent report33 on "Interlaboratory
Lead Analyses of Standardized Samples of Seawater"
thoroughly discusses clean—lab procedures, blanks,
contamination, high-purity reagents, laboratory water,
etc. as these affect analytical results. This report
should be of particular value to workers interested in
metal analyses in the 1-100 ng/kg level, using such
techniques as atomic absorption and anodic stripping
voltammetry.
The total system of sampling, storage, laboratory
processing, pre-concentration, and analysis needs care-
ful scrutiny for contamination, loss, and factors
affecting the magnitude and variability of the blank.
Contaminant mgtal levels in filter materials and
impactor surfaces are now much better known (s®®> e.g.,
Dams, et al., 3 Hwang?5 and Dzubay and Stevens44).
Organic membranes and polystyrene fiber filters tend
to be much superior to glass fiber and silver membrane
materials from the contaminant aspect. Filters may
need to be chemically cleaned (e.g., with aqua regia)
before use. Millipore filter chromium blanks have been
reported to be sometimes high.®
About Sampling
It is axiomatic that samples ought to be represen-
tative of the water, air, sediment, or waste stream
that are being monitored. Great care must be taken to
avoid contamination from boats, pumps, metal wires, oil
slicks, unjleaned sampling bottles, etc. Again, refer-
ence to Patterson and Settle's recommendations4^
regarding water sampling is suggested. The care taken
should be consonant with the levels of metals being
monitored and with the various contaminant risks
involved (watch out for helicopters burning leaded
gasoline, and avoid sampling astern!). Standard
Methods3' and the EPA Methods book38 provide valuable
guidelines for not-too-stringent water sampling situa-
tions. Effluent samples are often composited over
day-long periods.
In the water environment it is sometimes useful
to distinguish between dissolved and particulate frac-
tions, although this distinction is not yet clear-cut
in conceptual terms. Millipore filters of 0.45 ym,
0.22 um, or 0.1 um have been used for more than a
decade to separate "dissolved" and particulate. It is
now established that the newer Nucleopore filters be-
have more nearly as screens (i.e., they discriminate on
basis of particle size) than do the cellulose acetate
membranes (Millipore). Thus, estimates of particle
size distributions are probably better made with Nucleo-
pore filters^5 so long as they are lightly loaded.
Cellulose acetate filters have smaller effective pore
sizes than nominal, and are thus quite effective for
total particle retention.4® Chen's work^O 0n sewage
particles is noteworthy.
Sedimentation analysis is of potential value in
segregating aqueous particles into settling velocity
classes. Metal content (and other elemental content)
of particles in the different velocity classes can then
be determined.^ Sediment collectors positioned at
various depths in coastal waters are being used in the
Southern California Bight32 to estimate in situ sedi-
mentation rate of sewage effluent - natural particle
mixtures. We use high-speed centrifugation in our
laboratory to complement membrane and Nucleopore filter
separations, but we have not learned of any systematic
use of the centrifuge to classify particles in water
samples.
Air sampling of particulates is a time-averaging
procedure, unlike most ambient water sampling. The
traditional high-volume air samplers with large filter
surfaces filter on the order of 2 m3/min.35 The aver-
aging over a day is often of less interest than diurnal
variations at two or three hour intervals. Particles
have been collected for 1-3 hour periods on smaller
(10-20 cm^) filters of low contaminant content and
metals determined by sensitive methods such as INAA4?
NFAA, » ® XRF,5® and anodic stripping voltammetry
(ASV).51 Air sampling to gain particle size and trace
element concentration data can be carried out with
various kinds of impactors over various time spans.4>^>
34,50 A dichotomous sampler (0-2 um and 2-10 um) was
applied by Dzubay and Stevens to fractionate 13 elements
in urban aerosols.44
A metal sampling activity which bridges the atmo-
sphere and hydrosphere environments is rainfall collec-
tion and analysis. It would seem that distinguishing
between metal in the precipitation itself and that added
by dry fallout accumulations and mechanical devices
could be a problem, but I have not seen a great deal of
work in this area.
Of course, in considering air sampling of particu-
lates, one is obliged to mention the legendary problem
of copper contamination from electric machinery.
Storage
Storage of atmospheric samples (filters, impactor
surface films) out of contact with possible industrial
contaminants is an obvious requirement for low concen-
tration systems. Samples stored in liquids prior to
analysis need to be checked for losses or contamination
from vessels, etc. If particles are not dissolved be-
fore storage, adhesion to container walls must be antici-
pated. It is commonly observed that water samples lose
both particles and dissolved metals to surfaces of glass,
plastic, and Variously-coated containers used for
storage. Often, adsorption of cations is observed to be
more rapid and more extensive at neutral and alkaline
pH than at pH of the order of 1-3 (see, e.g., Tolg4^).
The EPA procedure for water sample storage entails addi-
tion of nitric acid.2-'- Addition of complexing agents,
e.g., EDTA, can also prove useful for preventing adsorp-
tive loss. However, leaching from container surfaces
must then also be considered. Durst and Duhart
observed (using a silver ion selective electrode) that
50/S of the initial (0.2 mg/1) silver ion was lost to FEP
Teflon from water over 46 days. Teflon had a slower
rate of uptake than did Pyrex and polyethylene. Recovery
studies for individual metals should be carried out so
that adsorptive losses and contamination effects can be
identified and brought under control.
Storage under natural pH or in acidified solutions
results in chemical speciation changes to different
degrees. For example, phytoplankton growth under natural
pH conditions of storage with illumination can lead to
uptake of metals, eventual release of soluble organic
matter, complexation, etc. These effects have been
noticed for Pb in seawater samples.33 Acidification
leads to dissociation of complexes, and perhaps of
greater practical significance, to desorption of adsorbed
ions from particle surfaces.I3,53. The pH effects of
adsorption-desorption on walls and system particles are,of
course, qualitatively similar. I tend more and more to
the view that a reliable total trace metal determination
and, where possible without contamination, an operation-
ally reproducible fractionation into dissolved and par-
ticulate is sufficient for most chemical interpretations
involving adsorption and metal-ligand complexes.
Concentration Procedures
For vapor forms of metal, absorption ,35,54 adsorp-
tion, and amalgamation (e.g., Hg with silver) are prac-
4
1-2
-------
tical concentration procedures. The 1975 "Air Pollu-
tion" review by Saltzman and Cuddeback55 contains a
number of recent references on metal vapor sampling.
Ross and Sievers^b applied chelation, extraction, gas
chromatographic separation, and electron capture detec-
tion to air analysis of beryllium. Talmi and Andren57
recently used chelation, toluene extraction, GC separa-
tion, and microwave emission spectrometric detection
for the determination of selenium in vapors, water, and
solid samples, with a detection limit of 40 picograms
(pg) •
For concentration of dissolved metal forms in
water, chelation and solvent extraction with dithlzone-
CCI4 or ammonium pyrolliditiedithiocarbamate-methyliso-
butylketone chelator-solvent systems have been prac-
tically employed by many workers. ^ »38;32 ,58,59
Burrell has gathered together many of the procedures
used for pre-concentration.30
Chelating resins (mainly Chelex 100) have been
used to concentrate dissolved metals from seawater and
river waters.60,61,31 Samples need to be filtered
first^3f for interpretation of recovery of portions of
total metals in suspended form is not possible. Should
individual metals be present to appreciable degrees in
stable and inert (slowly exchanging ligands) it is
difficult to assure complete recovery by this method.
Spiked metals will not test for this aspect of metal
recovery.53 I feel strongly that controlled experi-
mentation with well-defined model systems in the
laboratory is badly needed to gain insights into the
practical possibilities of this and similar concentra-
tion schemes.
Co-precipitation with Fe(011)3 has been used by
Young32 to separate Cr(III) species from Cr(VI) in
sewage effluent. A concentration procedure in which
Cu in seawater is co-precipitated with cobalt
pyrollidinedithiocarbamate and the organic-metals
precipitate then dissolved in MEK after filtration has
been used by Boyle and Edmond^ to determine Cu at
0.1 pg/kg levels in seawater. Flame AA is used for the
analysis step. It appears that the principle can be
extended to the co-precipitation of Pb, Zn, Cd, and Ag
as well as Cu, with NTAA as the final analysis step.*"
Analysis by XRF or IIXE once the metals precipitate
has been gathered on a membrane filter appears to be a
feasible method.6^
Lyophilization of natural water samples has been
employed by Harrison, et al.^ prior to INAA. Fifteen
elements were determined.
Losses and Contamination in Processing
The loss of metals such as Cd and Hg in dry ashing
of solids (particulates, sediments, biological samples)
is a danger in preparing samples for analysis by solu-
tion phase analysis (AAS, NFAA, ASV, colorimetry, etc.).
Important losses have been reported by various
workers;8,35,49 others have reported procedures to avoid
such losses&& (conversion to sulfate salts); Ranweiler
and Moyers6? dry-ashed atmospheric particulates on poly-
styrene hl-vol filters, followed by dissolution of the
residue in an acid digestion bomb, and reported no loss
of volatile metals. It seems that low temperature
ashing35 or wet ashing is to be preferred, unless it is
clearly shown that high temperature ashing does not
cause loss of the metals of interest.
Mercury and other metals may be lost as chlorides
if stored in HC1. Arsenic and antimony can be lost as
hydrides. Oxidizing acids are to be preferred; addi-
tional oxidant may be needed to prevent loss of redu-
cible elements (e.g., KMnO^ to prevent Hg loss).
An interesting tie-in of atmospheric metal contami-
nation and analysis of metal-containing samples in the
laboratory is afforded by the recent availability of
dry fallout rates for Pb in Los Angeles. Huntzicker,
et a_l. 3 • ^ determined a Pasadena area receptor deposi-
tion flux of approximately 45 ng/cm^/day. If that
applies to my laboratory it is clear that I must take
some care to protect against invasion of my samples
during processing, if my interest is in the 1-100 ng Pb
range. Reagents, glassware, and filters can become
contaminated easily unless precautions are taken, as
already noted. At the 5-50 ng Pb range, however, I
must look to reagent impurities, contaminations from
high-impurity labware, metallic lab equipment, and
possible contamination during sample collection opera-
tions if I find high and highly variable blanks. Know-
ing the sample metal concentration of interest in any
environmental sample, it is a reasonable objective to
reduce the procedural blanks in relation to the amount
of sample metal taken for analysis. It is clearly
undesirable to be determining an amount of metal of the
same magnitude as t^he procedural blank.
An example of differences between instrument-noise-
determined limits of detection and blank-determined
limits has been reported by Batley and Florence for
differential pulse anodic stripping voltammetry analysis
(DPASV) of Pb.6® The noise limit is 190 ng/1; the blank
limit is 10 ng/1. These data are interesting in view
of the report that Pb in surface seawater is in the
range of 40-80 ng/1.^3
Analysis Methods. Actual and Potential
"...the detection limit is a rather deceptive
butterfly which should be chased warily.
Chemistry, and in particular analytical
chemistry, is a practical science based on
quantification."
T. S. West69
As discussed in connection with Standard Methods,
the water analysts have settled on atomic absorption
and colorimetry in the arena of surveillance and spot-
check monitoring. There is reason to think (increased
sensitivity with no loss in selectivity) that non-flame
atomic absorption will become used more, particularly
for fresh waters where high salt backgrounds which pre-
vent direct injection of marine samples may not inter-
fere.
It remains to be seen whether anodic stripping
voltantmetry, pluse polarography, diJferential pulse ASV,
neutron activation analysis, x-ray emission methods,
multi-element atomic fluorescence, plasma emission
spectroscopy, and spark source mass spectrometry can
achieve wider and practical monitoring applications.
We have had direct experience with ASV, PP, and DPASV
in our laboratory and find these methods useful in re-
search and non-routine monitoring. They have been
tested in practical situations by a number of invest!-
gators|®>^ 1 >51,72,19,53 W},en blank precautions are
properly taken, ASV, PP, and DPASV are useful for look-
ing at low concentrations6®of Pb, Cd, Cu, Zn, Bi,
and a few other metals. The capital cost is not high,
but there must be a premium put on clean laboratory
conditions, pure reagents, and careful attention to
chemical details by knowledgeable personnel. And one
gets a limited number of metals.
For comprehensive comparisons of the capabilities
for water analysis of the various multi-element, it is
suggested that reviews by Minear and Murray^ Mancy,^®
and the LBL Environmental Monitoring Group'9 be
consulted.
Wary of detection limits as I have learned to be
5
1-2
-------
(vide supra), I have nonetheless ventured to have a
look at the suitability of several multi-element
methods for water analysis of my list of 13 interesting
metals (Be, V, Cr, Ni, Cu, Zn, As, Se, Ag, Cd, Ba, Hg,
and Pb). The methods are: INAA, PES (plasma emission
spectrometry), ASV, SSMS (spark source mass spectrome-
try) , XRF (and IIXE). In terms of capability of
analysis without prior concentration at the levels of
the drinking water standards or representative concen-
trations in f sh waters I tentatively identify the
following inadequacies:
Method Inadequate for Doubtful
INAA Be,Cr,Pb Zn,Se,Ba
PES Be, Se
ASV Be,V,Cr,Ni Eg
Se.Ba
SSMS Be,As,Se,Ag,
Cd,Ba,Hg
XRF Be Cd
NFAA Be
NFAA has been included for comparison, even though it
is an individual metal method. On this same basis of
evaluation, presently available selective ion electrode
methods would be suitable only for Ag, and would be
doubtful for Cd and Pb. Flame AAS is unsuitable with-
out preconcentration for all but Ag, Ba, and perhaps
Cd. We have used sensitivity and detection limit esti-
mates from ajnumber of sources.29»30,74,75 por sewage
effluents, sludges and sediments containing higher
levels of metals than those in water supplies and other
fresh waters, many of the multi-element methods may be
found suitable with no preconcentration. Detection
limits can be used to check such possibilities to a
first approximation.
Recent Analyses of Metals in Air Particulates
Surveillance of up to seventeen metals in the U.S.
has been by emission spectrograph of hi-vol filter
samples for some time, a semi-quantitative procedure
with inadequate detection limits for perhaps half of
the metals?2,76 Over the last five years three instru-
mental methods have been applied to aerosol metal
measurements with considerable success. The methods
are instrumental neutron activation analysis (INAA),
x-ray fluorescence (XRF), and atomic absorption, both
flame and non-flame (MS and NFAA). INAA and XRF are
both non-destructive, multi-element methods; AAS and
NFAA are destructive (atomization process) and indi-
vidual element methods. Not all metals of interest in
environmental monitoring are within the capability of
each of these techniques. For a comprehensive monitor-
ing program, two or more of these methods will need to
be used. In selecting a method it will thus be impor-
tant to define the monitoring objectives very carefully.
It is to be mentioned that the detection limits of INAA
and AAS vary rather widely for different metals,
whereas those of XRF (and IIXE) tend to be in the range
of roughly 1 to 100 ng/cm2 for the twelve or so metals
accessible to the method.
Some AAS and NFAA Applications: Matousek and
Brodie^B determined Pb in air particulates directly by
placing the collecting filter in a graphite cup. The
absolute sensitivity was 17 pg, requiring only small
volumes of filtered air to measure low levels of urban
Pb. Ranweiler and Moyere^? determined 22 metals
(including Al, Fe, Pb, Cu, Ti, Zn, Sr, Ni, V, Mn, Cr,
Bi, Co, Cs, and Be) by AAS with matrix-controlling
additions. Polystyrene filters were used to collect
2000 m-* of air with hi-vol samplers. Begnoche and
Risby149 used a carbon rod NFAA technique to analyze for
Al, Ca, Cd, Co, Cr, Fe, Mg, Mh, Ni, Pb, and Zn. Parti-
cles were dislodged from porous polymer filters in acid
medium by sonication. A method of standard additions
was used for calibration. The detection limits in terms
of ng metal per 47 mm diameter filter ranged from about
1 to 100.
Some INAA Applications: Dams, et^ al¦developed
and applied a method capable of determining 33 elements
in atmospheric particulates (Pb, Cd, and Be were not
included) with detection limits ranging from 1 to 1000
ng. Polystyrene, filters were used to collect samples.
Weslowski, et al.^ studied diurnal variations of a
number of metals (including Cd) by collecting particu-
lates on Whatman No. 41 filters over two-hour intervals
at a rate of 43.0 l/min. Gladney, et^alU? collected
atmospheric particulates with a six-stage cascade
impactor and determined 18 elements with INAA. "Small
particle distributions" were shown by V, Br, Se, Sb,
and Zn, while Al, Sc, and Fe showed "large particle
distributions." Ondov, et_ al..2^ used INAA to determine
3? elements in NBS SRM coal and 41 elements in NBS SRM
fly ash (with Pb in fly ash done by proton activation
analysis). Cd and Be were not done for either sample.
(Some of the unattractive features of the INAA technique
are candidly discussed. Of particular significance is
the need for nuclear facilities and well-trained
personnel in nuclear science.)
Some XRF Applications: Rhodes, et_ al.described
energy diBpersive XRF for air particulates over many
locations in Texas. Sixteen metals (and Br) were deter-
mined: Ca, Ti, V, Cr, Mn, Fe- Co, Ni, Cu, Zn, Hg, Pb,
As, Sr, Zn, and Mo. XRF and AAS results were compared
for Mn, Fe, Cu, Zn, and Pb. Glauque, et al.^O applied
XRF to analysis of California aerosols gathered on
filters, Lundgren-impactor stages, and impactor after
filters. Diurnal patterns at two-hour intervals were
obtained for Fe, Ca, Pb, Zn, Br, Ti, and Cu. Detection
limits for 13 (not including Cd) metals were in the
range 1-15 ng/m^ for aerosol deposits on mylar. This
study was a part of the larger project investigating
the chemical characteristics of the California aerosol,^
in which XRF has been used to determine 17 elements and
INAA 17. Both methods were used to determine 12 of the
metals. Dzubay and Stevens^ determined the following
metals following fractionation of urban aerosol with a
dichotomous sampler: K, Ca, Ti, V, Mn, Fe, Zn, As, Se,
and Pb. Detection limits were in the range 10-100
ng/cm2 for filters of mixed esters of cellulose.
Hammerle^ collected aerosols by filtration and
determined the following metals by XRF and INAA: Ca,
Ti, Cr, Mn, Fe, Ni, Cu, and Zn. Lead by XRF was com-
pared with AAS results. The aerosol masses collected
on small filters were from 0.6 to 2.2 mg. Results of
the two methods agreed to within 20%.
A Rough Comparison of Methods
I have made a rough comparison of what I think
might be practical detection limits for INAA, XRF, AAS,
and NFAA methods for metals in ai-r particulates
gathered on a small, low-blank filter, using low flow
rates and short sampling times. The metals included
are those which appear to have the greatest present
interest for air surveillance. For comparison purposes
10 cm2 filter areas were assumed, and 10 cm^ of the
final solution for submitting metals to AAS or NFAA
measurements. The mass of metal on the filter is c x Q
x t, where c is the air concentration of the metal, Q is
the air flow rate through the filter, and t is the samp-
ling interval (the same sort of comparison can apply to
impactors). For a low-volume flow rate of 55 l/min and
a time of 3 hours, the volume of air sampled is 10
6
1-2
-------
or 1 The following table indicates the capa-
bilities for metal aerosol analysis in a rough sense.
In some cases I was unable to find data or to learn
whether a particular metal could be done by a given
method. It is clear enough that NFAA is versatile
for metal analysis, but may not be sensitive enough for
some of the metals. The INAA method is also versatile,
but has widely different detection limits. AAS will
be useful for sources^ or for long-term hi-vol samp-
ling, not diurnal or particle-size investigations.
3*
Approximate Detection Limits, ng/m
Element
AAS
NFAA
XRF
INAA
Be
6
1
_
-
Ti
2000
10
2
3
V
1000
20
2
0.03
Cr
70
1
1
30
Mn
10
0.1
1
0.001
Fe
50
0.6
1
1000
Co
60
1
1
0.003
Ni
50
2
0.5
1
Cu
30
1
0.5
0.03
Zn
10
0.02
0.5
3
As
300
20
0.5
0.03
Se
600
20
0.5
-
Mo
300
8
-
3
Cd
5
0.02
-
1
Sn
1000
10
-
3
Sb
600
6
-
0.1
Ba
100
?
-
1
Pb
60
1
1
300
Bi
200
1
7
10
Pd
500
AO
?
-
Hg
-
20
1
-
*Q - 55 1/min, filter area ¦ 10 cm^, t « 3 hrs,
AAS and NFAA analysis volumes » 10 cnr*. NFAA
and AAS detection limits from Hwang?5 Burrell,30
or LBL.29 INAA limits from Hidy, e£ al • or
Dams, et al.^® XRF limits from Giauque, et
a!-56 ~
Concluding Remarks
I would like to attempt a last comparison of AAS,
NFAA, XRF, and INAA for the group of metals that appear
to have the greatest immediate multi-medium signifi-
cance in environmental monitoring. The following table
compares estimated practical minimum detectable amounts
of metals (assuming the same liquid analysis volumes
and filter areas as before, and drawing upon the same
sources of detection limit information as before).
Detection Limit Amounts, ng
Element
AAS
NFAA
XRF
INAA
Be
60
10
-
_
V
10,000
200
20
0.3
Cr
700
10
10
300
Ni
500
20
5
10
Cu
300
10
5
0.3
As
3000
200
5
0.3
Se
6000
200
5
-
Ag
500
1
-
0.3
Cd
50
0.2
1
10
Ba
1000
-
10
Hg
-
200
10
-
Pb
600
10
10
3000
It should be feasible to apply these techniques singly
or in combination to cover a wide range of environ-
mental monitoring tasks. Indeed, this is already being
demonstrated. One of the major considerations in
method selection is the through-put rate of samples,
and whether multi-element capabilities are required.
It seems that INAA is going to require large through-
put, and will be used at a limited number of central
regional laboratories. The same seems true, but to a
lesser degree perhaps, for XRF. The capital and opera-
ting costs are high. The attractiveness of AAS and
NFAA is the rather low cost and the flexibility of
operation if AAS and NFAA facilities are both available.
It was remarked not too long ago at an NBS panel
discussion on analytical problems in water pollution
control?*} that trace element technology is not a mature
science, and that an analyst must still be extraordi-
narily successful to measure impurities at these (low)
concentration levels. I think that not enough attention
has been devoted to blanks, sampling contamination, and
validation of methods through use of intralaboratory
and Intralaboratory comparisons. The methods of the
physicists will help to alleviate chemical difficulties,
but blanks will still be a problem. We do not lack the
analytical capability, in most cases, to carry out
element determinations for metals of interest; we have
not, however, been sufficiently thorough in evaluating
recoveries, identifying Interferences in the real
sample matrix, and using reference materials ("Gather
ye orchard leaves, fly ash, coal, bovine liver, and
. . . ?"80).
As we consolidate and validate elemental analysis
capabilities for all environmential media, we will need
to push hard in the area of chemical speciation. As an
example, it is not total Cu per se that governs the
toxicity of that element for phytoplankton. It is the
free and simpler labile forms." Bioassay, specific
ion electrode, and ASV techniques tend to support this
interpretation. But we still do not have a firm under-
standing of what kind of speciation information we can
get from ASV and related techniques. The reason, I
think, is that one cannot understand as complex a
system as seawater or lake water without looking at some
simpler chemical versions in parallel. The risk of
artifacts is great. As Hume and Carter®^ have pointed
out, "... the theoretical interpretation of anodic
stripping curves in terms of metal speciation has been
of an essentially qualitative nature. There has been
no convincing quantitative demonstration of the validity
of the approach, even with a known system containing a
single chelating species." Model experimental systems
will help us to break through in some cases, I believe.
Equilibrium models®3 ought to be used Iteratively with
special methods of analysis, such as specific ion elec-
trodes and ASV, to gain increased understanding of spe-
ciation. Bioassays should go together with such chemi-
cal work. Work already accomplished with organic forms
of Hg and Pb should serve as an example for further
developments of methods to get information about
molecular forms of metals and robust chelates.
The physical "speciation" of metals in water is
an unconquered field. One is inclined to be suspicious
about particle size-chemical element distributions
obtained by filters up to the present. Careful work
with physcially well-defined filters and screens is
needed to gain improved understanding of joint distri-
butions. More application can be made of the ultra-
centrifuge to distinguish between particulate and
dissolved forms in water. At the present time it seems
that only a more or less satisfactory fractionation
procedure for total vs dissolved forms of metals is
available to us. In this regard we have a long way to
go.
7
1-2
-------
et_ al_., Environ. Sci. Technol., 9,
References 35.
36.
1. Bertine, K. K. and E, D. Goldberg, SCIENCE, 173,
233 (1971). 37.
2. Dyrssen, D., et al., in* A Guide to Marine Pollution,
Goldberg,edT, Gordon and Breach, N.Y. (1972). 38.
3. Huntzicker, J. J., £t_ a1- Final report to the
Rockefeller Foundation, Grant RF-72049, August 1975,
Caltech, Pasadena. 39.
4. Huntzicker, J. J., eit al., Environ. Sci. Technol., 40.
9, 448 (1975).
5. Friedlander, S. K., Environ. Sci. Technol., 7, 235
(1973). 41.
6. Gordon, G. E. , et_ al_., in Trace Substances in
Environmental Health VII, Hemphill, ed., Univ. of 42.
Missouri, Columbia (1973).
7. Gladney, E. S., et al,, Environ. Sci. Technol., 8,
551 (1974). 43.
8. Bruland, K. W., Environ. Sci., Technol., 8, 425 44,
(1974).
9. So. Calif. Coastal Water Res. Proj. , The Ecology of 45,
the Southern California Bight, TR 104, March 1973, 46.
El Segundo.
10. Chen, K. Y., J. Water Poll. Contr. Fed., 46, 2663
(1974).
11. Morel, F. M.
756 (1975). 47.
12. Morgan, J. J. and T. H. Sibley, Proceedings Amer.
Soc. Civ. Engrs. Conference, Ocean Engineering,
Univ. of Delaware, 1975. 4g_
13. Stumm, W. and P. A. Brauner, in Chemical Ocean-
ography, 2nd ed., Riley and Skirrow, eds., Academic 49.
Press, N.Y. (1975).
14. Nationals Academy of Sciences, Principles for Evalua- 50.
ting Chemicals in the Environment, Washington, D.C.
(1975). 51.
15. Johnson, D. L. and Braman, R. S., Environ. Sci. 52.
Technol., 8, 1003 (1974).
16. Bisogni, J. J. and A. W. Lawrence, Environ. Sci. 53,
Technol., 8, 850 (1974).
17. Olson, K. W. and R. K. Skogerboe, Environ. Sci. 54
Technol., 9, 227 (1975).
18. National Academy of Sciences-National Academy of 55,
Engineering, Water Quality Criteria, Washington,
D.C., March 1973. 56.
19. Federal Register, 40, n. 51, Part II, March 14,1975.
20. Federal Register, 38, n. 199, Title 40, Ch. 1, sub- 57.
chapter D, October 16, 1973.
21. Methods for Chemical Analysis of Water and Wastes, 58.
EPA, Analytical Quality Control Laboratory, Cine.,
Ohio (1971). 59i
22. EPA, Air Quality Data for Metals, Office Air Quality,
Research Triangle Park, North Carolina (June 1973). 60.
23. Morgan, G. B., et_ al_., in Determination of Air
Quality, Mamantov and Shults, eds., Plenum,N.Y.(1972) 61.
24. Xntersociety Committee, Methods of Air Sampling and
Analysis, APHA, Washington, D.C. (1972). 62.
25. Lee, D. H. K., ed., Metallic Contaminants and 63.
Human Health, Academic Press, N.Y. (1972).
26. Israel, H. and G. Israel, Trace Elements in the 64.
Atmosphere, Ann Arbor Science, Ann Arbor(1974).
27. Ondov, J. M., et al., Anal. Chem., 47, 1102 (1975). 65.
28. Rhodes, J. R., et al., Environ. Sci. Technol., 6, 66
922 (1972).
29. Lawrence Berkeley Laboratory, Instrumentation for 67.
Environmental Monitoring, LBL-1, v. 2, Berkeley(1973)
30. Burrell, D. C., Atomic Spectrometric Analysis of 68.
Heavy Metal Pollutants in Water, Ann Arbor Science,
Ann Arbor (1974). 69.
31. Chester, R. and J. H. Stoner, Mar. Chem.,2,17 (1974).
32. So Calif. Coastal Water Res. Proj., Annual Report,
El Segundo, Calif. (June 1975), 7q,
33. Anon., Mar. Chem., 2, 69 (1974).
34. Hidy, F. G., et al., Final Report to Air Resources
Board, State of California, Science Center, Rock- 7j,
well International, Thousand Oaks, Calif., 72,
(September 1974).
8
Hwang, J. Y., Anal. Chem., 44, 20A (1972).
ASTM, 1973 Annual Book of Standards, Part 23,
Water; Atmospheric Analysis, Philadelphia (1973).
APHA, AWWA, and WPCF, Standard Methods . . ., 13th
ed., APHA, Washington, D.C. (1971).
EPA, Methods for Chemical Analysis of Water and
Wastes, Office of Technology Transfer, Washington,
D.C. (1974).
Lisk, D. J., SCIENCE, 184, 1137 (1974).
Altshuller, A. P., in NBS Spec. Pub. 351, Meinke
and Taylor, eds., National Bur. Stds., Washington,
D.C. (1972).
Tolg, G., in Comprehensive Analystical Chemistry,
III, Elsevier, N.Y., 1975.
Patterson, C. C. and D. M. Settle, in Accuracy in
Trace Analysis, LaFluer, ed., NBS Spec. Pub., in
press (1975).
Dams, R. , et_ al., Environ. Sci. Technol. ,6,441(1972).
Dzubay, T.G. and R. K. Stevens, Environ. Sci.
Technol., 9, 663 (1975).
Sheldon, R. W., Limnol. Oceanogr., 17, 494 (1972).
Cranston, R. E. and Buckley, Application and Per-
formance of Microfilters in Analyses of Suspended
Particulate Matter, unpublished manuscript, Bedford
Inst. Oceanogr., Dartmouth, Nova Scotia (BI-R-72-7)
(October 1972).
Weslowski, J. J., in Trace Elements in the Environ-
ment, Adv. Chem. Ser. n. 123, Amer. Chem. Soc.,
Washington, D.C. (1973).
Matousek, J. P. and K. G. Brodie, Anal. Chem., 45,
1606 (1973).
Begnoche, B. C. and T. H. Risby, Anal. Chem., 47,
1041 (1975).
Giauque, R. D. , et^ al_, , Environ. Sci. Technol., 8,
436 (1974).
Golovis, G. , et^ al_. , Anal. Chim. Acta, 64, 457 (1973).
Durst, R. A. and B. T. Duhart, Anal. Chem., 42,
1002 (1970).
Florence, T. M. and G, E. Batley, Talanta, 22, 201
(1975).
Coulson, D. M., e_t jd., SRI Project PYU-1374,
Report, Stanford Res. Inst., Menlo Park,Calif.(1973).
Saltzman, B. E. and J. E. Cuddeback, Anal. Chem.,
47, 1R (1975).
Ross, W. D. and R. E. Sievers, Environ. Sci.
Technol., 6, 155 (1972).
Talmi, Y. and A. W. Andren, Anal. Chem,, 46, 2122
(1974).
Kubota, J.
(1974).
Segar, D.
Anal. Chem,,
Riley, J. P.
479 (1968).
Galloway, J. N., Dissertation, Chemistry, Univ.
Calif. San Diego (1972).
Boyle, E. and J. M. Edmond, Nature, 253, 107 (1975).
Holm, T. R,, personal communication, Keck Labs,
Caltech (1975).
Brady, F. P. and T. A. Cahill, UC Davis Report
(UCD-CNL 166) April 1973.
Harrison, S. H., et al., Anal. Chem,, 47,1685(1975).
Kometani, T. Y., et al., Environ. Sci. Technol., 6,
617 (1972).
Ranweiler, L. E. and J. L. Moyers, Environ. Sci.
Technol., 8, 152 (1974).
Batley, G. E. and T. M. Florence, J. Electroanal.
Chem., 55, 23 (1974).
West, T. S., in Atomic Absorption Spectroscopy,
Dagnall and Kirkbright, eds., Butterworth,
London (1970).
Mancy, K. H., in NBS Spec. Pub. n, 351, Maink* and
Taylor, eds., National Bur, Stds., Waihington, D.C.
(1972).
Copeland, T. R., et al., Anal. ChMt., 45, 2171(1973).
Florence, T.M,,J. Electroanal.Ch«»., 35, 237 {1972)!
1«2
et_ al., Environ. Sci. Technol., 8, 243
A.
and J.
2, 291
and D.
L. Gilio, Intern. J. Environ.
(1973).
Taylor, Anal. Chim. Acta, 40,
-------
73. Minear, R. A. and B. B. Murray, in Trace Metals
and Metal-Organic Interactions in Natural Waters,
Singer, ed., Ann Arbor Science, Ann Arbor (1973).
74. Salbu, B., et al., Anal. Chem., 47, 1011 (1975).
75. Crocker, I. H. and W. F. Merritt, Water Research,
6, 285 (1972).
76. Gilfrich, J. V. , et al., Anal. Chem., 2002 (1973).
77. Frant, M. S., Environ. Sci. Techno1., 8, 224
(1974).
78. Dams, R. , et_ al. , Anal. Chem., 42, 861 (1970).
79. Hammerle, R. H., Anal. Chem., 45, 1939 (1973).
80. NBS, Standard Reference Materials, National Bur.
Stds., Washington, D.C.
81. Jackson, G. A. and J. J. Morgan, unpublished manu-
script. Keck Lab., Caltech (1975).
82. Hume, D. N. and J. N. Carter, Chlmia Analitczna,
17, 747 (1972).
83. Morel, F. M., et al., in Trace Metals and Metal-
Organic Interactions in Natural Waters, Singer,
ed., Ann Arbor Science, Ann Arbor (1973).
9
-------
METAL CONCENTRATIONS IN MOSS,LEAVES AND OTHER
INDICATORS OF METAL EXPOSURE IN THE ENVIRONMENT
Germund Tyler
Department of Plant Ecology
University of Lund
Ostra Vallgatan 14
S-22361 LUND,Sweden
Summary
The natural exchange and chelation capa-
city of mosses is used in assessing atmospheric
deposition of heavy metals.This method has been
used in Sweden since 1968 and is now being
utilized in several countries.The accumulating
power of moss tissue makes integrated measure-
ments also of minute levels of metal pollution
possible.Many carpet-forming mosses lack con-
tact with the soil and are wholly depending
on minerals supplied by wet and dry deposition.
In some species growth rate and age may easily
be determined,which allows a calculation of
deposited amount per unit of time.
Lichens may sometimes be used in the same
way as mosses,but the time-scale of the
measurements cannot usually be determined.
Another disadvantage is the high sensitivity
of lichens to air pollutants.Vascular plants
are used extensively in monitoring metal
pollution-and may be suitable in sites with
a high deposition rate.However,root sorption
of metals from the soil may be a serious
source of error,when current deposition data
are desired.The use of the metal accumulating
capacity and stratification of peat-bogs as
an indicator of historical changes in heavy
metal deposition rates seems less promising
than,e.g.,analysis of herbarium mosses or ice-
sheet strata.
Introduction
During the last decade considerable
interest has been devoted to atmospheric
pollution by heavy metals.The knowledge that
even small amounts of certain metals may be
harmful to organisms and the increasing wel-
fare of the industrialized nations created a
justified demand for a clean environment.
Certain heavy metals are emitted in increasing
amounts in technical and industrial processes.
Till now,measurements of atmospheric deposi-
tion of heavy metals over extensive areas
have been few,chiefly due to the lack of a
sufficiently sensitive and unexpensive
technique for simultaneous registration on a
large number of stations.
There are numerous sources of heavy metal
emission to the atmosphere.The natural back-
ground is probably very low but the level of
this natural background can only be a matter
of speculation,as every part of the globe has
been subjected.to deposition originating from
human activity/Natural sources are terrestrial
and cosmic dust and salt spray from the oceans.
Heavy metals are emitted in a great variety of
technical and industrial processes.Metal
smelteries and alloying plants,petrochemical
industry,fertilizer and sulphuric acid plants
are a few examples of polluting industry.
Refuse combustion may be of at least local
importance and fossil fuels in domestic heat-
ing and motor-traffic are major sources of
heavy metals in many inhabited parts of the
world.
The nature of the deposition is one of the
reasons why reliable direct measurements are
difficult to obtain.The amount of heavy metals
that reaches the ground per day or even per
year is certainly small in most areas.The
variability of,e.g.,meteorological conditions
leads to a corresponding variation in deposi-
tion rate,which makes continuous registration
for longer periods necessary.However,few
substances have as great a power to accumulate
in the environment as the heavy metals.Quite
a low deposition rate may therefore,in course
of time,give rise to considerable metal con-
centrations in certain components or organisms
of the ecosystems,high enough to be of bio-
logical significance.The main reason for the
accumulation is the property of most heavy
metal ions to form stable complexes or che-
lates with organic matter.All ecosystem com-
ponents, where negatively charged organic
groups are easily available and exposed to
the deposition may concentrate heavy metals.
The process may be passive(non-biological)
but active biological accumulation may
accelerate it.
It should therefore be possible to use
various components of the ecosystems as
natural integrators of heavy metal deposition.
The organic matter of the soil(humus)would be
an excellent component but it is usually not
possible to relate the metal concentration of
humus to a time-scale and the natural differ-
ences in decomposition rate between sites
make comparisons difficult or impossible.
Mosses as integrators
of heavy metal deposition
The natural exchange and chelation capa-
city of mosses may be used in assessing
atmospheric deposition of heavy metals.This
method was developed in Sweden,mainly at the
Department of Plant Ecology,University of
Lund,during the 1960's and is now in its
original or modified form utilized in several
countries,including the United States,the
United Kingdom,and th^ Scandinavian countries.
For further information,see references 5-13.
The stability ofi1 metal organic complexes
and chelates and the great cation exchange
capacity of the tissues are primary conditions
for the sorption of heavy metals by mosses.
In experimental studies on riioss tissues the
capacity for sorption of certain heavy metal
ions from dilute solutions has been demon-
strated .From these experiments it may be
concluded that lead and copper,supplied from
the atmosphere,are almost completely retained
by the moss carpets and several other heavy
metal ions will also be retained to a large
extent.The degree.of retention proved to
decrease in the order Cu>Pb>Ni>Co>Cd>Zn,Mn,
reflecting the stability of metal organic
chelates.In addition to passive sorption,
1
1-3
-------
living moss tissues have the capacity of an
active retention of metal ions.This is quite
decisive for the supply of potassium to the
moss tissues and certain evidence exists that
it is of importance also for zinc and mangan-
ese .
Carpet-forming(pleurocarp)mosses have no
particular organs for the uptake of minerals
from the ground.In a well-established carpet
there is no contact between the living parts
and the mineral soil.Growth and synthesis of
fresh organic matter takes place from the
continuously dying parts of the moss.As dis-
tinguished from vascular plants the 'leaves'
of mosses have neither epidermal layer nor
cuticle.They usually form a simple sheet of
parenchymatous cells,which are exposed to the
atmosphere.Salts and aerosols supplied in
precipitation,dry deposition and leaching or
wash-off of salts and dust particles from
trees,shrubs and litter are the sources of
those minerals,including heavy metals,which
may reach the moss carpets.
The use of carpet-forming mosses as
recording integrators of heavy metal deposi-
tion is further favoured by the mode of
growth in certain species.The very widespread
woodland moss Hylocomium splendens produces
one distinct 'segment' each year and,usually,
at least five 1 segments1 ( i-. e., f ive year's
growth) may be distinguished.lt is therefore
possible,within the limits of two to five
years,to choose the deposition time to be
integrated.
Hylocomium splendens was used in a depo-
sition study,covering most of Finland,Norway,
and Sweden_.The5re^ults have been published
elsewhere ' and only a brief survey
will be given here.Whole carpets,representing
the deposition of the last five-seven years
were used.To reduce the influence of eco-
logical variability on the results subsamples
were taken from ten randomly distributed
points in each site and combined for the
metal analysis.The vicinity of roads,buildings
and other signs of human activity was avoided
to reduce effects of local pollution.
The results have shown that the concen-
tration of lead in.. Hylocomium decreases from
100-120 ppm (nig-kg ,dry weight)in the south-
western parts to 5-10 ppm in the most
northern parts of Scandinavia.The lead depo-
sition in southern Sweden is almost twice as
high as the total emission of lead from
motor traffic in this province,not counting
the amounts deposited in the vicinity of the
roads.As the combustion of petrol is,by far,
the greatest source of lead emission to the
atmosphere in this area,a considerable amount
must be imported from surrounding areas,
mainly the densely inhabited and heavily
industrialized parts of Europe.The theory of
a long-distance transport of lead is further
supported by the findings of increasing con-
centrations in the glacier ice sheets of
Greenland and the Antarctic .
Next to lead the greatest regional depo-
sition differences have been measured in
cadmium,the concentrations decreasing from
0.80-1.00 ppm in the southwest to ca.0.10
ppm in the very north .A common feature of
all heavy metals analyzed is that the lowest
concentrations have been measured in northern
Scandinavia.However,the maxima are not always
found in the southwestern parts.Particularly
nickel and chromium show high concentrations
in an area extending almost throughout south-
central Sweden .A variety of metallurgic
industry is located in this area and,evident-
ly, internal sources make a more important
contribution than external sources to the
deposition of nickel and chromium,at least in
parts of Scandinavia.
No or only small differences in the con-
centrations of the common alkali(sodium,
potassium)and earth-alkali metals(magnesium,
calcium) were tjifsa^ured between various parts
of Scandinavia ' .These ions are supplied
in great amount with the precipitation and
the degree of passive retention is low.
Therefore,it will not be possible to use moss
as an integrator of the deposition of these
elements.
Through the analysis of herbarium mosses
collected in southern Sweden from about 1860
and onwards it was possible to demonstrate
the increase l^e^vy metal deposition during
the last century .The increase has been
particularly pronounced in lead.From about 20
ppm in the 1860's the lead concentration
rises to ca.50 ppm at the turn of the century.
It remains on this level to about 1950,after
which there is a rapid increase to ca.100 ppm
in 1968-1970.The rise during the 19th century
is related to a rapid industrial expansion, in
Europe,the rise after 1950 to the introduction
and increased consumption of lead petrol.
The carpet-forming woodland mosses are
suitable for large-scale surveys,particularly
in areas with a moderate or low population
density.Their preference for well-established
and rather stable ecosystems makes them diffi-,
cult to find in urbanized or heavily polluted
areas.The tree and stone growing specie^
Hypnum cupressiforme; has been used for local
surveys in cities agd.in areas with heavy
metal industry ' ' ' .An approximately
straight,sloping regression line is usually
obtained if the logarithm of the metal con-
centration of the moss is plotted against
the distance from a well-defined source of
emission.lt is usually possible to draw maps
showing a very regular deposition pattern.
The method with stone growing Hypnum has
proved useful for a rapid detection of point
sources in urban areas,where the location of
such sources was little known.It is also
easy to compare different cities and regions
as to the general degree of pollution(e.g.,
from fuel oil combustion by analyzing the
vanadium content of Hypnum samples).In heavily
polluted areas,where even Hypnum is lacking,
it is usually possible to find mosses of the
genera Bryum,Pohlia or Ceratodon with a
similar,though not identical sorption capa-
city as the carpet growing species.If several
species have to be used in the same program,
possible differences in the sorption capacity
and the growth rate must be compensated for.
An intercalibration study is now being started
in Sweden for this purpose.
A way of modifying the method has been
indicated by Goodman and Roberts .Mosses,
including substrate may be transplanted from
less polluted sites to areas where mosses are
lacking.After a certain time of exposure in
the new site the increase in the metal con-
centrations of the tissues is determined.lt .
should be possible to calculate the amount
deposited of several heavy metals from a
knowledge of the moss biomass and the initial
2
1-3
-------
and final metal concentrations.lt is advisable
not to make the time of exposure too long,as
adverse conditions of the new site finally
often kill the transplanted moss.It is true
that passive sorption continues in the dead
tissues but difficulties in distinguishing
'moss' from substrate may soon arise and
other sources of errcr may also become of
importance.The position of transplanted
samples,e.g.,distance to the ground,is of
great importance to the result,according to
recent evidence.If a measure of the true
deposition is desired,the moss sample should
be exposed on the ground.Some measure of the
metal concentration of the air is obtainable
in urban areas by utilizing,e.g.,the filter-
ing capacity of moss^in hanging net bags or
similar arrangements .For comparable results
rigorous standardization is necessary.
Lichens
Lichens - tree,stone or soil growing
double-organisms composed of alga and fungus
- have been used_in several urban areas,e.g.,
Stockholm,Sweden ,to estimate the degree of
air pollution.Lichens are very susceptible
to gases,including sulfur dioxide,and most
species are lacking in urban areas.The im-
provement of the air quality with the distance
from the city is accompanied by an increasing
number of surviving species.Certain lichens
may accumulate considerable quantities of
heavy metals without being killed,but due to
this high" susceptibility to polluting gases
they are often lacking in built-up areas.
For regional surveys of heavy metal deposi-
tion in woodland or mountain areas reindeer
lichens(Cladonia spp.)could be used,but the
time-scale can usually not be determined
because of the difficulty of separating the
biomass fractions produced during different
years.If a metal ion or radionuclide is
suddenly released in the environment,how-
ever, the deposition may be followed by re-
peated analysis of Cladonia carpets,as shown
in several excellent studies by radioecolo-
gists .
Vascular plants
Vascular plants have been much utilized
as integrators of heavy metal deposition,
particularly in estimating lead pollution
along rpads.Hundreds of papers have been
published in this_field since the study of
Cannon and Bowles .The literature has re-.,
cently been reviewed by the present author .
Though not equally effective as mosses in
sorbing the lead particles,a sufficiently
large share of the deposition is retained by
the above ground parts of vascular plants to
make them useful for this purpose.The 'back-
ground' concentration of lead is low in
leaves(usually 0.5-5 ppm),as very little is
transported from the roots in most species.
The 'background',however,is by no means con-
stant. It depends on species and kind of soil.
Moreover,there are differences due to age,
season,duration of lead exposure,etc.Per-
ennial , exposed organs are characterized by
higher lead concentrations than annual
tissues and perennial herbs and grasses
usually by higher concentrations than related
annual species.Trees and bushes may accumu-
late much lead in bark and twigs,along the
roads also in leaves(needles),but little is
contained in the wood.Many species concent-
rate lead in their roots but not in below-
ground tuberous organs.Species with a large
relative leaf area or long-persisting leaves
have the highest concentrations.There is also
a large-scale passive uptake of lead in dying
plant parts,in litter and in superficial soil,
where it is retained in a very little leach-
able form.
Vascular plants are usually not suitable
for detecting and surveying diffuse patterns
of metal deposition.Root sorption cannot
usually be distinguished from foliar uptake
and the 'background' concentration range may
be great.The possibility of particular bio-
accumulation in certain species must be con-
sidered. The 'background' concentration of
cadmium in roots of the wood anemone(Anemone
nemorosa)from southern Sweden is usually 10-
30 ppm in acid sites with low soil concent-
rations and only 0.8-1 ppm in the moss carpets.
The'background'concentration of lead in leaves
of vascular plants is almost constantly higher
in acid sites than in sites with a neutral
soil reaction.The examples could be multiplied.
However,in industrialized areas with a con-
siderable metal deposition,e.g.,in the United
Kingdom,vascular plants have been used success-
fully in surveying the degree of pollution .
For elements with a very low 'background'a
comparatively low degree of pollution may be
estimated by the analysis of vascular plants.
Tree leaves have even been used in estimating
mercury pollution of urban sites
Litter and soil
As mentioned above,litter(decomposing
organic material)and the organic matter of
the soil have a great capacity of retaining
heavy metals.In every site,where some humus
is present in the topsoil virtually every-
thing is trapped for a longer or shorter
period before it is finally lost from the
ecosystem,usually by leaching.Very little
is still known about the residence time of
various heavy metal ions in different soils.
The mean residence time of lead in an organic
topsoil is several decades,calculated as the
time necessary to reduce a given amount to
half its value.
A general feature of litter formation
is that the heavy metal concentrations(cal-
culated on a dry weight basis)usually in-
crease in two different ways.The loss of dry
matter(mainly carbon,oxygen and hydrogen)
automatically leads to a concentration in-
crease, as little or nothing is lost of those
metal elements which are sorbed most strongly.
The second2wav,is a real uptake of metals by
the litter ' .Due to the breakdown of organic
matter it is usually not possible to use
litter and soils as integrators of metal
deposition.The decomposition rate of similar
or identical litter material varies greatly
between sites.Therefore,coittparable samples
are difficult to obtain.The best way is to
use soil samples of known volume and relate
the metal content not to the dry weight of
the samples but to the volume or(even better)
the area covered by the sample.By repeated
sampling of the same sites it is possible to
detect at least a major increase of the metal
deposition.Because of the considerable amounts
of heavy metals present in the soil,either
naturally or as a consequence of diffuse
deposition,and of the difficulties in obtain-
ing comparable samples,the use of soils will
3
1-3
-------
probably never become an equally sensitive
and reliable method as the use of mosses.
Since many decades the use of peat-bogs
as historical documents has been a great
scientific field in quaternary geology.
Major trends in the development of the post-
glacial landscape have been elucidated through
pollen analytical studies and the peat strati-
graphy of each undisturbed bog reveals its
development.The inorganic chemistry of the
peat strata is considered to have been pre-
served to a certain extent,reflecting the
original mineral composition.Therefore,it has
been suggested that the vertical differences
in the heavy metal concentrations of the peat
could reflect changes in the deposition rate
during postglacial time.It is true that the
degree of humification may vary rather much
between different strata,but these difficulties
could be overcome by dating of the strata by
C-technique and determination of differences
in volume-to-weight ratios.However,the metal
accumulation pattern evidenced by ombrotrophic
bogs in southern Sweden makes the usefulness
of this tool rather doubtful.A very distinct
maximum in the concentrations of several
metals has been found in the peat of hummock
communities at a depth of 20-35 cm.The surface
concentrations are distinctly lower and the
maxima are not accompanied, by any correspond-
ing maximum in cation exchange capacity or
ash content of the peat.It must be concluded
that some kind of vertical transport is
possible in the peat,at least under acid,
reducing conditions.The accumulation surface
at about the level of the water-table is
probably a reduction-oxidation horizon,where
iron,lead,cadmium,copper,and zinc, and probably
also other elements,are precipitated.The
exact mechanism of this accumulation is
still unknown.
Conclusions
Biological materials as indicators of
metal exposure are valuable tools in environ-
mental pollution studies.Mosses as integrators
of heavy metals are often more suitable than
technical equipment in deposition studies,
particularly in areas with a moderate or low
deposition rate.For a rapid detection of
point sources in industrialized regions,where
the location and magnitude of the sources are
insufficiently known,the method of moss
analysis is more rapid,less expensive and
probably more reliable than other methods.
Biological monitoring is a major field
of research in environmental science.The
biological diversity is almost unlimited and
it will certainly be possible to find several
indicator organisms with unsuspected possi-
bilities .This survey was limited to certain
plants,litter and soil.The use of birds as
indicators and accumulators of environmental
mercury pollution is nowadays well-established
through the7work of several Scandinavian
scientists .A variety of animal groups is
now being examined in different parts of the
world with respect to possibility of metal
pollution monitors.lt is also certain that
future research will be able to present
several bacteria and fungi of great value in
biological monitoring.
References
1 Murozumi,M.,Chow,T.J.,and Patterson,C.1969.
Geochim.Cosmochim.Acta 33,1247.
2 Ruhling,A.,and Tyler,G.1968.Bot.Notiser
121,321.
3 Rtihling, A. ,and Tyler ,G .1969 . Ibid . 122, 248 .
4 Tyler,G.(1970)1971.2nd Intern.Clean Air
Congr.,Wash.,D.C.,Proc.,p.129.
5 Goodman,G.T.,and Roberts,T.M.1971,Nature
231,287.
6 Lee,J.A.1972.Nature 238,165.
7 Yeaple,D.S.1972.Nature 235,229.
8 Gorham,E.,and Tilton,D.L.1972.Bull.Ecol.
Soc.Amer.53,33.
9 Lee,J.A.,and Tallis,J.H.V.1973.Nature 245,
216.
10 Huckabee,J.W.A.197 3.Atmos.Environ.7,749.
11 Wallin,T.rand Jernel5v,A.1973.Rapport B
17 3,Inst.f.Vatten-och Luftvlrdsforsk-
ning,GSteborg,Sweden.
12 Keller ,T. 1-^74 .Schweiz .Zeitschr .f .Forstw.
125,719.
13 Little,P.,and Martin,M.H.1974.Environ.
Pollut.6,1.
14 Ruhling,A.,and Tyler,G,1970.Oikos 21,92.
15 Ruhling,A.,and Tyler,G.1971.J.Appl.Ecol.
8,497.
16 Tyler,G,197 2.Ambio 1,52.
17 Riihling,A. ,and Tyler,G.1973 .Water,Airfoil
Pollut.2,445.
18 Ruhling,A.1971.Modern Kemi 11,32.
19 Ruhling,A.1971.Sveriges Natur 5,226.
20 Skye,E.1968.Acta Phytog.Suec.52(123 pp.)
21 Persson,B.1970.Tellus 22,564.(References
to related studies are given in this
paper).
22 Cannon,H.L.,and Bowles,J.M.1962.Science
137,765.
23 Knutsson,G. ,Backman,L. ,Hedgren,S.,Rtihling,
A.,and Tyler,G.1974.National Swedish
Environment Protection Board P M 476.
24 Smith,W.H.1972.Science 176,1237.
25 Tyler,G.1971.Oikos 22,265.
26 Nilsson,I.1972.Oikos 23,132.
27 Jensen,S.,Johnels,A.,01sson,M.,and Wester-
mark,T. (1970)1972.Proc.15th Intern.
Ornithol.Congr.,the Hague,p.455.
4
1-3
-------
METAL CONCENTRATIONS IN BLOOD, URINE, HAIR
AND OTHER TISSUES AS INDICATORS OF METAL ACCUMULATION IN THE BODY.
Thomas W. Clarkson
Department of Radiation Biology & Biophysics
University of Rochester School of Medicine
Rochester, New York 14642
Summary
The parameters relating absorption, accumulation,
and excretion in man (the metabolic model) are best un-
derstood for methylmercury. Concentrations in blood
and hair are reliable indicators of body burdens in in-
dividuals in steady state and undergoing changes in
body burden. The hair sample also allows recapitula-
tion of past exposure. The metabolic model for mercury
vapor suggests that urinary excretion is a useful indi-
cator of body burden but more tracer data are needed on
man. The metabolic model for lead is more complex in-
dicating the presence -of "mobile" and "stable" frac-
tions of lead in the body. Recent evidence indicates
that the logarithm of the mobile fraction is propor-
tional to urinary concentration. Cadmium is distin-
guished by having a long biological half-life in man in
the range of 10 to 30 years. Renal accumulation ap-
pears to continue throughout the first 50 years of life.
Urinary concentrations may indicate, on a group basis,
the concentration of cadmium in the major organs of ac-
cumulation. (liver and kidneys) providing that renal
damage has not ensued.
General Principles
Tlie accumulation of metals in the body cannot be
directly measured during the lifetime of an individual.
The sole exception is Y-ray emitting isotopes of met-
als where the body burden may be estimated by whole
body counting. Thus it is customary to make us of met-
al determinations in so-called indicator media such as
blood, urine, and hair. The validity of this approach
depends upon how accurately the metabolic model for
each metal is established. Specifically we need to
know the metabolic parameters that allow a quantitative
description of the processes of absorption, distribu-
tion, accumulation, and excretion.
A schematic and simplified model relating the
processes of absorption, accumulation in and loss from
the body is depicted in figure 1.
DAILY INTAKE
a
ACCUMULATED
AMOUNT
A
LOSS
b.A
A » (a/b) (l-exp(-b.t)
A ¦ a/b
CD
(2)
Figure 1. Single compartment model of uptake accumula-
tion and excretion assuming linear kinetics,
A constant average daily amount absorbed (ja)re-
sults in the accumulation of an amount in the body
(body burden) at any time Jt. The rate of elimination
from the body is assumed to be proportional to the body
burden, where b_ is the elimination constant. The net
accumulation is related to the time of exposure by
equation (1), The rate of excretion will increase with
increasing body burdens until daily intake and excre-
tion balance. The body burden has then assumed a final
steady state value, , the value of which ia deter-
mined by the daily intake and elimination constant as
described in equation (2). The elimination half-time,
T, is related to the elimination constant _b by equa-
tion (3).
T - In 2/b (3)
The same equations may be used to describe the accumu-
lation of metal in compartments in the body where, a,
in this case is the fraction of the daily dose going to
chat compartment. The mathematical expressions based
on this model were developed by the Task Group on Met-
al Accumulation1 who also discussed the limitations of
this model.
A. Metabolic Model of Methylmercury
Studies on volunteers2' 3 given a single tracer
dose of methylmercury indicate that this model may be
valid for this form of mercury. Some of the key param-
eters for the metabolic model of methylmercury for the
"standard 70 kg man" are listed in Table 1.
Table 1, Metabolic Model for Methylmercury
Absorption efficiency
Intake to 1 liter blood
Elimination from whole body
Clearance from blood
Cone, in hair/conc. in blood
95% oral intake
1% oral intake
50-93 days half-time
45-105 days half-time
250
The tracer studies revealed that 95% or more of
the oral dose was absorbed, that approximately 1% of
the daily intake was found in one liter of whole blood3
and that the elimination half-time for the whole body
fell in the range of 52 to 93 days. The half-time of
clearance from blood averaged 50 days in tracer stud-
ies but field studies'*<5 of individuals exposed to
methylmercury in their diet revealed a wide range of
half-times from 45-105 days but one person had a clear-
ance half-time from red blood cells of 164 days.
The validity of this model may be tested by exam-
ination of data relating average daily intake to steady
state blood concentrations in Individuals exposed to
methylmercury in their diet. The linear relationship
predicted by equation (2) has been confirmed by several
studies of fish eating populations. These results are
summarized in Table 2 where the least squares linear
regression relationship is given for steady state blood
levels y ng Hg/ml and the average daily intakes, x,
Ug Hg/70 kg. Th.e tracer parameters quoted in Table 1
would predict that the coefficient of x is unity - for
convenience the elimination half-time from the whole
body and the clearance half-time from whole blood was
assumed to be 69 days. The field studies indicate the
coefficient of x to be between 0.5 to 0.7, except for
one study yielding the low value of 0.3. The latter
value has been attributed to errors in estimation of
dietary intake8. Given the difficulties in making ac-
curate studies in exposed populations, the agreement
seems reasonable between the relationship calculated
from the metabolic ncdel and the observed relationship
in people exposed to methylmercury from fish.
1-5
-------
Taole 2. Steady State blood Level (y, ng Hg/ml) and
Average Daily Intake (x, Pg Hg/day/70kg)in Fish
Eating Populations
Number Linear Regression Reference ¦#
15 y = l.Qx 3
32 y = 0.7x + 1 4
165 y = 0.3x +5 5
20 y = 0.8x + 1 6
22 y = 0.5x + 10 7
Considerable evidence now exists that the concen-
tration of mercury in hair is a reliable indicator of
the mercury concentration in blood in people exposed to
methylmercury from fish. Table 3 reports the linear
regression relationship between the mercury concentra-
tion in hair (y pg/g) in people who experienced long-
term daily intake of methylmercury from fish. The co-
efficient of x indicates that the average concentration
ratio of hair to blood was between 230 and 280. The
ratio also appears to be constant in individuals under-
going changing blood levels10.
These limited metabolic data in man may be com-
pared with the extensive data on industrial workers ex-
posed to elemental mercury vapor. The metabolic param~
eters quoted in Table 4 would indicate that a steady
state body burden should be attained after about one
year's exposure (five times the whole body elimination
half-time). Therefore, urinary excretion should acr
count for 25 to 33% of the daily amount retained in the
lung. Assuming five working days per week, a daily
ventilation at work of 10m3 and the retention of 75%
of the inhaled amount, workers exposed to a time
weighted average air concentration of 0.1 mg Hg/m^
should have an average urinary concentration of 134 to
177 pg/liter assuming an average 24 hour urinary volume
of one liter. A study by Smith £t al.15 of 642 workers
exposed to elemental mercury in the chloralkali indus-
try indicated a linear relationship between average
urinary concentrations and average time weighted ex-
posures and that workers exposed to 0.1 mg Hg/m^ had an
average urinary concentration of 220 yg Hg/liter.
These results suggest that further tracer studies on
volunteers would'be useful in predicting the outcome of
long-term industrial, and environmental exposures.
Table 3, Relationship Between Concentration of Mercury
in Hair (y ug/g) and Whole Blood (x Mg/g)
No. of
Subjects
51
12
60
Linear Regression
y = 230x + 7
y = 280x - 1
y - 230x - 4
Reference #
9
6
7
Studies of the recent Iraq outbreak of methylmer-
cury indicated that concentration of mercury in human
milk closely correlates with maternal blood concentra-
tions but urinary concentrations do not correlate with
those in blood11.
In summary, both tracer studies in volunteers and
observations on individuals having high exposures to
.methylmercury Indicate that the mercury concentration
in blood and hair is a reliable indicator of body bur-
den. The hair sample has the advantage of allowing re-
capitulation of past exposure.
B. Metabolic Model of Mercury Vapor
Pulmonary retention of inhaled vapor in man is ap-
proximately 75% consistent; with complete absorption
from alveolar air Elimination from the whole
body has been estimated in two individuals who acciden-
tally inhaled radioactive vapor over a period of a few
hours15. The whole body elimination half-times were 26
and 28 days observed over a post-exposure period of 60
days. A study of five volunteers inhaling radioactive
vapor for approximately 20 minutes revealed a mean
half-time of 58 days (range 35-90 days).
Urinary excretion was found to account for between
25% and 33% of total excretion 7 to 19 days after ac-
cidental exposure of the two persons referred to
above15. These findings are supported by unpublished
observations by Cherian et: al. on five volunteers in-
haling radioactive vapor. In this study, urinary ex-
cretion accounted for 25% of total excretion. The ob-
servations on pulmonary retention, biological half-
times and urinary excretion are summarized in Table 4.
Table 4. Metabolic Model for Mercury Vapor
Absorption efficiency
Elimination from whole body
Urinary excretion
75% Inhaled amount
58 Days half-time
25-33% Total excretion
C. Metabolic Model of Lead
The metabolic model for lead in man has recently
1 17
Gas t roInt es tInal
been reviewed by expert groups
absorption of lead is between 5-15% based on balance
studies. Data on three volunteers consuming a single
oral dose of 212Pb indicate wide individual variation
i.3% to 16%1®. Factors influencing the efficiency
of gastrointestinal absorption have been reviewed
elsewhere17. Pulmonary retention of a 212Pb aerosol
by volunteers was in the range of 14 to 45%^9. The
importance of the physical state of lead in air is dem-
onstrated by Booker et _al.2" who found 99% retention of
inhaled lead vapor versus 34 to 60% retention of in-
haled lead attached to sub-micron particles,
Rabinowitz et al,21 have used stable isotopes as
tracers of lead in man. Their results, quoted in Table
5, indicate a three compartment model. Lead in blood
and possibly in an additional soft tissue compartment
has a mean half-time of 19 days. Lead in soft tissue
and in. a rapidly exchanging compartment in bone has
a half-time of 21 days. Lead in the stable fraction of
bone has a half-time of about 21 years. The long half-
time in bone confirms previous conclusions1.
Table 5. Metabolic Model for Lead
Absorption efficiency
Absorption efficiency
Blood
Soft tissue
Bone
5-152 oral intake
14-45% inhaled amount
19 Days half-time
21 Days half-time
21 years
The widely different biological half-times in
various compartments of the body complicates any at-
tempts to relate concentrations in indicator media to
body burden of lead and to concentrations of lead in
brain, bone marrow, and other critical tissues and or-
gans. Empirical approaches have been tried to relate
blood levels to "exchangeable" lead in the body.
Chisholm et al.w estimated mobile tissue lead by the
measurement of the quantity of lead excreted Jti urine
in the 24 hour period following administration of cal-
cium EDTA in a dose of 25 mg/kg to children. The mo-
bile lead increased exponentially as the blood lead
increased linearly in the range of 8-75 yg Pb/100 g
blood.
2
1-5
-------
D. Metabolic Model of Cadmium
Data on animals and on volunteers who Ingested a
single dose of '^"cd (for detailed review, see Friberg
et al.23) indicate that an average of approximately 6%
of an oral dcrse of cadmium is absorbed and that indi-
vidual variation probably lies in the range of 3 to 10%
depending on such factors as calcium and protein con-
tent of the diet. No data are available for pulmonary
retention of cadmium in man. Friberg et al." con-
cluded that retention would be in the range, of 10 to
50% based on general principles of aerosol deposition
in lung.
The biological half-time of cadmium in man is much
greater that that for mercury and is of the order of
half-times reported for the most stable compartment of
lead. Because of the slow excretion of cadmium, tracer
studies in volunteers can only give a lower limit to
the value of the biological half-time. For example,
the data of Rahola _al.2i+would give a limit of 130
days. Friberg et al.23 estimated the biological half-
time to between 13 to 47 years by relating average
urinary excretion to body burdens based on autopsy
data. Other estimates of biological half-life fall in
the range of 10-30 years2
A review of autopsy material from a normal popula-
tion at the age of 50 years suggests that approximate-
ly one third of ttie body burden is found in the kidneys,
Th.e metabolic parameters are given in Table 6.
Table 6. Metabolic Model for Cadmium
Absorption efficiency 3-10% oral intake
Absorption efficiency 5-50% inhaled amount
Elimination from whole body 10-30 years half-life
Fraction kidney deposition 33% body burden
It has been calculated2^ that renal accumulation of
cadmium will attain a value of about 50 ug/g at the
age of 50 years in people having a daily intake of cad-
mium of 24 ng/calorie (equivalent to average environ-
mental exposure) assuming that 5% was absorbed daily
and one-third is deposited in the kidneys.
An expert grouplhas noted that renal effects of
cadmium may become detectable at kidney cortex concen-
trations as low as 200 »g Cd/g wet weight.
The relationship of cadmium concentrations in
blood and urine to the body burden and organ concentra-
tion is not well established. Evidence has been re-
cently reviewed indicating that urinary concentrations,
on a group basis, may correlate with cadmium concentra-
tions in the main storage organs (liver and kidneys)
provided that renal damage is absent and that blood
concentrations probably reflect recent exposure.
References
6. Tejning, S. (1969) Report 68 05 29 From Dept.
Occup. Med. University Hospital, Lund.
7. Skerfving, S. (1974) Toxicology 2 2-23
8. Swedish Expert Group (1971) Nord.Hyg. Tidskr. Supple.
4
9. Tejning, S. (1967) Report 67 08 31 From Dept.
Occup, tied. University Hospital, Lutid,
10. Amin-Zaki et al. (1975) Amer. J. Dis. Child
(In press)
11. Bakir, F. et al. (1973) Science 181 230-241
12. Nielsen-Kudsk, F. (1965) Acta Pharmacol. 23
250-262
13. Teisinger, J. and V. Fiserova-Bergerova (1965)
Indus. Med. Surg. 34 580-584
14. Hursh, J. et al. (1975) Arch. Environ. Health
(In press)
15. Brown, K. W. et al. (1975) Health Physics ^28 1-4
16. Smith, R.G. et al. (1970) Amer. Ind. Hyg. Assoc.
J. 687-700.
17. Nordberg, G. (1974) Proceedings from an internar
tional meeting organized by the subcommittee on
the toxiciology of metals. Tokyo, Nov. 18-23,
1974.
18. Hursh, J. and Suomela (1967) Acta. Radiol.
(Scand) I 108-120
19. Hursh, J.B. et al. (1969) Health Physics 16
257-267
20. Booker, D.V. et al. (1969) Brit. J, Radiol. 42
457-466
21. Rabinowitz, M.B, et al. (1973) Science 182
725-727
22. Chisholm, J.J. Jr. et al. (1975) Quoted by
reference 1.
23. Friberg, L. et al ( 1974) "Cadmium in the Environ-
ment" 2nd edition CRC Press, Cleveland, Ohio
24. Rahola, T. et al. (1972) In "Assessment of Radio-
active Contamination in Man". l.k.Z.k., Vienna
1972
1. Task group on metal accumulation (1973) Environ.
Physiol Bio'chem. _3 65-107
2.Aberg, B. et al. (1969) Arch. Environ. Health 19
478-484 ~~
3. Miettinen, J.K. (1973) In:"Mercury, Mercurials and
Mercaptans"(M.W. Miller and T.W.Clarkson eds) pp
233-243 Charles C. Thomas, Publisher, Spring£ieldv
.Illinois
4. Birke, G. et al (1967) Arch Environ. Health ^5 77
5. Tejning, S, (1967) Report 67 02 06 From Dept,
Occup. Med., University Hospital, Lund
3
1-5
-------
THE ROLE OF EPIDEMIOLOGY IN
ASSESSING THE HEALTH EFFECTS OF METALS
1 J 7
Kenneth Bridbord , Joseph K. Wagoner , Hector P. Blejer ,
Philip J. Landrigan^, and Richard A. Lemen^
Introduction
The process of assessing the health effects of
human exposure to metals, or any other substance, often
requires consideration of data available from a number
of study disciplines. In this regard, one must consi-
der the health effects data available from a variety of
disciplines including results from epidemiologic, clin-
ical, industrial hygiene and animal toxicologic re-
search. Whereas each of these study approaches offers
certain advantages, each also has certain limitations.
Consequently, exclusive reliance upon any single ap-
proach does not provide all or even most of the needed
answers to the many occupational and environmental
health questions which are raised.
As no single study approach can provide all or
even most of the needed health information, so must a
specific approach, such as epidemiology, consider the
results of studies involving a number of populations
exposed to common materials. Such is frequently the
case where workers are exposed to chemicals inside a
plant and where residents of communities contiguous to
the plant are similarly exposed, though usually to a
lesser degree. In such instances, studies of workers
in the plant can frequently provide valuable clues as
to effects upon residents in adjacent communities.
That is, occupational health data can be considered the
sentinels of the effects of chemicals among the general
population. Conversely, studies of the general popula-
tion can at times provide clues as to health effects
among workers exposed to similar substances.
One of the well recognized advantages of the epi-
demiologic approach is that it offers the potential to
determine effects o£ long-term, low-level exposures to
chemicals in man under actual conditions of exposure,
i.e., as they occur in the "real world." However, one
of the major problems in epidemiologic research, either
in the workplace or the community, often involves the
difficulty in quantifying present exposures and, more
frequently, the even greater difficulty in adequately
determining previous exposures. For this reason, it is
usually not possible to obtain accurate dose-response
information of pollutant effects using epidemiology
alone, particularly when one is concerned about effects
which may develop many years following onset of expo-
sure. In contrast, whereas more accurate dose-response
information can be derived from toxicologic studies
using experimental animals, an investigator must con-
National Institute for Occupational Safety and Health
5600 Fishers Lane
Rockville, Maryland 20852
2
National Institute for Occupational Safety and Health
U.S. Post Office Building
Fifth and Main Streets
Cincinnati, Ohio 45202
3
Center for Disease Control
1600 Clifton Road, N.E.
Atlanta Georgia 30333
sider the question of how to extrapolate results from
experimental animals to man and how to extend the re-
sults from observed higher dose levels to lower dose
levels. Clinical studies, which involve exposure of
volunteers for brief periods to known concentrations of
single or simple mixtures of pollutants, are limited by
the toxicity of the materials which can be administered
and by the fact that results from only a few individ-
uals must be extrapolated to the entire population.
Clinical studies are not very useful for determining
chronic effects which frequently result from long-term
low-level exposures.
However, when all three study approaches are uti-
lized in a coordinated way, the benefits of each indi-
vidual approach may frequently be maintained and many
of the methodological weaknesses of a given approach
may be overcome.
The remainder of this paper illustrates the above
points and discusses (1) the usefulness of toxicologic
data in predicting long-term effects of chemicals in
man, in turn providing clues for epidemiologic studies,
(2) the usefulness of occupational studies as sentinels
of toxic effects among the general population, and (3)
the potential of community studies to provide clues as
to effects of chemicals upon workers.
The value of animal studies in predicting a cancer
risk for man has been corroborated by numerous examples
in which epidemiologic studies in man have confirmed
the results observed in animals. Illustrations of such
corroboration include vinyl chloride, B-naphthylamine,
benzidine, 4 aminodiphenyl, and bis (chloromethyl)
ether. There are, of course a large number of chemicals
which have been shown to be carcinogenic in animals but
for which epidemiologic evidence for cancer in humans is
generally weak or nonexistent. In these instances, one
must carefully examine available epidemiologic studies
to determine whether they were sufficiently well de-
signed to detect an increased cancer risk and whether
their data have been properly evaluated or given full
consider ation.
Beryllium
Beryllium, for example, is a substance which has
shown positive responses for cancer in multiple species
of experimental animals. In fact, few substances give
so consistent a carcinogenic response in so many animal
species as beryllium. Additionally, numerous studies
have repeaterly shown beryllium to be carcinogenic by
most routes of .'-iminlstration, i.e., inhalation; intra-
venous, intraperitoneal, subcutaneous, and intratra-
cheal.
As early as 1946, Gardner and Heslington'' reported
the induction of osteosarcoma in rabbits following in-
travenous injection of zinc beryllium silicate or beryl-
lium oxide. The intravenous administration of zinc sili-
cate, zinc oxide, and silicitic acid to rabbits, however,
failed to elicit this carcinogenic response. Over the
years osteosarcoma has been induced in rabbits by these
same and other beryllium compounds using intravenous or
intramedullary administration. In 1951, Dutra et al.^
reported the induction of osteosarcoma in one of six
rabbits following inhalation of beryllium oxide dust.
1-6
-------
Vorwald in 1966 reported that rats developed pul-
monary neoplasia following inhalation of beryllium sul-
fate. Similar findings have also been observed in rats
by other investigators using this same and other com-
pounds of beryllium As recently as 1972, Groth
reported the induction of pulmonary cancer in rats in-
jected intratracheally with several beryllium compounds.
Beryllium has also been shown to be carcinogenic
to primates '®. Vorwald^, in 1966, reported the induc-
tion of pulmonary cancer in monkeys exposed by intra-
bronchial and/or bronchopleural implantation of beryl-
lium oxide and by inhalation of beryllium sulfate.
As a result of this consistent evidence of carcin-
ogenic effects caused by beryllium compounds in animals
and of the generally assumed negative, though admitted-
ly limited epidemiologic evidence to date, beryllium
has often been cited as an example of a compound which
is carcinogenic only in experimental animals. Reexami-
nation of the available epidemiologic evidence, however,
raises questions on concluding that beryllium is not a
carcinogen for man.
For example, a study by Bayliss® which observed
that excess respiratory tract cancers did not occur a-
mong 4,000 workers exposed to beryllium might lead one
at first to negate the possible role of beryllium in
human cancer. However, the majority of workers in this
study had not been*followed more than 15 years since
onset of exposure to beryllium, a latency period which
is not long enough to demonstrate a related risk of
lung cancer.
Hardy et. al.^® in 1967 reported on the complica-
tions of beryllium disease among 535 individuals entered
into the Beryllium Case Registry during 1952-1966. Of
14 subjects known to have subsequently developed cancer
among this group, three were diagnosed as having primary
cancer of the bone. The observation of three such cases
was striking since the International Agency for Research
on Cancer of the World Health Organization^-1-, has estima-
ted that no more than 0.1 case of bone cancer would have
been expected to occur among these 535 individuals.
12
In 1970, Mancuso reported on the relation of dura-
tion of beryllium exposure and beryllium-induced bron-
chitis and pneumonitis to the etiology of lung cancer
among beryllium production workers. Among 142 cases of
beryllium-related bronchitis and pneumonitis identified
during 1940-1948, six deaths due to lung cancer subse-
quently occurred for an age-adjusted lung cancer mortal-
ity rate of 284.3 per 100,000 population. This rate con-
trasted witb 77.7 for all white males employed in the same
beryllium production facility during 1937-1948, and led
Mancuso to conclude that "prior chemical respiratory ill-
ness does influence the subsequent development of lung
cancer among beryllium workers."
These observations constitute a continuum, starting
with experimental bioassys for carcinogenesis, through
epidemiological studies in man which is highly consistent
and supportive of the position that beryllium is not only
carcinogenic to animals but can be carcinogenic in man as
well. Although additional evidence is certainly needed
before any definitive conclusions can be made, it is evi-
dent that had epidemiologists and clinicians followed
through sooner on the strongly suggestive body of evi-
dence from experimental animals, the question of beryl-
lium carcinogenicity for man might already be fully re-
solved.
Molybdenum
The role that molybdenum may play in gout offers
a second example of animal studies which have provided
valuable clues to epidemiologists. A number of these
studies has shown molybdenum to affect the xanthine ox-
idase-uric acid metabolic pathway. By comparison, an
epidemiologic study from the Russian province of Anka-
van, where dietary molybdenum intake ranges from 10-15
mg per day due to high natural levels, has shown ele-
vated blood xanthine oxidase activity and increased
uric acid concentrations in the blood compared to con-
trols living in an area with lower molybdenum expo-
sure-'- . A high percentage (31%) of the examined pop-
ulation from the Ankavan province was diagnosed to
have a gout-like disease. These animal data combined
with the community study provide substantive evidence
for the need to evaluate workers exposed to molybdenum
for a possible increased prevalence of gout. This
is an example not only of toxicologic data providing
clues for epidemiologic research, but also of a comr-
munity study providing clues as to a possible occupa-
tional health problem.
Arsenic
A recent example of complementary epidemiologic
studies of both workers and residents in contiguous pop-
ulations exposed to industrial emissions invo.lves inor-
ganic arsenic.
14
As reviewed by Blejer and Wagoner , nine out of
eleven epidemiologic studies reported from 1948 to
1975 have shown, initially or upon reassessment, signi-
ficant excess mortality from respiratory cancer among
diverse occupations exposed to various inorganic arsen-
icals. Two of the nine studies showed concomitant, sig-
nificant excess mortality from lymphatic cancer; and
another, from skin cancer. Additionally, two of these
studies revealed a dose-response relationship between
arsenical exposure and lung cancer. In the first, re-
ported in 1969, the relationship was semi-quantitative
with a possible interactive role by sulfur dioxide or
other contaminants^. The other demonstrated an in-
creased lung cancer mortality risk , apparently at ar-
senic concentrations above 1 Vg/m^, when the total dosage
was recalculated as the 8-hour time weighted average
daily exposure over a 40-year working life-^. However,
these and related data did not reveal a definite no-
effect exposure level. Arsenic is an example where an
animal model for lung cancer has not yet been found.
A recent study by Blot and Fraumeni''"^ of residents
of counties with copper, lead, or zinc smelting and
refining industries, but not of counties where other non-
ferrous ores are processed, showed significantly in-
creased lung cancer rates for white males and females
based upon data compiled from 1950 to 1969. These find-
ings could not be attributed to differences in geograph-
ic region, population density, urbanization, socioeco-
nomic status, or other manufacturing process. The au-
thors concluded that exposure of residents to community
air contaminated with industrial emissions containing
inorganic arsenic may have contributed to these excesses
of mortality from lung cancer. A study by Milham-'-®
showing excess arsenic absorption among children resi-
ding near a smelter in Tacoma, Washington, provides
additional bases for concern.
Lead
Lead has been shown to produce kidney cancer when
administered to experimental animals-'--'-. Data from the
U.S. Cancer Mortality by County from 1950-1969 show
that the rate of kidney cancer among males, but not
females in Shoshone County, Idaho, site of a lead, zinc,
and cadmium smelter, is elevated (100%) compared to the
state of Idaho-*-®. The recent data Bhowing increased
absorption of lead among residents near this same smel-
ter are also relevant. Landrigan, et al.^®, have dem~
2
1-6
-------
onstrated that 99% of 1-9-year-old children living with-
in 1.6km of the smelter had blood lead levels of 40 pg/
100ml or greater indicating increased absorption, and
22% had levels of 80 PS/100ml or More. The prevalence
of elevated blood lead levels decreased with increased
distance of residence from the smelter. Free erythro-
cyte protoporphyrin levels also increased with blood
lead levels, and 17% of children with lead levels of
80 yg/100ml or greater were anemic. While there was
no overt neurologic toxicity, toxic effects of lead
were found based upon decreased nerve conduction veloc-
ity in children with blood lead levels of 40 Ug/lOOml
or greater. A study by Cooper and Gaffey21 has shown
increased mortality from lung cancer among workers in
lead smelters suggesting the need to evaluate lung
cancer mortality among workers in and residents near
this smelter.
Cadmium
The first indication of potential cancer risk in
man from cadmium dates back to Potts in 1965 . He re-
ported eight deaths from a group of 74 workers exposed
for over 10 years to cadmium oxide dust generated in
the manufacture of alkaline batteries. Potts noted that
three of the deaths were due to carcinoma of the pros-
tate, one was due to carcinoma of the bronchus, and one
was due to carcinomatosis.
Kipling and Waterhouse^ in. 1967 studied 248 cad-
mium workers who had been exposed for a minimum of one
year to cadmium oxide. They found a high incidence of
prostatic carcinoma (4 cases) compared to 0.58 cases
which would have been expected. Other studies by
Humperdinck"^, Holden^5( and Friberg2^, however, failed
to find excess cancer mortality among men occupationally
exposed to cadmium.
These data in part prompted Lemen, et. al.27, to
initiate a mortality study of workers exposed to cad-
mium. The smelter selected for study began production
in 1880 as a lead smelter. It switched to arsenic pro-
duction at the end of World War I and ceased that pro-
duction in 1925 to begin production of cadmium. The
smelter is currently engaged almost exclusively in the
production of cadmium metal and compounds. Some pro-
duction of lead occurs, but this represents only a one
or two-man operation. An industrial hygiene study at
the plant conducted by the state of Colorado's Depart-
ment of Health in 1960 showed that cadmium air concen-
trations in the plant ranged from 3 to 210 Vg/m . The
latest analyses of samples taken in 1973 indicated cad-
mium air concentrations in the premeld department of
90 and 75 Vg/m^, and in the retort department of 1,105
Ug'm^ while arsenic levels were only 1.1 ug/m3, 0.3 yg/
nH, and 1.4 respectively. Exposure at this
smelter was therefore primarily to cadmium oxide with
relatively low concentrations of arsenic.
To evaluate the mortality experience of the work-
ers at this smelter, Leraen, et. al., ascertained occu-
pational histories from company employment files. The
cohort under study was restricted to 29Z males who had
achieved three years employment in the plant between
January 1, 1940, and December 31, 1969. As of August 1,
1975, 180 cohort workers were alive, 92 dead and 20 un-
traced. For analysis purposes those untraced were con-
sidered alive, thus making cause-specific results con-
servative. Total person-years in the study were 6,968.
The total of 92 deaths observed compared to 99.32 ex-
pected deaths based upon U.S. white male rates matched
for age, sex, and race.
Malignant neoplasms, however, accounted for 27 of
the observed deaths whereas only 17.57 would have been
expected (p<0.05). Twelve of the 27 cancer deaths were
due to lung cancer (p<0.05) and four deaths were due to
prostatic cancer (p<0.05).
These results which can be considered conservative,
show significantly increased mortality due to total ma-
lignancies, lung cancer, and prostatic cancer among cad-
mium smelter workers. These results confirm earlier
reported studies and strongly implicate cadmium exposure
as a contributing cause to lung and prostatic cancer.
Conclusions
Occupational and environmental epidemiologic stud-
ies can provide useful and complementary health informa-
tion. At times, however, there has been reluctance to
consider both study approaches when addressing a given
problem. Although the setting of exposure standards for
the general population considers a number of factors not
usually addressed in setting exposure standards for
workers, this does not mean that studies of workers and
communities cannot be interrelated to achieve better
assessment of the effects of chemicals upon human health.
In this regard, iata available from a variety of disci-
plines including epidemiologic, clinical, industrial
hygiene, and animal toxicologic research can provide
valuable information.
REFERENCES
1. Gardner, L.U. & Heslington, H.F. (1946) Osteosarco-
ma from intravenous beryllium compounds in rab-
bits. Fed. Proc., ji>, 221.
2. Dutra, F.R., Largent, E.J. & Roth, J.L. (1951) Os-
teogenic sarcoma after inhalation of beryllium
oxide. Arch. Path,, 51, 473.
3. Vorwald, A.J., Reeves, A.L. & Urban, E.C.J. (1966)
Experimental beryllium toxicology. In: Stokin-
ger, H.E., ed., Beryllium: its industrial hy-
giene aspects, New York, Academic Press
4. Schepers, G.W.H., Durkan, T.M., Delahant, A.B. &
Creedon, F.T. (1957) The biological action of in-
haled beryllium sulfate. Arch. Industr. tilth.,
15, 32.
5. Reeves, A.L., Deitch, D. & Vorwald, A.J. (1967)
Beryllium carcinogenesis. 1. Inhalation exposure
of rats to beryllium sulfate aerosol. Cancer
Res., 27, 439.
6. Reeves, A.L. & Vorwald, A.J. (1967) Beryllium car-
cinogenesis. II. Pulmonary deposition and clear-
ance of inhaled beryllium sulfate in the rat.
Cancer Res., 27, 446.
7. Groth, D.H., Scheel, L.D. & Mackay, G.R. (1972)
Comparative Pulmonary Effects of Beryllium and
Arsenic Compounds in Rats. Lab. Invest. 26, 477.
8. Schepers, G.W.H. (1964) Biological action of beryl-
lium: Reaction of the monkey to inhaled aerosols.
Industr. Med. Surg., 33. 1.
9. Occupational Exposure to Beryllium, National
Institute for Occupational Safety and Health, 1972
10. Hardy, H.L., Rabe, E.W. & Lorch, S. (1967) United
States Beryllium Case Registry (1952-1966):
Review of its method and utility. J. Occup. Med.,
9, 271.
3
1-6
-------
11. IARC Honographs on the Evaluation of Carcinogenic
Risk to Man. Volume 1, International Agency for
Research oh Cancer, World Health Organization
Lyon 1972.
12. Mancuso, T.S1. (1970) Relation of duration of em-
ployment and prior respiratory illness to res-
piratory cancer among beryllium workers.
Envlronm. Res. _3, 251.
13. Kovalsky, V.V. et. al. (1961) Changes in Purine
Metabolism in Man and Animals in Various Molyb-
denum Rich Geochemical Provinces. Transae, of
WHO, from Z Obsc. Biol.. 22, 179.
14. Blejer, H.P., and Wagoner, J. "Inorganic Arsenic";
Ambient level approach of occupational canceri-
genic exposures,: paper presented at the Confer-
ence on Occupational Carcinogenesis, Hew York
Academy of Sciences, New York City, March 25,
1975.
15. Lee, A.M., and Fraumeni, J.F., Jr. (1969) "Arsenic
and Respiratory Cancer in Man - An Occupational
Study," J. Nat. Cancer Inst.. 42, 1045.
16. Ott, M., Holder, B., & Gordon, H. (1974) "Respira-
tory Cancer and Occupational Exposure to Arseni-
cals," Arch. Environ. Health. 29, 250.
17. Blot, W.J. and Fraumeni, J.F. (1975) Arsenical Air
Pollution and Lung Cancer, Lancet. IX, 142.
18. Milham.^S. and Strong, X., Human Arsenic Exposure in
Relation to a Copper Smelter (1974) Environ. Res.
It 176.
19. U.S. Cancer Mortality by County; (1950-1969) U.S.
Department of Health, Education, and Welfare,
National Cancer Institute, DHEW Publication No.
(NIH) 74-615.
20. Landrigan, P. et al., (1975) "Increased Lead Ab-
sorption with Anemia and Subclinical Neuropathy In
Children Near a Lead Smelter," Center for Disease
Control, Atlanta, Ga.
21. Cooper, W.C. and Gaffey, W.R. (1975) Mortality of
Lead Workers, J. Occup. Med. 17, 100.
22. Potts, C.L. (1965) "Cadmium Proteinuria—the Health
of Battery Workers Exposed to Cadmium Oxide Dust,"
Ann. Occup. Hyg., jS, 55.
23. Kipling, M.D., and VJaterhouse, (1967) "Cad-
mium and Prostatic Carcinoma," Lancet. j[, 730,
24. Humperdlnck, (1968), "Kadmium and Kungkrebs,"
Med. Klin., 63, 768.
25. Hoiden, (1969) "Cadmium Toxicology," Lancet. II.
57.
26. Friberg, L., Piscator, M. and Nordberg, G. (1971)
Cadmium in the Environment. Chemical Rubber Com-
pany Press, Cleveland, Ohio,
27. Lemen, R.A., Lee, J.S. & Wagoner, J.K., "Mortality
Among Workers Exposed to Cadmium," paper presen-
ted at the Conference on Occupational Carcinogen-
esis, New York Academy of Sciences, New York,
March 27, 1975. (Results updated to August 1,
1975).
4
1-6
-------
METHYLMERCURY: FORMATION IN PLANT TISSUES
Don D. Gay
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
P.O. Box 15027
Las Vegas, Nevada 89114
SUMMARY
This investigation identified methylmercury
(CH3Hg) as one of the mercury species present in peas.
The study also showed that methylmercury was formed by
the plants (a) when grown in vermlculite and the leaves
sprayed with as little as 90 milliliters of 10 micro-
grams/gram of mercuric nitrate (Hg(N03)2)> (b) when
grown in soil with mercuric nitrate or phenylmercury
added, or (c) when sections of the pea plants were
surface sterilized with 5 percent Chlorox or Gentamicin
added to the incubation medium and incubated 20 hours
in 10 micrograms/gram mercuric nitrate or 10 micro-
grams/gram phenylmercury acetate. The age of the peas
plays a role in the amount of methylmercury formed.
Older pea tissues produce lees methylmercury than young
tissues when these tissues are infiltrated and incuba-
ted with 10 micrograms/gram of mercuric nitrate.
To check for the presence of methylmercury in
plants grown in a natural environment but near a
mercury source, samples of three species of plants were
collected around an abandoned mercury mine. All three
of the species show the presence of methylmercury but
the influorescence of Bromua rubena was especially high
in methylmercury. These plants were collected in early
May 1975 before the summer heat had dried, all aerial
portions.
INTRODUCTION
With the continued increase in mercury pollution
of the entire ecosystem through natural and man-made
Sources, increased study of the exposure of vegetation
is warranted. Data on the effects of mercury pollution,
the uptake and distribution of mercury, the transform-
ation and identification of mercury species, and the
cycling of mercury in vegetation are vitally needed to
complete the environmental picture.
Most plant studies are concerned with the uptake
of mercury by various plant species1! 3> 5» 6> 8»
13, 1H, 15, 17, 19, 20, 23, 24, 27, 29, 30, 31, 40,
In addition to the translocation studies of phenyl-
tnercury compounds in rice plants, Fukunaga et at.13
found traces of methylmercury present in phenylmercury
treated rice plants,
Clendenning and North®, working with the giant
kelp Maarooyetie pyrifera, showed that 100 nanograms/
gram of mercuric chloride (HgCl2> in water caused a
50 percent inactivation of photosynthesis. Harriss et
al- reported that concentrations of organomercurial
fungicides as low as 0.1 nanograms/gram in water re-
duced photosynthesis and growth of plankton. The fresh
water planktonic diatom, Synedva ulna, was investigated
by Fujita and Hashizume12 for its ability to accumulate
mercury. The uptake of mercury from the medium occur-
red rapidly during the first seven hours. The mercury
was found mainly adsorbed to the surface of the cells.
An inhibition of photophosphorylation by isolated
chloroplasts was found by Bradeen et at,7 to be inhib-
ited by the addition of mercuric chloride. Only one
of the two functionally isolated sites of photophos-
phorylation coupled to electron transport, the site
located between the oxidation of plastoquinone and the
reduction of cytochrome was found to be sensitive to
mercuric chloride. Watling-Payne and Selwyn31* found
that phenylmercuric acetate was a poor inhibitor of
photophosphorylation. Radmer and Kok25 showed that
mercuric chloride, when added to isolated chloroplasts,
inhibited the electron flow between Photosystems I
and II. This was due to the inactivation of plasto-
cyanin, an electron carrier close to P7oo>
\
Ahmed and Grant2 showed cytological abmormalities
occurring in root tips of Tradeeoantia and Vieia faba
following treatment of the tips with 1 to 5 micrograms/
gram of Panogen 15, a mercurial fungicide. Rao et at,
25, working with pea roots in mercuric and phenylmer-
curic acetate solutions showed a cellular distribution
of mercury based on differential centrifugation of
homogenized root tips into nuclear, mitochondrial,
microsomal, and soluble fractions.
Anelli et at.4 investigated the influence of metal-
lic mercury vapor on the amino acid content of tobacco
leaves. Their selection of tobacco is based on the
work of Zimmerman and Crocker111 which showed that to- '
bacco is one of the most resistant plants to mercury
poisoning since it can accumulate more than 3,000 micro-
grams/gram of mercury in leaves and still be viable at
maturity. Anelli et al. ** showed that the Insoluble
proteic amino acid content increased throughout the
experiment (after a slight decrease due to initial poi-
soning of mercury). The amounts of cysteine, glutamic
acid and glycine appeared exceptionally high. The
level of systeine in mercury vapor treated plants was
over six times as great as the level in control plants.
The possibility of mercury as a trace element for
plant growth and development was suggested by Dobrol-
yubskii10. In his work with grapevines sprayed twice
with mercuric sulfate (HgSOij) , he noted an increase in
the weight of 100 berries sprayed vs. control, an in-
creased sugar content, decreased berry acidity as
tartaric acid and an acceleration of grape ripening as
shown by an increase in its glucoacidometric index.
The chlorophyll content in treated leaves increased
over the control. The carbohydrate metabolism was
changed after the addition of mercuric sulfate. The
sugar content increased at the expense of sucrose
which was all converted to glucose and fructose. This
conversion Is due to invertase activity which likewise
increased in the treated plants.
Siegel et al,2S noted analytical inconsistencies
in mercury content of plant parts from a single col-
lection of plants when assayed over a period of sever-
al hours. They have found an unknown volatile,
organic-solvent soluble mercury compound that j.s not
methyl- or dimethylmercury as determined by gas chro-
matography. The loss of the volatile compounds from
hexane, methanol, and water follows the volatility
series hexane>methanol>water.
The areas of metabolism and biotransformations of
mercury in plants need further study. The purpose of
this study was to determine the chemical forms of
1-7
-------
mercury in. plants. Subsequent to this determination
was the elucidation of the role of the plant in trans-
forming one chemical form of mercury to another.
MATERIALS AND METHODS
PLANTS
Seeds of Pigum sativum L. var. Alaska and Little
Marvel were purchased from Burpee Seed Company in
Riverside, California. The supplier stated that the
expected germination was 92 percent.
Plants collected from the mercury mine area were
(a) Bromue rubens; (b) Spharalosa ambigua; (c) a
Boraginaaeae.
GROWING MEDIA
Vermiculite and soil which was collected from the
University of Nevada Agricultural Experiment Station
at Logandale, Nevada, were used. The soil, a fine
sandy loam) has the following characteristics:
(a) 54 percent sand
(b) 11 percent clay
(c) 35 percent silt
(d) 1.3 percent organic carbon
(e) pH 8 .*6
PLANTING METHOD
Plastic. trays, 29.16 x 36.78 x 8.84 cm (11.5 x
14.5 x 3.5 inches), were filled to within 1.22 cm (%
inch) of the top with a medium of vermiculite or soil.
Pea seeds were placed 2.54 to 3.76 cm (1 to 1.5 inches)
apart and pushed down into the medium even with the
surface. The trays were then filled to the top with
medium, lightly tamped, and watered thoroughly. Addi-
tional watering was done via an automatic system twice
a day at 8 a.m. and 4 p.m. If peas were grown longer
than two weeks, complete nutrient solution was added
weekly.
GERMINATION METHODS
Pea seeds were placed in a beaker of water or
mercury solution, depending on the experiment. A
stream of air was bubbled through the solution for the
entire imbibing period. The swollen seeds were washed
thoroughly with distilled water and placed in pans
lined with moistened paper towels. The pans were cov-
ered with plastic-coated paper and the paper secured
with the aid of rubber bands. The pans with seeds were
then placed inside an oven set at 25° C. At the end
of the germination period, the young pea plants were
separated into epicotyls, roots, and cotyledons, and
each weighed. The lengths of epicotyls and roots were
also measured and recorded.
TISSUE EXTRACTION PROCEDURE
All tissue was washed with distilled water prior
to use. The tissue was homogenized in a Waring blender
with 2.2 N HCL (1 ml HCl/g tissue) for two minutes.
The resulting brei was centrifuged at 10,000 rpm for
30 to 45 minutes. The supernatant solution was decant-
ed and filtered, and extracted with three volumes of
nanograde quality benzene, each equal to the volume of
superqate. All benzene fractions were passed through
Whatman Phase Separating paper and combined. A freshly
prepared 1 percent cysteine solution (Westoo35-38) was
titrated to pH 8.3 to 8.4 with 5 N NaOH immediately
before use. Ten ml of the cysteine solution per 100 ml
of benzene was added to the benzene and vigorously
shaken. After separation of the aqueous and organic
layers, the cysteine layer was drawn off and acidified
to pH 0,5 to 0.7 with 5 N NCI. This acidified solution
was allowed to stand for 15 minutes after which 10 ml
of benzene was added and vigorously shaken. After
separation, the benzene layer was drawn off into a
glass vial with anhydrous Na2S0it covering the bottom.
MODIFICATION OF WESTOO'S CYSTEINE EXTRACTION SOLUTION
Westoo's cysteine extraction solution consists of
cysteine • HC1, sodium acetate, and anhydrous sodium
sulfate. The pH of the resulting solution is "\<3.8.
Using a freshly prepared methylmercury solution, only
50 to 60 percent recovery of the methylmercury could
be obtained. Wong et ai.39 found that binding of
methylmercury to another sulfur-containing amino acid,
penicillamine, occurred via deprotonation of the sulf-
hydryl group. The ionization of cysteine at pH 7 is
only 8 percent. Raising the pH to 8.3 results in 50
percent ionization of the sulfhydryl group of cysteine.
Weatoo's cysteine solution pH of 3.8 was adjusted to
pH 8.3 to 8.4 with 5 percent NaOH and used immediately.
The recovery of methylmercury increased to 86 percent
INFILTRATION AND INCUBATION METHODS
Selected plant sections were placed in a flask
containing the mercury solution. This flask was
placed inside a plastic dessicator fitted with a stop-
cock. The clear plastic lid was covered with aluminum
foil to make the chamber light-free. An in-line act-
ivated charcoal filter was placed in the pressure
tubing between a vacuum pump and the dessicator. The
tissue was placed under a vacuum for 15 minutes, the
vacuum pump turned off, and the inside of the desBica-
tor allowed to return to an equilibrium with the
ambient pressure. The stopcock was then closed and
the incubation allowed to proceed inside the darkened
chamber.
GAS CHROMATOGRAPHY
A Hewlett-Packard Gas Chromatograph Model 5713 (1)
with a fi3Ni linearizing electron-capture detector and
a 6-foot column packed with 5 percent HIEFF was used
to separate organic mercurial halides. The carrier
gas was 95 percent argon-5 percent methane obtained
from Matheson (2) and used at a flow rate of 60 ml/
minute. The oveu temperature was 170° C. and the
detector and injection port temperatures were 200° C.
The column was on-line in the injection port.
CHELATING RESIN
In order to remove ionic mercury from samples
before injection onto the gas chromatograph, Sraflon
NMRR (3) chelating resin was employed. Preliminary
work by Law22 indicated the possibility of separating
ionic mercury from methylmercury as a function of pH.
Experiments at the Environmental Monitoring and Support
Laboratory-Las Vogas confirm tl\is separation. The
resin chelates ionic mercury at low pH (pH M3.5) and
chelates methylmercury at pH 6.4. The methodology of
utilization of the resin is as follows. The resin Is
washed repeatedly with glass distilled water, allowed
to settle for 10 minutes, and the liquid decanted.
(1) Hewlett-Packard, 11300 Lomas Blvd, Albuquerque,
NM 87123.
(2) Matheson Gas Products, 8800 Utica Avenue,
Cucatnonga, CA 91730
(3) Ayalon Water Conditioning Co., P.O. Box 586,
Haifa, Israel.
2
1-7
-------
After washing, 10 to 20 grams of the resin is poured
into a column. A glass fiber filter is cut to fit the
inside diameter of the column. With the stopcock open
the filter is gently pushed down on top of the resin
bed. The pH of the solution to be placed on the
column is critical for chelation of the desired
mercury species.
In this case, the chelation of ionic mercury was
desired. The sample (10 to 20 ml) was placed on the
column and flow rate adjusted to 0.5 to 0.6 ml/minute.
After the void volume of the column had passed through,
the volume collected was equal to the volume of sample
plus 10 ml of the glass distilled water.
Elution of the ionic species of mercury from the
column was accomplished by the addition of a 5 percent
thiourea in 0.06 N HC1. When this elution was complete
the column was flushed with 300 ml of glass distilled
water and regenerated for future use.
MERCURY TRAP SYSTEM
In order to determine whether ethyl- or methyl-
mercuric chloride was released during the incubation
of tissue in mercuric nitrate solution, a modification
of the Kimura and Miller21 mercury entrapment procedure
was used. Nitrogen gas was bubbled through the incub-
ation solution and through two traps. The first was
carbonate-phosphate solution followed by a 1 percent
cysteine solution at pH 8.3. The flow rate of nitrogen
was 112 ml/minute for 20 hours of incubation. The
carbonate-phosphate trap removes volatile ethyl and
methylmercury chloride. The cysteine trap is used as
a check on the primary carbonate-phosphate trap.
VERIFICATION OF METHYLMERCURY FROM GAS CHROMATOGRAPHY
To verify that Hg(N03>2 and/or the concentration
of Hg(N03)2 used in the experiments was not producing
a peak on the chromatogram at the same time as CH3HgCl,
Hg(N03>2 at the concentration of 10 yg/g was dissolved
in distilled water and 90 ml of this solution was
taken through the extraction procedure. A 5-yl aliquot
of the extract was injected into the gas chromatograph.
No CH3HgCl peak appeared.
With the formation of the chloride species of
mercury in the extraction procedure and the possibility
of forming HgCl2 from the Hg(N03)2, a HgCl2 solution
at the concentration of 170 yg/g was made in benzene.
When 5 ml of this solution was injected onto the gas
chromatographic column, the HgCl2 peak appeared at the
same time as the reference CHjHgCl peak. This 170
(ig/g HgCl2 benzene solution was extracted via the
cysteine procedure and 5 yl of the final 10 ml benzene
layer was injected onto the gas chromatograph column.
In this case, no peak corresponding to the CH3HgCl
reference peak appeared.
PURIFICATION OF REAGENT GRADE MERCURIC NITRATE
Reagent grade Hg(N03)2 was found to contain HgCl2
impurity. To remove this impurity, a small portion of
the Hg(N03)2 was poured into a flask and washed three
times with benzene. The mixture was shaken, allowed
to settle, and the benzene was carefully decanted.
After three such treatments, diethyl ether was added
and carefully swirled. After allowing to settle and
decanting the ether, acetone was added and the same
procedure employed except that the acetone-Hg(N03)2
was in the filter, the mercury was stirred while under
vacuum. The Hg(N03)2 was transferred to a dark bottle
and placed in the freezer for future use.
GAS CHROMATOGRAPHIC ANALYSIS OF STOCK ORGANIC MERCURY
COMPOUNDS
Solutions of 50 yg/g were made from the stock
500-ug/g benzene solutions of ethyl-, methyl-, and
phenylmercury chlorides. Five yl of each was injected
onto the gas chromatograph column. The contamination
of ethylmercuric chloride with methylmercuric chloride
was confirmed. The phenylmercury chloride showed no
contamination of methyl- or ethylmercury compounds.
GAS CHROMATOGRAPH UNIFORM INJECTION PROCEDURE
To insure consistent and known volumes injected
onto a gas chromatographic column, a 0- to 10-yl syringe
was used. The procedure is as follows:
(a) Rinse the syringe twice with solution to be
injected and flush.
(b) Draw 0.6 yl of air into the syringe.
(c) Draw 5.0 yl of solution to be injected into
the syringe after the 0.6 yl of air and then
inject into the gas chromatograph.
Peak heights of standard methylmercury-benzene
solutions are reproducible to within 1 percent.
RESULTS AND DISCUSSION
GERMINATION EXPERIMENTS
The sensitivity of Little Marvel pea seeds to
concentrations of ionic mercury compounds was deter-
mined by germinating pea seeds in 1-, 10-, and 25-yg/g
distilled water solutions of mercuric acetate and mer-
curic nitrate for 6 hours with aeration. They were
placed in separate germination pans after extensive
washing with distilled water. The seeds were germin-
ated at 25° C. for 4 days completely in the dark. At
the end of this time, the young pea plants were separa-
ted into three sections—epicotyls, roots, and cotyle-
dons . The epicotyls and roots were measured and the
combined tissues of each section were weighed. The
percent germination of each lot of peas was recorded.
Burpee Seeds states a 92 percent germination for this
batch of peas. Each section for each condition was
extracted according to the modified Westoo procedure
to determine if any detectable organic mercury halides
were present.
The results are given, in Table 1. No organic
mercury was detected. The data show that the nitrate
and acetate forms of ionic mercury do not elicit com-
parable responses in the same variety of peas. Only
52 percent germination occurred for peas in 25 ug/g
Hg(N03)2 while 68 percent was noted for 25 yg/g HgOAc.
The total weights of epicotyls and roots in the 25 yg/g
mercuric nitrate were 100 percent and 73 percent great-
er, respectively, than those in 25 yg/g mercuric
acetate. The overall response of the peas to the
mercuric acetate was a diminution in growth and weight
with an increase in HgOAc concentration. The response
to mercuric nitrate definitely did not follow a diminu-
tion pattern as concentration of Hg(N03)2 increased.
Based on these experiments, Hg(N03)2 was chosen as the
ionic form of mercury to be used in subsequent invest-
igations and 10 yg/g was chosen for the preliminary
concentration with which to work.
FOLIAR APPLICATION OF MERCURY
Little Marvel peas were grown in vermiculite in
exposure chambers for 34 days with weekly addition of
a complete nutrient solution. On the 34th day after
3
1-7
-------
planting, the leaves of the peas were sprayed by atomi-
zer over an 8-hour period with a total of 90 ml of 10
lig/g Hg(N03)2 solution made up in distilled water.
Watering was done at the vermiculite level, so little
if any of the tfg(N03)2 was on the vermiculite available
to root absorption. One flat of peas was used and the
dense canopy of leaves coupled with the small volume
sprayed at any one time and the fineness of the mist
eliminated any run-off to the vermiculite. The peas
were grown an additional 5 days with no further addi-
tion of mercury. On the 39th day after planting, all
tissue above the third node was harvested and extracted
according to the modified Westoo cysteine procedure.
A total of 125.53 grams of tissue fresh weight was ob-
tained. When 5 vl of the final benzene layer was anal-
yzed by gas chromatography, a peak at the time of the
methylmercury chloride reference appeared. Relating
the peak height to standards gave a concentration of
0.16 ng ± 0.06 ng of methylmercury chloride/gram of
tissue fresh weight based on a conservative 60 percent
efficiency for the entire extraction procedure.
MERCURY ADDITION TO SOIL MEDIUM
Mercuric nitrate and phenylmercurie acetate in
powder forms were each mixed with 6 kg of sandy-loam
soil, pH 8.6, with characteristics given previously,
to obtain 10 and 100 vg/g soil-mercury mixture of each.
This contaminated Soil was placed in trays and Little
Marvel seeds were planted in each 2.54 cm (1 inch)
apart. After 14 days, the aerial portions were harvest-
ed and rinsed thoroughly. The modified cysteine ex-
traction procedure was used to extract the tissue.
Table 2 gives the amount of methylmercury found
per gram of aerial tissue from peas grown in the mer-
curic nitrate or pheynlmercuric acetate contaminated
soil. The amount of methylmercury detected in the tis-
sue increased 100 percent from 10 ug/g to 100 vg/g
mercuric nitrate in the soil. The amount of methyl-
mercury decreased with corresponding concentrations of
phenylmercuric acetate in the soil, indicating that the
two forms of mercury have different effects or sites
of activity within the plant.
Various investigators9'16'32'33 working with soil
types and soil pH have found the root uptake of mercury
into plants to decrease with neutral or alkaline soils.
The use of an alkaline soil was to determine if mercury
uptake would occur and, if so, if the small amount taken
up would be transformed to methylmercury in concentra-
tions large enough to be detected with the system at
hand.
INFILTRATION AND INCUBATION EXPERIMENTS WITH MERCURIC
NITRATE
The detection and identification of methylmercury
in pea plants after the addition of Hg(N03)2 to leaves
or applied to the soil could be the result of microbial
conversion of the ionic to an organic mercury species
with the concomitant uptake of the methylmercury com-
pound by the plant. An experimental procedure was
devised to determine if the plant tissue was capable of
transforming ionic mercury to methylmercury.
Alaska peas were grown in vermiculite and after 20
days from planting various internodes were harvested,
weighed, thoroughly rinsed with distilled water, and
placesj in flasks with 100 ml of 10 ug/g Hg(N03)2 in
glass distilled water. The flasks were placed under
vacuum from a water aspirator for 30 minutes. The
flasks were stoppered and incubated for 89>s hours. The
tissue in each flask was thoroughly rinsed with distill-
ed water (a minimum of five complete rinses was used)
and extracted via the modified Westoo procedure.
Gas chromatographic analysis of 5 pi of each final
benzene extract revealed the presence of CH3HgCl from
all stem sections. Table 3 gives the weight of tissue
in each fraction and the amount of CH3HgCl detected.
The amount of CH3Hg in internode sections up to the
apical segments is three times as much as the subtend-
ing internodes.
Little Marvel peas were harvested after 17 days
growth into stems and apical region. No leaves or
laterals were used. The same infiltration procedure
was used. The incubation period was only 20 hours.
After thoroughly rinsing the tissue, each was extracted
according to the modified Westoo procedure. A compar-
able amount of tissue was used for a control. The
stems plus apices were thoroughly washed and infiltra-
ted the same length of time in glass distilled water
in separate chambers with in-line activated charcoal
to eliminate the possibility of mercury contamination.
The control tissue was incubated 18 hours after which
it was washed and homogenized in 2.2 N HC1 in a blender
and centrifuged. The supernatant solution was extract-
ed via the modified Westoo procedure. The pellet was
resuspended in glass distilled water by magnetically
stirring for 15 minutes. The suspension was extracted
via the modified Westoo procedure.
The results of the tissue incubated in 10 yg/g
Hg(N03)2 and the control tissue are given in Table 4.
After 20 hours of incubation in Hg(N03)2, stems and
apices form CH3Hg with the stems forming more per gram
than apical region. The control plants with no addi-
tion of Hg(N03)2 show no CH3HgCl in either the super-
nate or the pellet.
The formation of methylmercury from mercuric ni-
trate by plant tissue within 20 hours would seem to
rule out bacterial contamination as causing the trans-
formation because the time involved is so short. How-
ever, to determine if bacteria associated with the pea
stem are causing this transformation, two experiments
were performed using two different sterilization pro-
cedures. Stems from Little Marvel peas grown 18 days
were harvested and divided into two 10-gram fractions.
Both fractions were washed thoroughly with distilled
water and one fraction was washed again for 5 minutes
in 5 percent Chlorox solution. The other fraction was
the control. The Chlorox was rinsed off thoroughly
and both were infiltrated under vacuum in separate
chambers for 15 minutes and incubated for 20 hours in
100 ml of 10 vg/g Hg(N03)2. The stems were thoroughly
rinsed again to remove any mercury on the surfaces and
extracted via the modified Westoo procedure. The sec-
ond experiment involved stems from Little Marvel peas
grown 38 days. Two 10-gram fractions were obtained.
Both fractions were washed, one placed in 100 ml of 10
Ug/g Hg(N03)2 and the other placed in 100 ml of 10 ug/g
Hg(N03)2 to which Gentamicin, an antibacterial agent,
was added (2 mg/ml). Both were infiltrated and incuba-
ted in separate chambers for 20 hours. The stems were
again rinsed and extracted via the modified Westoo
procedure. Also via of the modified Westoo procedure,
50 ml of concentrated Chlorox was extracted to deter-
mine if any methylmercury was present as a contaminant
in the Chlorox.
The results of both experiments are given in Table
5. There are no significant differences between com-
parable experiments. The age of tissue seems to have
an effect on the ability to transform ionic mercury
into methylmercury. Stems 38 days old produced only
61 percent as much methylmercury as did the stems, 18
days old (using the average of both treatments).
The influence of pH on the transformation of ionic
mercury to methylmercury was determined by adjusting
the pH of the incubation solutions to 3.5, 4.5, 5,5,
1-7
-------
6,5, 7.5 or 8.5. A physiologic concentration of 0.01
M was used for l^HPOij and Kl^POi,. The lowest pH (3.5)
was obtained by adding l^PO^. Little Marvel peas,
stems and apices 13 and 35 days old were used. The
tissue from each age group of peas was divided into 6
10-gram fractions, infiltrated, and incubated 20 hours.
The results as given in Figure 1 show that the age of
the tissue has some effect on the transformation of
ionic mercury to methylmercury. It is interesting to
note that the peak heights of methylmercury at the two
extremes of pH comparing age of tissue are almost the
same. Obviously, other factors that are not present or
suppressed in the physiologic pH range are influencing
the formation of methylmercury at these extremes.
With the possibility of the loss of some volatile
organic mercury compounds from the incubation flask
during incubation, an experiment was performed to trap
these volatile organic mercury compounds. Little Mar-
vel pea stems 13 days old were harvested and 21.9 grams
obtained. The stems were thoroughly rinsed and infil-
trated in a glass distilled water incubation medium
with 10 pg/g Hg(N03)2. The incubation flask with stems
was placed in a nitrogen gas train as described in the
methods. The incubation period was 20 hours after
which the tissue was rinsed thoroughly and extracted
via the modified Westoo procedure. The carbonate-phos-
phate solution and cysteine solution were removed from
the traps and also extracted via the modified procedure
Gaa chromatographic analysis of -the traps revealed the
possibility of a very small amount of methylmercury
present. If such peaks are indeed present on the chro-
matograms, they are almost at the level of background
noise. The stems showed 8.3 ± 2.5 ng/g tissue methyl-
mercury .
INFILTRATION AND INCUBATION EXPERIMENTS WITH PHENYL-
MERCURIC ACETATE
Rice plants treated with phenylmercuric acetate
reportedly produce small quantities of methylmercury
(Fukunaga et a£.)13. To determine if phenylmercury is
.transformed to methylmercury, Little Marvel stems with
apices and leaves were harvested and 10 grams of each
tissue was infiltrated and incubated in separate cham-
bers with phenylmercuric acetate. Two 10-ug/g aqueous
solutions were attempted but after stirring for 48
hours an estimated 20 percent of the phenylmercury in
each may have gone into solution. The solutions were
allowed to settle and the aqueous solution decanted
without carrying any of the undissolved material over
into the incubation flasks. After 20 hours incubation,
the tissues were thoroughly rinsed and extracted via
modified Westoo procedure. Gas chromatographic anal-
ysis of the final benzene extracts shows the presence
of methylmercury in stems with apices at 10 ± 3 ng/g
and leaves at 8 ± 2.4 ng/g. This level of transform-
ation occurred with an estimated 2-ug/g solution of
phenylmercuric acetate.
EXTRACTION OF PLANTS FROM A MERCURY MINE AREA
Methylmercury has been shown to be present in
plant tissue from foliar application of ionic mercury,
from root uptake of ionic and phenylmercury and from
tissues incubated in ionic and phenylmercury solutions.
Is methylmercury present in plants growning naturally
in an area with elevated levels of mercury? To find
out, several plants were collected from the mercury
mine area at the Nevada Test Site in mid-May 1975.
Bromue rubene and Sphavaloea ccmbigua were two of the
species collected. A Boraginaaeae was also collected.
In plant samples collected in November-December
1974 from the Four Corners area around the coal-burning
power station, no appreciably high levels of mercury
were noted. The physiology of the plants during this
collection period is far different from that of the
same plants in the spring. In order to survive the
long, hot, dry, summers in the desert, the plants enter
a dormant state. In the spring after a few showers the
physiology changes to an active growth period. Mid-May
was selected for collection to obtain plants still act-
ively growing and also so that small leafy plants could
be obtained.
The plants were extracted in 2.2 N HC1 which was
extracted with benzene. The modified Westoo procedure
was used for the extraction and concentration of organ-
ic mercury compounds from the benzene. A 5—yl aliquot
of the final 10-ml benzene layer was injected onto the
gas chromatographic column. All three plants showed a
methylmercury peak, but that from Bromue rubens was
exceptionally large. To determine that the peak was,
in fact, methylmercury, 6 ml of the final benzene layer
was extracted with 5 ml of the 1 percent cysteine solu-
tion at pH 8.4. The cysteine-mercury complex was brok-
en by the addition of 5 N HC1 to a pH 0.7. This acidi-
fied fraction was placed on a column of Srafion NMRR
resin. The eluate was extracted into benzene (6 ml)
and a 5-pl sample injected on the gas chromatographic
column. Again the methylmercury peak showed up almost
as large as before. The amount of methylmercury present
in Bvomus rubena was 10 ± 3.3 ng/g, for the Boragina-
aeae, 7.5 ± 2.3 ng/g, and for Spharalaea arribigua, 2.5
±0.8 ng/g.
REFERENCES
1. Aakrog, A., and Lippert, J. "Direct Contamination
of Barley with 51Cr, 59Fe, 58Co, 65Zn, 203Hg,
210Pb," Radiat. Bot., 11, No. 6, pp 462-472 (1971).
2. Ahmed, M., and Grant, W. F., "Cytological Effects
of the Mercurial Fungicide Panogen 15 on Tradeeaan-
tia and Vioia fciba root tips," Mutat. Res., 14, pp
391-396 (1972).
3. Alvarez, R., "Sub-microgram per Gram Concentrations
of Mercury in Orchard Leaves Determined by Isotope
Dilution and Spark-Source Mass Spectrometry,"
Anal. Chem. Aata, 73, pp 33-38 (1974).
4. Anelli, G., Pelosl, P., and Galoppini, C., "Influ-
ence of Mercury on the Amino Acid Composition of
Tobacco Leaves," Agri. Biol, Chem., 37, No. 7, pp
1579-1582 (1973).
5. Bache, C. A., Gutenmann, W. H., St John, L. E., et
al., "Mercury and Methylmercury Content of Agricul-
tural Crops Grown on Soils Treated with Various
Mercury Compounds," J. Agr. Food Chem., 21, No. 4,
pp 607-613 (1973).
6. Barker, W. G., "Toxicity Levels of Mercury, Lead,
Copper, and Zinc in Tissue Culture Systems of Cauli-
flower, Lettuce, Potato, and Carrot," Can. J. Bot.,
50, pp 973-976 (1972).
7. Bradeen, D. A., Winget, G. D., and Gould, J. M.,
"Site Specific Inhibition of Photophosporylation
in Isolated Spinach Chloroplasts by Mercuric Chlor-
ide," Plant Physiol52, No. 6, pp 680-682 (1973).
8. Clendenning, K.A., and North, W. J., "Effects of
Wastes on the Giant Kelp, Macrooyetie pyrifera
pp 82-91, In E. D. Pearson (Ed.), Proa. Int. Conf.
Waste Disposal Marine Environ., Berkeley, California,
Pergamon Press, New York (1958).
9. D'ltri, F. M., The Environmental Mercury Problemj
CRC Press, Cleveland, Ohio, pp 63-67 (1972).
5
1-7
-------
10. Dobrolyubskii, O.K., "Mercury Trace Element as
the Influencing Factor on Plant Organisms,"
Doklady Akad, Nauk SSSR (Bot. Sai. Sea. Transl.),
128, No. (1/6), PP 253-255 (1959). Translated:
(from) Doklady Akad. Hauk SSSR, 125, No. 5 (1960).
11. Estes, G. 0., Knoop, W.E., Houghton, F.D,, "Soil-
plant Response to Surface Applied Mercury," J.
Environ. Qual., 2, No. 4, pp 451-452 (1973).
12. Fujita, M., and Hashizume, K., "Accumulation of
Mercury by Fresh Water Planktonic Diatom,"
Chemosphere, 1, No. 5, pp 203-207 (1972).
13. Fukunaga, K., Teukano, Y., and Kanazwa, J.,
"Residue Analysis of Organomercury Fungicides
Sprayed on Rice Plants," Environ. Toxicol. of
Pesticides, Academic Press, New York, pp 177-191
(1972).
14. Gerdes, R. A., Hardcastle, J. E., and Stabenow,
K. T., "Mercury Content of Fresh Fruits and Vege-
tables," Chemosphere, No. 1, pp 13-18 (1974).
15. Gilmour, J. T., and Miller, M. S., "Fate of a
Mercuric-mercurous Fungicide Added to Turfgrass,"
J. Environ. Qual., 2, No. 1, pp 145-148 (1973).
16. Hale, V. Q., ^and Wallace, A., "Effect of Chelates
on Uptake of 'Some Heavy Metal Radionuclides from
Soil by Bush Beans," Soil Sai., 109, No. 4, pp
262-263 (1970)
17. Haney,., A., and Lipsey, R. L., "Accumulation and
Effects of Methylmercury Hydroxide in a Terres-
trial Food Chain Under Laboratory Conditions,"
Environ. Pollution, 5, pp 305-316 (1973).
18. Harriss, R. C., White, D. B., and Macfarlane, R.
B., "Mercury Compounds Reduce Photosynthesis by
Plankton," Science, 170, pp 736 (1970).
19. James, P. E., Lagerwerf, J. V., and Dudly, R. F.,
"Translocation of Mercury from Seed Treatment,"
Int. Symp. Ident. Meae. Envir. Pollut. Proa.,
pp 213-215 (1971).
20. John, M. K., "Mercury Uptake from Soil by Various
Plant Species," Bull. Environ. Contam. Toxicol.,
8, No. 2 pp 77-80 (1972).
21. Kimura, Y., and Miller, V, L., "Vapor Phase Sepa-
ration of Methyl or Ethyl Mercury Compounds and
Metallic Mercury," Anal. Chem., 32, pp 420 (1960).
22. Law, S. L., "Methyl Mercury and Inorganic Mercury
Collection by a Selective Chelating Resin,"
Science, 174, No. 4006, pp 285-287 (1971).
23. Lee, D. G., Thomas, B., Roughan, J. A., and
Watters, E. D., "Mercury Content of Some Food-
stuffs of Vegetable Origin," Pestic. Sci., 3,
pp 13-17 (1972).
24. Newsome, W. H., "Determination of Methylmercury
in Fish and Cereal Grain Products," J. Agr. Food
Chem., 19, pp 567 (1971).
25. Radmer, R., and Kok, B., "Kinetic Observation of
the System II Electron Acceptor Pool Isolated by
Mercuric Ion," Bioahem. Biophye. Acta, 357,
pp 177-180 (1974).
26. Rao, A. V., Fallin, E., and Fang, S. C., "Compara-
tive Study of Uptake and Cellular Distribution of
®03Hg-Labeled Phenyl-mercuric Acetate and
Mercuric Acetate by Pea Roots," Plant Physiol.,
41, pp 443 (1966).
27. Ross, R. G., and Stewart, D. K. R., "Movement and
Accumulation of Hg in Apple Trees and Soil,"
Can. J, Plant Sci., 42, pp 280-285 (1962).
28. Siegel, S. M., Puerner, N. J., and Speitel, T. W.,
"Release of Volatile Mercury from Vascular Plants,"
Physiol. Plant., 32, No. 2 pp 174 (1974).
29. Smart, N. A., "Mercury Residues in Potatoes Fol-
lowing Application of Foliar Spray Containing
Phenylmercuric Chloride," J. Sai. Food Agr.,
15, pp 102-108 (1964).
30. Smith, W. H., "Lead and Mercury Burden of Woody
Plants," Science, 176, No. 4040, pp 1237 (1972).
31. Tkachuk, R., and Kuzina, F. D., "Mercury Levels in
Wheat and Other Cereals, Oilseed and Biological
Samples," J. Sai. Food Agr., 23, No. 10 pp 1183-
1185 (1972).
32. Van Loon, J. C., "Agricultural Use of Sewage
Treatment Plant Sludge, a Potential Source of
Mercury Contamination," Environ. Lett., 4, No. 4,
pp 259-265 (1974).
33. Van Loon, J. C., "Mercury Contamination of Vege-
tation Due to Application of Sewage Sludge as
Fertilizer," Environ. Lett., 6, No. 3, pp 211-218
(1974).
34. Watling-Payne, A. S., and Selwyn, M. J., "Inhibi-
tion and Uncoupling of Photophosphorylation in
Isolated Chloroplasts by Organotin, Organomercury
and Diphenyleneiodonium Compounds," Bioahem. J.
142, pp 65-74 (1974).
35. Westoo, G., "Determination of Methylmercury Com-
pounds in Foodstuffs. I. Methylmercury Compounds
in Fish, Identification and Determination," Acta
Chem. Saand., 20, pp 2131 (1966).
36. Westoo, G., "Determination of Methylmercury Com-
pounds in Foodstuffs, II. Determination of
Methylmercury Compounds in Fish, Egg, Meat, and
Liver," Acta Chem. Saand., 21, pp 1790 (1967).
37. Westoo, G., "Determination of Methylmercury Salts
in Various Kinds of Biological Material," Acta
Chem. Saand., 22, pp 2277 (1968).
38. Westoo, G., "Methylmercury as Percentage of Total
Mercury in Flesh and Viscera of Salmon and Sea
Trout of Various Ages," Science, 181, No. 4099,
pp 567-568 (1973).
39. Wong, Y. S., Chieh, P. C., and Carty, A. J.,
"Interaction of Organomercury Pollutants with
Biologically Important Sites. X-ray Study of 2:1
Complex Between Methylmercury and Penicillamine,"
Can. J. Chem., 51, No. 15, pp 2597-2599 (1973b)
\
40. Yamada, T., "Uptake of Phenylmercuric Acetate
Through the Root of Rice and Distribution of
Mercury in the Rice Plant," Nippon Nogei Kaishi,
42, No. 7, pp 435-439 (1968).
41. Zimmerman, P.W., and Crocker, W., "Plant Injury
Caused by Vapors of Mercury and Compounds of
Mercury," Contvib. Boyae Thompson Inatit., 6,
pp 167 (1934).
6
1-7
-------
u
u
a) ta
9 u
rH a
JS w
Ja co
U i-l
0) h
* a
-H
u ai
J3 s
no G
•rt o
« St
¦i
a
&
24
22 +
20
18
16
14
12
10
8
Figure 1.
3.5 4.5 5.5 6.5 7.5 8.5
pH of Incubating Solution
Effect of pH of the phosphate buffer incu-
bation medium on the transformation of
ionic mercury to methylmercury in Little
Marvel Peas as determined by methylmercury
peak height.
A » LM peas 13 days old; 0 ™ peas 35 days
old.
Table 1. GERMINATION STUDIES OF LITTLE MARVEL PEAS IN VARIOUS CONCENTRATIONS
OF MERCURIC ACETATE OR MERCURIC NITRATE
HgOAc yg/g Hg(N03)2 yg/g
1
10
25
1
10
25
% Germination
76%
76%
68%
92%
88%
25%
Epicotyl length-av.
1.7 cm
1.6 cm
1.4 cm
2.2 cm
1.3 cm
1.6 cm
Root length-av.
3.4 cm
2.8 cm
2.1 cm
3.4 cm
2.6 cm
2. 7 cm
Epicotyl total wt.
0.93 g
0.96 g
0.62 g
1.62 g
0.84 g
6.0 g
Root total wt.
0.96 g
0.90 g
0.52 g
1.43 g
1.7 g
7.1 g
Cotyledon total wt.
12.35 g
15.87 g
16.6 g
14.1 g
16.13 g
15.8 g
CH.Hg Present
NDa
NDa
NDa
NDa
ND3
ND3
^ot detectable
Table 2. METHYLMERCURY FORMATION FROM LITTLE MARVEL PEAS GROWN 14 DAYS IN
10 and 100 yg/g MERCURIC NITRATE AND PHENYLMERCURIC ACETATE CONTAMINATED SOIL
Mercury Concentration Weight of Methylmercury present
contaminated in soil aerial tissue ng/g fresh weight
Hg(N03)2
10 yg/g
12.9 g
3.1 ± 0.9 ng/g
Hg(N03)2
100 yg/g
23.6 g
7.6 t 2.3 ng/g
Ph-Hg-OAc
10 yg/g
29.1 g
3.8 ± 1.1 ng/g
Ph-Hg-OAc
100 yg/g
24.1 g
2.1 ±0.6 ng/g
7
1-7
-------
Table 3. WEIGHTS OF ALASKA PEA INTERNODES USED IN
INCUBATION EXPERIMENT WITH Hg(NC>3)2 AND
THE AMOUNT OF CH3HgCl FORMED PER GRAM OF
TISSUE IN 90 HOURS
Stem
internode
Total
weight
CHjHgCl
ng/g tissue
0-3
14.6 g
3.4 + 1
4
15.0 g
3.3 ± 1
5
17.8 g
2.5 * 0.7
6
14.2 g
2.3 ± 0.7
Apical
8.1 g
10.9 t 3.3
6 LITTLE MARVEL STEMS AND APICES INFILTRATED
AND INCUBATED 20 HOURS WITH 10 ugjg Hg(N0,),
AND CONTROL PLANTS WITH NO ADDITION OF
MERCURY TO INCUBATION SOLUTION
Tissue Total CHgHgCl
weight ng/g tissue
Stems
22.3
g
7.3 t 2.2
Apice8
3.6
g
4.1 ± 1.2
Control
27.8
g
Supernatant None detected
Pellet None detected
FORMATION OF METKYLMERCURY AS INFLUENCED BY
SURFACE STERILIZATION AND GENTAMICIN
TREATMENT
treatment
Age of
peas
Weight
of stems
CH3HgCl formed
ng/g tissue
Chlorox
18 days
10 g
7.6 ± 2.3
Chlorox
18 days
10 g
5.6 ± 1.7
Gentamicin
38 days
10 g
4.0 ± 1.2
No Gentaaicin
38 days
10 g
4.0 t 1.2
Coi>c. Chlorox
None present
Table 5.
8
1-7
-------
MONITORING: THE TRIGGER FOR ACTION
Dr. I. Eugene Wallen
Office of Toxic Substances
Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
Vinyl chloride air emissions ... organics
in drinking water ... asbestos fibers in waste
water effluents ... polybrominated biphenyls
in fish ... these and many other toxic chemi-
cal concerns are being clarified by current
monitoring efforts of the Office of Toxic Sub-
stances. In each of these instances, the
speed and direction of the responses of Gov-
ernmental regulatory machinery are critically
dependent on monitoring data — particularly
recent data which can provide a reasonably
sound basis for generalizations.
Two of the key informational concerns in
addressing chemical problems are toxicity and
exposure, and monitoring data provide the
backbone of exposure estimates. While toxici-
ty data may often be inadequate, acceptable
monitoring data are rarely available at stand-
ard-setting time.
What are the problems inhibiting the ready
availability of solid monitoring data? Why
are the available data generally not well
suited for standards setting? Why are we con-
tinually surprised by unanticipated chemical
problems? Who has the responsibility for
addressing currently unattended problems? Who
has the capability? Who has the resources?
These questions go to the heart of this con-
ference. Their answers inevitably involve a
blurry mix of scientific, technical, organiza-
tional, and budgetary issues. Perhaps a brief
review of a few recent case histories, togeth-
er with examples of current efforts, will help
to provide a perspective for the applications
of monitoring data in the regulatory decision
process.
Examples of Recent Toxic
Substances Monitoring Efforts
Vinyl Chloride
Initial EPA monitoring activities were
directed primarily to air emissions at PVC
polymerization facilities. Upon discovery of
the relationship of vinyl chloride to Angio-
sarcoma, monitoring patterns, sampling and
analysis approaches and equipment, and data
interpretation techniques were hastily devel-
oped by several EPA laboratories in early 1974.
A preliminary sampling effort at one facility
provided much needed experience in refining
the approaches to be used in more refined
efforts at seven other industrial complexes.
The entire sampling and analysis program was
carried out in less than three months, and
there is no doubt that the results had a major
impact on the EPA decision to set an air stand-
ard for vinyl chloride. Details of the ap-
proach and the results are set forth in the
report Preliminary Assessment of the Environ-
mental Problems Associated with Vinyl Chloride
and Polyvinyl Chloride, Environmental Protec-
tion Agency, September 1974. In retrospect,
this ad hoc monitoring effort was probably
about as good as could have been expected given
the time and resource constraints and the lack
of previous experience with vinyl chloride.
One major shortcoming was the failure to con-
vene a meeting of all the sampling and analy-
sis teams early in the program to identify
problem areas and promote greater uniformity
in the approaches that were used.
The recent discovery of vinyl chloride in
drinking water and near waste disposal sites
has caused a reexamination of the initial con-
clusions that these were not areas of serious
concern. Initial monitoring data, coupled
with tests of^the behavior of vinyl chloride
in water, suggested that there was little
likelihood that vinyl chloride would reach
drinking water supplies as the result of vinyl
chloride effluent discharges. With regard to
landfills, it was assumed that wastes from PVC
plants and discarded PVC products would proba-
bly not be problems. Monitoring of selected
sites hopefully would clarify the need for reg-
ulatory actions other than current steps di-
rected to air emissions.
A third type of concern is related to the
possible migration of vinyl chloride out of
finished PVC products into the environment,
thus possibly establishing a background levels
of the vinyl chloride monomer in air. While
skeptics have doubted this possibility, there
are several unconfirmed and unexplained reports
of detection of vinyl chloride distant from
either vinyl chloride or PVC plants. Sampling
for vinyl chloride is planned at a number of
potential sites where monomer migration might
have taken place.
Organics in Drinking Water
The report Preliminary Assessment of Sus-
pected Carcinogens in Drinking Water, Environ-
mental Protection Agency, June 1975, detailed
the initial results of an EPA monitoring sur-
vey of organics in drinking water. Eighty
water supplies were chosen and six specific
organics of potential concern ... the four tri-
halomethanes (chloroform, bromodichloromethane,
dibromochloromethane, and bromoform), carbon
tetrachloride, and 1,2-dichloroethane ... were
found. These 80 supplies provided a reasonably
representative sample of community drinking
water supplies that chlorinate their water,
and represented a wide variety of raw water
sources, treatment techniques, and geographi-
cal locations. Based on the survey findings,
it appeared that chlorination may have contri-
buted to the formation of the four trihalometh-
anes.
For a more comprehensive survey of the
organic content of finished water, a Second
portion of the Survey investigated 10 of the
80 cities with sites representing 5 major
categories of raw water sources. Preliminary
analyses of the drinking water of the first
five cities have identified over 85 organics.
The water supplies of the five remaining cities
2-1
-------
were surveyed more recently, and the results
are expected to be reported by the end of the
year.
The results of this monitoring effort are
having a major impact on the regulatory and
research efforts of EPA. Indeed, it has pro-
bably been the most important single activity
in shaping future Agency actions concerning
organics in drinking water.
Environmental Levels of Asbestos
A nation-wide asbestos sampling program
is underway to determine the environmental
levels of asbestos resulting from discharges
from various sources. Thirty-two sampling
locations have been chosen to include four
major categories of asbestos dischargers. A
natural site category was selected because of
asbestos rock formations which may contribute
significant amounts of asbestos in run-off or
emissions as a result of natural weathering
processes. Other site categories are asbestos
mining; mining of other ores (such as talc
and vermiculite which may also be sources of
asbestos); and asbestos manufacturing. For all
categories, both air and water samples are
being taken. Over 60 sampling sites have been
chosen, including water supplies of several
major cities, such as San Francisco, Denver,
Chicago, Atlanta, and Dallas. The sampling
has begun and a report detailing the analyses
should be available by mid 1976.
Previsouly Unidentified Pollutants in Water
As a part of the early warning efforts of
the Office of Toxic Substances, 200 water sam-
ples from sites in industrial areas through-
out the country are being collected and ana-
lyzed. It is anticipated that collections
will be made at about 100 sites in 30 to 40
industrial areas. Each sample will be sub-
jected to three subsample analyses to screen
for (1) inorganics, (2) volatile organics, and
(3) non-volatile organics. The purpose of
this monitoring program is to identify speci-
fic chemicals in surface waters having an in-
dustrial effluent origin, and then to clarify
which should be of concern in view of their
potential to reach man through drinking water,
fish products, or recreational exposure. The
monitoring information, together with data on
the toxicity of the identified chemicals,
should provide a basis for initial judgments
on the possible need for regulatory actions.
Multimedia Exposure to Halogenated Organics
A fifth program involves monitoring stu-
dies to determine the levels of selected halo-
genated organics in air, water, soil, and food
in four limited geographic areas, and parallel
investigations to determine the levels of the
same chemicals in the blood, tissue, and urine
of human populations in these areas. A review
will be undertaken of the causes of death in
these areas in comparison with death rates in
other areas, as well as with national averages.
The sources of environmental discharges of the
chemicals also will be identified.
This project is intended to utilize moni-
toring, epidemiological, statistical, and oth-
er analytical techniques in clarifying the de-
gree to which different types of pollutant
sources may contribute to human health pro-
blems .
The Use of Monitoring Data
in Setting Standards
An estimate of likely exposure should, of
course, be central in deciding the need for,
character of, and impact resulting from a pro-
posed regulatory action to curtail the dis-
charge of a chemical into the environment.
Current environmental levels, fluctuations,
and trends are of interest in assessing the
benefits to the environment of curtailing ex-
posure, as well as in detailing the costs in-
volved in achieving reduced exposure levels.
In grappling with the costs of regulation, —
for example, in setting maximum contaminant
levels in drinking water, allowable chemical
concentrations in water effluents, or pesti-
cides tolerances — the severity and urgency
of the problem, the numbers of activities
which would be affected by different types of
limitations, and the feasibility and costs of
compliance are all critically dependent on
monitoring data. More often than not, the
available data are so inconclusive that these
questions are only answered in a very cursory
way. However, as economic concerns over regu-
latory actions heighten, the insistences by
the Congress and the courts as well as by the
Executive branch of the Government that hard
evidence back up these actions increase.
Review and Evaluation of
Existing Monitoring Data
Sources of monitoring information gener-
ally were prepared for purposes other than reg-
ulation of specific chemicals. When these
data are compiled as a part of the considera-
tion of proposed regulation, they must be re-
assembled and reformatted so as to contribute
to calculations of the exposure of man and the
environment to a particular substance. Some
of the considerations used in such calcula-
tions must include the relative importance of
the alternative pathways from the environment
to man or other affected organism; the geo-
graphic sites of exposure; the internal varia-
bility of the affected populations; the effects
of acute or chronic rates of exposure; the per-
sistence of the chemical; the toxicity of its
breakdown products; and other similar charac-
teristics. Although monitoring data may be
used to gain significant insight into several
of these considerations, they generally must
be selected according to application.
Monitoring information to be gathered will
include concentrations and recent trends in
air, surface and drinking water, groundwater,
soil, food, sediment, aquatic and terrestrial
organisms, and human tissues and body fluids.
Details will also be needed on the methods of
collecting the samples, interferences, meteor-
ological data, and analytical methods employed.
Already existing monitoring data will
come from computerized searches, specialized
data centers, university programs, federal
programs, state programs, and science informa-
tion exchanges. These data must be keyed to
the specific types of limitations necessary
to reduce discharges into the environment,
thus reducing the risks of environmental ex-
posure. Data will be collected on a highly
2-1
-------
selective basis, with no necessity of obtain-
ing comprehensive monitoring data for all
sources and all media.
Data Gaps
It will be necessary to identify data
gaps and develop an appropriate environmental
sampling plan designed to yield the data re-
quired. Major air, water, soil, sediment,
and biota monitoring programs may have to be
undertaken. In identifying suitable sampling
sites, pertinent meteorological, hydrogeolo-
gical, and demographic considerations, along
with identification and evaluation of the var-
ious sources of discharge, must be considered.
Acceptable criteria must be developed with a
defensible rationale for selecting the sites
and the types of sampling to be undertaken.
It often may be desirable to establish season-
al trends of toxic substances levels at var-
ious locations and the concentration levels
in different media. In some cases cross-media
transport mechanisms must be considered.
Sampling Equipment
A wide spectrum of sampling equipment
must be available and used to collect and pre-
serve the integrity of the samples. Concen-
trations of the sample may-be required in the
field for some toxic chemicals. The analytic
laboratory must be sufficiently equipped to
perform quantitative trace chemical analyses
in various media. Readily available instru-
mentation should include atomic absorption
spectrophotometer, spark source mass spectro-
graph, high and low resolution mass spectro-
meters, gas chromatograph/mass spectrometer,
high resolution infrared spectrometer, UV-vi-
sible spectrometer, gas chromatograph, and
associated items.
Data Package
The data must be fitted into integrated
packages that will profile human and environ-
mental exposure to these chemicals in the
media of concern, and assess the degree of
harmful exposure of animals, plants and
microorganisms from the production, use and
disposal of these chemicals. The profile must
be keyed to the available approaches for re-
ducing environmental levels by controlling
sources.
Relevant data are required on the levels
of the chemical which are identified in air,
drinking and surface water, groundwater, dust,
sediment, soil, and other media. Reported
sources of contamination should be indicated
and, for those substances that occur naturally,
background levels should be identified. Maps
should be provided showing the distribution
of sampling areas and levels identified wher-
ever possible.
Data also are needed concerning (1) be-
havior of the chemicals in the environment;
(2) occurrence of the chemical in food and
other products that come in contact with man;
(3) exposure and biological accumulation; and
(4) environmental trends.
In conclusion, although monitoring data
may be collected for research, demonstration,
management, or other purposes, properly
massaged data are useful to the regulator in
a variety of contexts and they may trigger or
obivate several types of regulatory considera*-
tions.
3
2-1
-------
MONITORING VINYL CHLORIDE
IN THE VICINITY OF
POLYVINYL CHLORIDE FABRICATION PLANTS
Lawrence Elfers
PEDCo-Environmental Specialists, Inc.
Cincinnati, Ohio
and
Harold Richter
Environmental Protection Agency
National Environmental Research Center
Raleigh, North Carolina
Summary
Because vinyl chloride is a known carcinogen,
the U.S. EPA is investigating the need for ambient
air quality standards for this pollutant and
similar materials. Under contract with EPA,
(Contract No. 68-02-1375, Task No. 20) PEDCo-
Environraental Specialists, Inc. conducted a
program of ambient air monitoring near six plastics
fabrication plants in the eastern United States.
Five of the plants use polyvinyl chloride (PVC) in
their manufacturing processes; the sixth was
investigated as a 'control' study.
PEDCo designed the custom-built samplers for the
project. Four sampling sites were established for
each plant, representing four directions from the
plant whenever possible, with two samplers placed
at the site that is normally downwind. Sampling
was done in the spring of 1975, over a 2-week
period at each plant. Meteorological data concurrent
with the sampling period were obtained from nearby
agencies or other sources. A local operator was
trained at each plant location to perform 24-hour
integrated sampling in accordance with procedures
outlined in EPA's "Tenative Methods for Determining
Vinyl Chloride in the Atmosphere."
Analysis was performed at PEDCo-Environmental
laboratories under a stringent program of quality
assurance, including use of NBS reference samples,
analyses of duplicate samples by PEDCo and an
independent laboratory, and comparative calibration
of sampling flow rates.
Vinyl chloride concentrations at two of the
PVC fabrication plants and at the 'control' plant
were below the detection limits of 0.5 parts per
billion. At the other three plants, vinyl chloride
concentrations generally ranged from 0.5 to 5 ppb,
levels considered to be relatively low. The
highest concentration measured in the study was 7
ppb. Current knowledge of health-related effects
of exposure to vinyl chloride indicates that at
concentrations measured in this study the PVC
fabrication plants are not significant hazardous
emissions sources.
Introduction
Recent research has demonstrated that vinyl
chloride is a carcinogen. Further, this substance
has been detected in the atmosphere near vinyl
chloride monomer plants and polymerization plants.
As a result of these discoveries, the U.S. EPA is
investigating the need for ambient air standards
for vinyl chloride and similar materials.
The presence of vinyl chloride in the atmosphere
near plants that fabricate polyvinyl chloride
(PVC) products has not yet been demonstrated. If
it is present in the atmosphere in significant
amounts, the responsible agencies in states where
such plants are located may need to modify the
State Implementation Plans (SIP) to require atmo-
spheric surveillance in the vicinity of the plants.
Tn order to determine the magnitude of vinyl
chloride concentrations near PVC fabrication
plants, EPA contracted with PEDCo-Environmental
Specialists, Inc. to undertake a program of ambient
air monitoring.
The monitoring program, conducted in the
spring of 1975, encompassed six sampling areas in
the eastern United States. Five of the sampling
areas were in the vicinity of PVC fabrication
plants. The sixth was a "control" area near a
plant that fabricates plastic products but does
not use PVC. Locations for sampling were selected
by the Air Surveillance and Monitoring Branch of
the Environmental Protection Agency. Sampling at
each plant was conducted for 14 consecutive days,
during which integrated 24-hour samples were
collected on activated charcoal adsorption tubes.
The samples were shipped to the PEDCo-Environmental
Laboratory and analyzed for vinyl chloride content
within 48 hours of sampling. Descriptions of the
sampling sites, the techniques of sampling and
analysis, and quality control procedures are given
in this report with a summary of results. No
attempt is made to evaluate the health-related
implications of the data.
Sampling Locations
Ford Company Vinyl Plant
The Ford Company's PVC fabrication plant is
located at the outskirts of Mount Clemens, Michigan
(population 21,000). Residential areas adjoin the
plant property on three sides. Meteorological
data for this sampling period were obtained from
Selfridge Air National Guard Base located 3 miles
east of the plant.
1
2-2
-------
The Ford plant manufactures vinyl upholstery
material for the automobile industry. Overall
production of PVC during the 14-day sampling
period was estimated at 35 percent of capacity.
During this time approximately 360,000 pounds of
PVC plastics were producted by three processes:
suspension polymerization, emulsion polymerization,
and solution polymerization.
S.ylvania Electric Corporation
The Sylvania plant is located in the industrial
riverfront area at Warren, Pennsylvania (population
13,000). Industrial properties border the plant
on two sides. The residential areas lie to the
northwest, north, and northeast. Although no
reliable meteorological data were available for
this sampling area, observations by the field
operator indicated that winds during the sampling
period were generally northeasterly or southwesterly.
The Sylvania Electric Products plant at
Warren manufactures wire and cable products.
Since it does not use polyvinly chloride plastic in
the manufacturing process, this plant serves as a
control for the study. During the sampling
period the plant was operating at 80 percent of
capacity.
Congo!eum Company
The Congoleum plant is located in Marcus Hook,
Pennsylvania, one of the small but highly industrial-
ized towns along the Delaware River south of
Philadelphia. Industrial and residential properties
are intermixed in the vicinity of the plant.
Meteorological data for the area were obtained
from a system located at the main gate of the Sun
Oil Company, 1 mile south of the Congoleum plant.
The Congoleum plant at Marcus Hook manufactures
flooring material. Rates for PVC usage were not
available from the plant management. Overall
production rates at the plant during the sampling
period were estimated at 80 percent of capacity.
Reynolds Metals Company
The Reynolds Metals plant is located at the
south edge of the town of Grottoes, Virginia
(population 1,200). The area is primarily rural.
Meteorological data were obtained from the Shenandoah
Valley Airport 3 miles west of the plant.
The Reynolds Metals plant at Grottoes produces
plastic film for the packaging industry. Overall
production for the sampling period was estimated
at 40 percent of capacity. During this time
approximately 295,000 pounds of PVC was produced
by the suspension polymerization process.
Charlotte Plastics
The Charlotte Plastics plant is located in a
rural area 4 miles northwest of Monroe, North
Carolina. Meteorological data were obtained from
Douglas Airport at Charlotte, North Carolina, 25
miles northwest of the sampling area.
The Charlotte Plastics plant at Monroe fabricates
plastic pipe and fittings. During the sampling
period approximately 2 million pounds of PVC was
used in production operations. Plant management
did not indicate the production process.
Royal Electric, Division of ITT
The Royal Electric plant is located in the
northeast section of Pawtucket, Rhode Island
(population 88,000). Residential areas lie on
three sides of the plant. Adjacent to the west
side is a small industrial property. Meteorological
data were obtained from an air pollution monitoring
station of the Rhode Island Department of Health.
The station is located at Providence, 5 miles
southwest of the sampling area.
The Royal Electric plant manufactures rubber-
and vinyl-coated wire and cable. During the
sampling period approximately 176,000 pounds of
PVC was used in production operations. Plant
management did not indicated the production process.
Sampling and Analysis
Sampling Equipment
Vinyl chloride sampling was conducted with a
specially designed sampler, custom-built for the
project. The sampler incorporates a vacuum pump
to maintain a vacuum greater than 0.5 atmosphere,
measured by means of a vacuum gage. A critical
orifice consisting of a 27-gage hypodermic needle
controls the sample air flow at approximately 175
ml/min. A borosilicate glass absorption tube (18-
in. long, 10 mm 0D) filled with activated charcoal,
provides the collection medium. This tube is
connected to the vacuum source by an appropriate
fitting.
In this sampling program a rotameter, calibrated
at 25°C and 760 mm Hg with a positive displacement
type calibrator (traceable to an NBS standard),
was used to measure the sampling rates before and
after each 24-hour sampling period. A tubular
shield was provided to protect the absorption tube
from light and rain during sampling. The sampler
was normally mounted on a wooden stand with the
sampler inlet 5 feet above the ground. This
sampling configuration incorporated all operational
parameters recommended in the EPA publication,
"Tenative Methods for Determination of Vinyl
Chloride in the Atmosphere."
Sampling Procedure
Sampling was conducted at four sites near
each plant, located wherever possible to provide
exposures in four directions from the plant. In
all, five samplers were operated in each plant
survey, since two were placed at the site in the
prevailing downwind direction to provide duplicate
samples for use in PEDCo's quality assurance
program. Specific location of sampling sites was
determined on the basis of prevailing winds,
proximity to the plant, availability of electric
power, and security.
Sampling was performed at each plant by a
local operator, who was instructed in operation of
the sampling devices and in sampling procedures (a
printed operators guide was also provided). The
operator visited each of the four sites at the
same time each day; the sampling time at each, site
was thus as close to 24 hours as possible. After
installing a new adsorption tube each day, the
operator recorded the vacuum gage reading, tube
number, flow rate, and exact time. He recorded
similar readings at the end of each sampling
period before removing the adsorption tube. The
2
2-2
-------
exposed adsorption 'tubes were capped and stored in
a light-tight container. On alternate days the
operator shipped the 10 exposed adsorption tubes
in special padded boxes by air mail to the PEDCo-
Environmental laboratory for analysis.
Analytical Procedures
All standards were prepared by the procedure
recommended by the National Institute of Occupational
Safety and Health, method P&CAM 127. Analytical
procedures followed EPA's "Tentative Method for
the Determination of Vinyl~Chloride "in the Atmosphere
by 24-Hour Integrated SamplTivgT"
Preparation of Standards. Vinyl chloride
standards were prepared by dissolving 99.9 percent
vinyl chloride gas in spectrophotometric-grade
carbon disulfide. A 100-ml gas sampling bulb was
flushed and filled with vinyl chloride gas and
allowed to equilibrate to room temperature and
pressure. A 250-microliter aliquoit was withdrawn
with a gas-tight syringe and injected into a vial
containing a known volume of carbon disulfide.
This solution was used as a master stock.
Fifty microliters of the master stock solution
was injected into a second vial containing a known
volume of carbon disulfide. The concentration of
vinyl chloride in this solution was calculated
(corrected to STP) and this solution was used as a
working standard.
Initially, standards were prepared in the
range of 0.001 mg vinyl chloride per microliter of
solution. Since these standards tended to degrade
rapidly, standards were intentionally prepared at
lower concentrations (=0.0003 mg/microliter solution).
New standards were prepared for each set of analyses.
When these were compared with the original low-
level standards, no significant change was observed
in the concentration of the original standard
during the 4-month period of this study. The
standard was stored at 0°C in a Teflon-sealed vial
when not in use.
Desorption of Vinyl Chloride from Charcoal.
A 25-ml graduated cylinder fitted with a ground
glass stopper was used as the desorption vessel.
Twenty-five milliliters of carbon disulfide was
pipetted into the cylinder, which was then stoppered
and cooled to 0°C in a refrigerated bath. An
exposed charcoal absorption tube was opened, the
glass wool plug removed from one end, and the
contents shaken as rapidly as possible into the
cylinder, which was restoppered and swirled In the
ice bath to remove the heat of reaction from the
charcoal contacting the solvent. This process was
repeated for all of the exposed sample tubes. The
cylinders were left in the bath for 1 hour to
desorb the charcoal.
After 1 hour the tubes were removed and
agitated; a portion of the contents was transferred
into a vial and sealed with a Teflon-lined septum.
This aliquot of the sample was then used for analysis.
Analytical Method. A 2.5-m glass column, filled
with 0.4 percent Carbowax 1500 on Carbopak A, was
used with a Perkin Elmer model 990 gas chromotograph
equipped with a flame ionization detector. Helium
was us'fed as carrier gas, inlet pressure was 60 psig,
and flow rate, 73 ml per minute. Detector temperature
and injection block temperature were maintained at
150°C. Column temperature was maintained at 70°C
and the column was operated isothermally. Approxi-
mately 3 microliters of sample was injected for
analysis. Figure 1 is a typical chromatogram.
CARBON DISULFIDE
£ *
S o
Y
lABBON DISULFIDE
Y
I
:ABB0N DISULFIDE
I
IARB0N DISULFIDE
V
UL
INJECTION
INJECTION INJECTION INJECTION
Y
v_L
Figure 1. Chromatogram for the determination of
vinyl chloride.
Desorption Efficiency. Desorption efficiency
was determined by injection of metered concentrations
of vinyl chloride gas into blank exposure tubes
under controlled conditions of exposure. The
tubes were then desorbed by the procedure described
and analysis done by the same gas chromatographic
technique. The desorption efficiency for 28 samples,
calculated from the amount of vinyl chloride
injected versus the amount recovered, was determined
to be 89 percent.
Quality Assurance
PEDCo undertook a rigorous quality assurance
program to ensure the validity of data generated
during this project. The program was designed and
maintained to indicate any analytical error and
thus to provide a base for evaluation of the data.
NBS Reference Samples. Charcoal absorption
tubes identical to those used for collection of
samples were prepared by the National Bureau of
Standards and supplied to PEDCo-Environmental by
the EPA for quality control checks. These tubes
were spiked with vinyl chloride^in the range of 12
and 60 yg. These randomly selected quality control
samples were analyzed along with each batch of
samples from the field by the same procedure. In
comparison of values in the 12-yg range, the
average value determined by PEDCo on 14 samples
was 11.4 ug versus 11.9 ug reported by NBS, an
average difference of 4.4 percent. In the 60 ug
range, the average value determined by PEDCo on 13
samples was 61.4 ug versus 63.4 ug reported by
NBS, an average difference of 3.2 percent.
3
2-2
-------
Duplicate Samples. At one site in each study
area two samplers were operated simultaneously. Upon
receipt of the duplicate samples, PEDCo analyzed
one by standard procedures. If vinyl chloride was
detected, the duplicate sample was sent to the EPA
Project Officer for analysis by an independent
laboratory. Evaluation of data from analysis of
nine sets of duplicate tubes by PEDCo and the
independent laboratory showed an average difference
of 0.9 + 0.6 yg. One set of values for which the
difference exceeded the expected range was not
included in the statistical evaluation.
Calibration of Flow Rates. A third phase of
PEDCo's quality assurance program entailed checking
the flow rate through the 27-gage hypodermic needle
used as the critical orifice in the sampler. Six
needles were selected at random and given to the EPA
Project Officer for comparative evaluation. Flow
rates determined by EPA were compared with those
determined by PEDCo, all values corrected to 760 mm
Hg and 25°C. The average difference, based on an
average flow rate of 200 cc/min, was 1.5 percent.
Results and Conclusions
Results
Vinyl chloride concentrations detected in the
vicinity of PVC fabrication plants generally ranged
from <0.5 ppb to 5 ppb, considered to be relatively
low levels. The maximum concentration for any
single 24-hour period was 7 ppb. Vinyl chloride was
detected at three of the six study areas. Levels
measured at the Sylvania Electric plant (the 'control'
plant) and at the Charlotte Plastics plant and the
Royal Electric plant were below the detection
limit of 0.5 ppb.
Ford Vinyl Plant. During most of the sampling
period, vinyl chloride was not detected at the Ford
plant. On one sampling day, however, the highest
value obtained in this study was observed. A
vinyl chloride concentration of 7 ppb was observed
at sampling site B which is downwind of the plant
(see Figure 2),
Congoleum Company. During the first 6 days of
sampling at the Congo!eum plant, vinyl chloride
was twice detected in the range of 1 ppb. On the 8
days following, vinyl chloride was detected during
every 24-hour sampling period. Measurements of vinyl
chloride obtained at the three sites farthest from
the plant ranged from 1 to 3 ppb. Values obtained
at the site nearest the plant (100 feet) were higher,
ranging from 3 to 4 ppb (see Figure 3).
Reynolds Metals. On seven days in the middle
of the sampling period, vinyl chloride was detected
during each 24-hour period at the Reynolds plant.
The levels were consistently in the 1 to 2 ppb
range at each site (see Figure 4).
FORO PAINT
PLANT
LITTLE LEAGUE,, FREQUENCY
-s BALL PARK
\ AVE.
FORD VIN"
PLANT
'WATER TANK
ST. JOSEPH HOSPITAL
JONES ST.
HURBARO AVE.
CALM 1-5 6-10 U-?0 21
WIND SPEED mplt
Figure 2. Meteorological data at the Ford Vinyl plant
for February 19, 1975, Warren, Michigan.
Conclusions
This study demonstrates the presence of vinyl
chloride in the ambient air in the vicinity of many
polyvinyl chloride fabrication plants. Concentrations
measured in the study were relatively low, ranging
from less than 0.5 ppb to approximately 5 ppb. At
present, no Ambient Air Quality Standards for
vinyl chloride have been promulagated by the
Environmental Protection Agency. In the absence
of a standard, evaluation of known health-related
effects of exposure to vinyl chloride indicates
that PVC fabrication plants are not significant
hazardous emission sources.
4
2-2
-------
i FREQUENCY
0 300 600
SOSl! 1
CALM 1-5 6-10 11-20 21«
WIHCSntP mnh
1 In. - 600 ft.
CAIM 1-5 6-10 11-20 21
WINS SPEED mph
% FREQUENCY
600 ft.
Figure 3. Meteorological data for the period March 18
through March 25, 1975 at the Congoleum
plant, Marcus Hook, Pennsylvania.
Figure 4. Meteorological data for the period March 19
through March 25, 1975 at the Reynolds
Metals Company, Grottoes, Virginia.
5
2-2
-------
environmental levels OF PCB's
Doris J. Ruopp
Vincent J. DeCarlo
U.S. Environmental Protection Agency
Washington, D.C. 20460
ABSTRACT
Since the 1966 discovery in Sweden that
chlorinated biphenyls were widely dispersed in the
environment their environmental levels have been
the subject of many studies with the results
indicating that PCB's can have adverse ecological and
toxicological effects. However, a well planned
national approach to environmental sampling has
not been attempted thus hindering making a national
assessment of the PCB problem. This paper reviews
the levels that are currently being found in the
environment.
INTRODUCTION
Since the 1966 discovery in Sweden that
chlorinated biphenyls were widely dispersed in the
environment, their levels have been the subject of
many studies both abroad and in the United States.
Extensive laboratory and environmental measurements
indicate that PCB's can have adverse ecological and
toxicological effects, are . very resistant to
environmental degradation and are being discharged by
many different sources.
This paper will focus on the current PCB data
base in an effort to assess the PCB levels in the
environment on a national basis.
ENVIRONMENTAL PCB SOURCES
Before the PCB data base is examined, it would
be of interest to review how these chemicals get into
the environment. The major sources contributing to
the environmental levels of PCB's are industrial
effluents, manufacturing processes, consumer and
industrial waste materials, sewage treatment
facilities and accidental spills. Lacking extensive
air and soil data from these sources, available water
data are shown in Table 1. The 5 major PCB spills in
1973-75 all involved transformers. In 4 of those
incidents, the PCB's were spilled on soil while in the
fifth the transformer was dropped on a pier and the
contents were spilled into the water. In the latter
incident, 283 gallons were spilled and it was
estimated that only 70-90 gallons were recovered.
In all cases the material that could be recovered
was drummed up and entombed at a cost approaching
$2.3 million.
REVIEW OF THE PCB DATA BASH
To assess the scope and extent of PCB's in
the environment, all national surveys and national
data bases maintained by the U.S. Environmental
Protection Agency (EPA) were examined, including
STORET, the national water quality storage and
retrieval system, the National Soils Monitoring Program
for Pesticide Residues and the Human Monitoring
Program. In addition, data from special data bases
maintained by the United States Geological Survey and
the open literature were also examined along
with available unpublished reports. Using these
data, maps were constructed showing the
concentrations of PCB's in urban soils and the
aquatic environment.
The data from the national surveys show that
a great deal of effort has been expended and continues
to be expended but that the data are very limited.
On examination, all States are believed to show some
level of PCB contamination. In Table 2 States
which had significant levels in at least one
medium are listed with their reported PCB levels for
surface and ground water, bottom sediments and fish.
In Table 3, a number of localized studies are listed
along with the reported environmental data. In all
these studies the most extensive measurements have
been in fish.
TABLE 1
ENVIRONMENTAL SOURCES OF PCB's
SOURCE
PAPER MILLS
EFFLUENT WATER CON-
CENTRATIONS (ppb)
Wisconsin
WASTE PAPER MILLS
Wisconsin
INDUSTRIAL EFFLUENTS
California2
2
Wisconsin
Ohio2 3
Michigan
MUNICIPAL WASTE WATER TREATMENT
PLANTS
Michigan
Wisconsin
Ohio2 2
California
CAPACITOR AND TRANSFORMER
FACILITIES ,
New York ^
Massachusetts
PCB MANUFACTURING FACILITY
East St. Louis
SPILLS
0.1 - 18.5
18.5
0 - 76
.04 - 0.25
0 - 17
.1 - 7000
0.5
.05 -
10
0.16 -
17.0
17
76
2800
42.5
.87
Variable
1.
2.
3.
5.
Kleinert, S. J., Environmental Status of
PCB in Wisconsin, May 8, 1975, Wisconsin
Department of Natural Resources.
PCB's and the Environmental, Report of the
Interdepartmental Task Force on PCB's,
National Technical Information Service, 1972.
Statement of Concerns of the Lake Michigan
Toxic Substances Committee Related to Poly-
chlorinated Biphenyls, June 1975. Prepared by
Karl F. Bremer, USEPA, Chicago, 111.
Unpublished Data - Royal J. Nadeau and Robert
P. Davis, Investigation of Polychlorinated
Biphenyls in the Hudson River, Hudson Falla-
Ft. Edward Area, August 1974.
Field Sampling and Analysis of Toxic Pollutants
Interim Report, Battelle, Pacific Northwest
Laboratories, August 1974.
2-3
-------
WATER
On the basis of the national aquatic
environmental data collected in 1971-1972 and in
1974, a continuing widespread accumulation of PCB's in
water, sediment* and fish appears to he occurring.
However, no trend analysis is possible with the
available measurements. For example, whole water
measurements have been taken throughout most of the
country but those states reporting non zero
readings are few in relation to the number of states
showing zero concentrations. This is due both to the
low solubility of PCB's and to the usual analytical
procedure that limits detectability to the 0.1 ppb
level. More meaningful water concentrations were
obtained in the Lake Ontario1 and Orange County,
California,2 studies where concentrations in water at
the ppt (ng/1) level were measured. At these
levels, changes in PCB concentrations could be
found with distance and time and could be
related to other measurable parameters such as PCB
concentrations in sediment, flora and fauna.
SEDIMENTS
Since PCB's are relatively insoluble, it is
not surprising that bottom deposits have shown
significant concentrations. Although 30 states
collected samples in the 1974 study, 15 had less than
4 stations reporting and showed zero readings. Of
the remaining 15, with at least 4 reporting
stations, 13 showed detectable concentrations.
Since the J.3 states involved were not necessarily the
same in both studies, only a very broad comparison
may be made, i.e., five states had lower
concentrations, four states were higher and four did
not sample again.3 We may conclude, then, that in
those states with any monitoring effort in the 1974
study, PCB's are present in bottom deposits and the
levels are not any lower than in the 1971-72 study.
Considering the persistence of PCB's, a significant
proportion of the nation's waters are now affected
and will continue to be.
SOILS
The National Soils Monitoring Program is a
small sampling effort studying only 5 urban areas
each year. However, PCB's were detected in three of
the five cities sampled in each of the years for
which data are available, 1971-73, Of the 22
positive readings 17 of them were below 1 ppm. Of
the cities sampled in 1973, Pittsfield, Mass., was of
particular interest because it has a large transformer
and condenser plant using large quantities of PCB's.
Six different sampling sites within one mile of
the plant show no detectable PCB residues. In
contrast, the soil surrounding another facility in
Illinois using PCB's in the manufacture of investment
casting waxes was recently measured for PCB content.
Samples taken in an area one-quarter mile radius
around the plant ranged in value from 0.77-5.2 ppm.
The PCB's identified were mixtures of Aroclor 1260 and
decachIot obipheny1. Levels up to 1.8 ppm Aroclor
1260 were found at 1-1/4 miles from the plant.
Samples were collected from this facility as part
of a study being conducted for the Office of Toxic
Substances at sites suspected to have PCB
concentrations. Other sites sampled in March of 1975
were in the vicinity of an investment casting
company in Michigan and the PCB manufacturer in
Illinois. Surface soil samples were collected at
each site up to a distance of approximately one mile
from the plant boundary, at 1/4 mile intervals in all
directions. PCB's were detected in soils at both
sampling locations. Concentrations ranged from
the detection limit of 0.001 ppm to over 20 ppm.
The distribution of all PCB's analyzed appears higher
near the plant site and generally decreases with
distance from the slte.^ Details of the sampling
sites and the concentration levels measured are shown
in Figure 1.
TABLE 2
SELECTED STATE DATA ON ENVIRONMENTAL LEVELS OF
PCB's
State
Surface and
Groun^
Water
Ug/1
Bottom FishJ
Sediment ppm
Ug/kg
AL
AR
CA
CO
CT
FL
GA
IL
IA
MD
MA
MI
MN
MS
SB
NJ
NY
OH
OR
PA
PR
SC
TX
UT
VA
WV
WI
0.1
0.3
0.1-0.2
0.1-2.1
0.1
0.2
0.1-0.3
0.1
0.1-4.0
0.2
0.1
0.1-3.0
0.1
20-2,400
20-190
5-3502
5-3,200
10-1,300
10-1,200
50-170
3-800
3-13,0002
1-5-140
6-7002
30-200
7.9-290
5-80
10
1.53-5.48
1.69-3.88
2.16-5.34
0.10-1.25
0.52-1.18
1.21-11.3
0.35-1.41
0.56-1.31
4.00-11.7
0.44-1.09
0.47-4.58
0.10-4.00
2.68-9.50
1.73-8.07
0.71-3.62
1.94-2.48
0.10-7.3
0.10- .22
0.15-2.14
0.31_1.20
1.24-14.8
1. All data in this column taken from Hans J.
Crump-Wiesner, Herman R. Feltz and Marvin L.
Yates, A Study of the Distribution of Poly-
chlorinated Biphenyls in the Aquatic Environ-
ment , Jour. Research U.S. Geol. Survey JL,
603 (1972, unless otherwise noted.
2. USGS Sediment data, 1974.)
3. All data in this column taken from Croswell
Henderson, Anthony Inglis and Wendell L. Johnson,
Organoehlorlne Insecticide Residues In Fish
Fall 1969 National Pesticide Monitoring Program,
Pesticide Monitoring Journal 5^ 1(1971).
2
2-3
-------
1,
2
3,
4
5,
6
7,
8,
9,
10
11
12,
TABLE 3
SELECTED PCB STUDIES
Fish
Water
Sediment
LAKES
Lake Ontario!. 2
Lake Erie^, 4
Lake Superior5
Lake Huron®
Lake Michigan®, 9
Cayuga LakelO
Lake St. Clair^-®
RIVERS
Iowa
Mississippi Riverll
Texas
Brazos River^
Utah
Green River
New York
Hudson
Michigan
Saginaw River® »
Kalamazoo River7•15
Portage Creek7
Wisconsin
Milwaukee ®
MARINE ENVIRONMENT
Atlantic Ocean1'> 18
Bay of Fundy-*-®
Gulf of Mexico/Carib-
bean Sea^0> 21
Southern California^
0.14
- 9.17
0.035 - 0.097
43
- 245
0.2
7.8
0.3
5.6
0.16
- 11.0
0.1
- 165.3
3.72
- 636
3.4
9.7
tr
3.8
0.1 - 0.2
0.9
- 5.5
Plankton
(PP°) ..
3.4 - 11.8
6.9
<0.01
- 350
- 165.3
- 164.56
0.1
0.1
1.0
0.1
0.1
0.1
.02
- 0.5
- 2800
- 0.21
- 0.45
- 0.48
- 2.17
6.6 - 6,700,000
100 - 560
10 - 475,400
0.038 -
0.07 -
0.032 -
0.04 -
0.190
1.54
0.059
6.6
0.0009 - 0.0036
0.0015 - 0.019
0.157 - 1.055
Clarence L. Haile, Gilman Veith, G. Fred Lee and
William C. Boyle, Chlorinated Hydrocarbons in
the Lake Ontario Ecosystem, EPA-660/3-75-002,
June 1975, U.S. E.P.A., Corvallis, Oregon.
Klaus L. Kaiser, Mirex, An Unrecognized Contami-
nant of Fishes from Lake Ontario, Science 18,
523 (1974).
Richard L. Carr, Charles E. Finsterwalder and
Michael J. Schibi, Pesticides Monitoring Jour-
nal 6:23 (1972).
John R. M. Kelso and Richard Frank, Organochlorine
Residues, Mercury, Copper and Cadmium in Yellow
Perch, White Bass and Smallmouth Bass, Long
Point Bay, Lake Erie, Trans. Amer, Fish. Soc.,
103:577 (1974).
Unpublished data. Gilman D. Veith and G. E. Glass,
PCB's and DDT in Fish from Western Lake Superior,
U.S. E.P.A., Duluth, Minnesota.
Unpublished Data - John L. Hesse, Bureau of Water
Management, Michigan Department of Natural Re-
sources, June 1973.
Unpublished data - State of Michigan Water Re-
sources Commission, Polychlorinated Biphenyl Sur-
vey of the Kalamazoo River and Portage Creek in
the Vicinity of the City of Kalamazoo, 1972.
Gilman D. Veith and Fred G. Lee, Chlorobiphenyls
in the Milwaukee River, Water Research, 1971.
Unpublished data - Schacht 1974, EPA 600/3-74-002.
Carl A. Bache, James W. Serum, William D. Youngs
and Donald J. Lisk, Polychlorinated Biphenyl
Residues: Accumulation in Cayuga Lake Trout
with Age, Science 177:1191 (1972).
Lauren G. Johnson and Robert L. Morris, Chlorin-
ated Insecticide Residues in the Eggs of Some
Freshwater Fish, Bulletin of Environmental Con-
tamination and Toxicology 11:503 (1974).
Jean A. Schulze, Douglas B. Manigold and Freeman
L, Andrews, Pesticides in Selected Western
Streams 1968-1971, Pesticides Monitoring Jour-
nal 7:73 (1973).
13. Unpublished data - Royal J. Nadeau and Robert P.
Davis, Investigation of Polychlorinated Bl-
phenyls in the Hudson River (Hudson Falls - Ft.
Edward Area, August 1974.
14. Unpublished data - John L. Hesse, Monitoring for
Polychlorinated Biphenyls in the Aquatic Envi-
ronment, May 1973.
15. Unpublished data - Michigan Water Resources Com-
mission, Evaluation of the Aquatic Environment,
of the Kalamazoo River Watershed, May 1972.
16. Michigan Department of Agriculture, Bureau of
Consumer Protection, 1973 Great Lakes Environ-
mental Contaminants Study.
17. T. F. Bidleman and C. E. Olney, Chlorinated Hyrdo-
carbons in the Sargasso Sea Atmosphere and Sur-
face Water, Science 183:516, 1973.
18 George R. Harvey, Helen P. Miklas, Vaughan T. Bowen
and William G. Steinhauer, Observations on the
Distribution of Chlorinated Hydrocarbons in
Atlantic Ocean Organisms, Journal of Marine Re-
search 32:103 (1973).
19. V. Zitko, 0. Hitzinger and P.M.K. Choi, Contamin-
ation of the Bay of Fundy - Gulf of Maine Area
with Polychlorinated Biphenyls, Polychlorinated
Terphenyls, Chlorinated Dibenzodioxins and Di-
benzofurans, Environmental Health Perspectives,
1:47 (1972).
20. C. S. Giam, A. R. Hanks, R. L. Richardson, W. M.
Sackett and M. K. Wong, DDT, DDE, and Polychlor-
inated Biphenyls in Biota from the Gulf of Mex-
ico and Caribbean Sea - 1971. Pesticides Moni-
toring Journal 6:139 (1972).
21. C. S. Giam, M. K. Wong, A. R. Hanks, W. M. Sackett
and R. L. Richardson, Chlorinated Hydrocarbons
in Plankton from the Gulf of Mexico and Northern
California, Bulletin of Environmental Contam-
ination and Toxicology 9:376 (1973). '
22. The Ecology of the Southern California Bight:
Implications for Water Quality Management,
Southern California Coastal Water Research Pro-
ject, 1500 East Imperial Highway, El Segundo,
1973.
3
2-3
-------
Figure 1
Concentration of PCB's in Soil with Distance
(ppm)
Aroclor 1260
Investment Casting Company
T«
Aroclor 1242
PCB Manufacturer
Aroclor 1260
Investment Casting Wax Manufacturer
tN
Aroclor 1260
PCB Manufacturer
decachlorobiphenyl
Investment Casting Wax Manufacturer
decachlorobiphenyl
PCB Manufacturer
4
2-3
-------
AIR
PCB measurements in air along with the
related transport studies are practically
nonexistent. Harvey and Steinhauer-' have reported
levels ranging from 3.9 - 5.3 ng/m3. Air
measurements by Bidleman and Olney for Rhode Island
ranged from 2.1-9.A ng/m3 and from 0.21 - 0.65 ng/m3
at Bermuda.6 Snow melt water data from Wisconsin
ranged from zero to 0.24 ppb. The suggestion has
been made that atmospheric fallout may be the most
significant source of PCB discharge to the waters of
the state of Wisconsin. 7
TABLE 4
LAKE ONTARIO ECOSYSTEM1
FISH, ug/g
Alewife
.14
- 4.36
Smelt
1.40
- 3.49
Slimy Sculpin
1.58
- 9.17
WATER, ng/1
38
- 97
SEDIMENT, ng/g
43
- 245
NET PLANKTON, pg/g
3.4
- 11.8
FISH
Fish have been studied on a nationwide basis by
the Department of the Interior since 1967. In the
1969 study, PCB levels were identified in fish
from 35 states. ® Data from the nationwide sampling
programs, 1970-1973, are currently being prepared
but were not available for this report.
Consequently, nationwide fish data are not
available any more currently than 1969, with the
exception of the isolated studies listed in Table 3,
The 1969 national study showed. PCB levels generally
within the FDA 5 ppm guideline. In the Great Lakes
area, PCB levels in fish from Lake Michigan were so
high, 7.6-10.9 ppm, that FDA seized shipments of coho
salmon in May 1975.^
Although transport mechanisms are not well
known and would vary through different ecosystems,
it is interesting to consider the Lake Ontario
Ecosystem Study summarized in Table 4. From this study
an estimate of biomagnification is possible giving
a sediment to fish ratio of 1:120. The FDA guideline
of 5 ppm in the edible portion of fish
corresponds to a sediment concentration of 41 ppb -
a figure exceeded in all 13 states reporting through
USGS in 1974,
HUMANS
With PCB's as widespread across various media
as they are, it is expected that levels would be
identified in humans as well. The Human Monitoring
Program, in 1972, found that 3035 out of 4102 samples
from 31 States showed levels ranging from less than 1
ppm to more than 3 ppm. In 1973, 964 out of 1277
samples from 28 States again showed levels in the same
range. In both years approximately 75% of the
adipose tissue analyzed contained some
polychlorinated biphenyls. Unfortunately we do not
know age, occupation or residential histories of the
cases involved so it is difficult to trace the levels
back to the potential sources.
CONCLUSIONS
Although there are many national and state
groups collecting environmental PCB data,
limitations in the current data base prevent one from
making a uniform national assessment of PCB
environmental levels. It appears that if a well
planned national approach to sampling were
attempted, the data base could be improved in a very
short period of time.
1. Haile, C. L., Veith, G. D., Lee, G. F.,
Boyle, W. C., Chlorinated Hydrocarbons in
the Lake Ontario Ecosystem (IFYGL) June,
197^.
REFERENCES
1. Haile, C.L., Veith, G.D., Lee, G.F., Boyle, W.C.,
Chlorinated Hydrocarbons in the Lake Ontario
Ecosystem (IFYGL), June 1975.
2. McDermott, D.J., and Hansen, T.C., Inputs of DDT,,
PCB and Trace Metals from Harbors, Coastal Water
Research Project Annual Report, 1975.
3. U.S. Geological Survey.
4. Environmental Polychlorinated Biphenyl Contamination
near Sites of Manufacture and Use, Environmental
Science and Engineering, Inc., 1975.
5. Harvey, G.R. and Steinhauer, W.G., Atmospheric
Transport of Polychlorinated Biphenyls to the
North Atlantic, Atmospheric Environment, Vol. 8,
1974.
6. Bidleman, T.F., and Olney, C.F., Chlorinated Hydro-
carbons in the Sargasso Sea Atmosphere and Sur-
face Water, Science, Vol. 183, February 1974.
7. Kleinert, S.J., Environmental Status of PCB's in
Wisconsin, May 8, 1975, Wisconsin Department of
Natural Resources.
8. Henderson, C., Inglis, A., Johnson, W.L., Organo-
chlorine Insecticide Residues in Fish - Fall 1969
National Pesticide Monitoring Program, Pesticides
Monitoring Journal, Vol. 5, No. 1, June 1971.
9. Bremer, Karl E., Draft copy of "Statement of Con-
cerns of the Lake Michigan Toxic Substances Com-
mittee related to PCB's," June 1975.
5
2-3
-------
HALOGENATED HYDROCARBONS AND THE ENVIRONMENT
John A. Zapp, Jr.
Haskell Laboratory for Toxicology and Industrial Medicine
E. I. du Pont de Nemours & Company
Wilmington, Delaware 19898
Summary
The chemical and physical properties of
halogenated organic compounds make them ex-
tremely useful in a variety of applications.
Among the most important of the properties are
their relatively high solubility in and for
fatty materials, and their relatively low
solubility in water. These same properties
lead to a tendency for accumulation in the
fatty materials in the environment, which in-
clude the fatty tissues of living organisms.
Because of the bio-magnification which occurs
in food chains, the higher organisms are par-
ticularly prone to accumulate in their fatty
tissues levels of the halogenated organic
compounds which are far higher than the
initial input into the environment would sug-
gest.
5.
The halogenated organics entered the
historical scene rather late, although
chlorine'and fluorine occur to the extent of
0.2% and 0.1% respectively in the earth's
crust while bromine and iodine occur to only
0.001% each. The halogens do not, however,
occur to any significant extent in nature as
the elemental molecules, CI2, F2, Br2 or I2.
Scheele, in 1774, prepared chlorine gas by
heating a mixture of hydrochloric acid with
manganese dioxide and called it "dephlogisti-
cated marine acid air." It was not until 1810
that Davy identified it correctly as elemental
chlorine. Fluorine was the last of the halo-
gens to be prepared in elemental form, and
this was in 1886 by Moissan.
Scheele had noted when he prepared
chlorine in 1774, that the gas bleached paper
colored with litmus; that it bleached green
vegetables, and red, blue and yellow flowers
nearly white. As chlorine became commercially
available, its first use was for bleaching of
paper and textiles. Some enterprising French
candle maker then discovered that chlorine
would also bleach candles. However, when the
candles bu: aed they gave off smoke and choking
fumes. The chemist Dumas investigated and,
in 1834, reported that chlorine had reacted
with the wax, displacing some hydrogen and
substituting chlorine. When halogenated
hydrocarbons burn, if at all, they give off
smoke and choking fumes. If enough hydrogen
is replaced with chlorine, they will not burn.
It can be seen from the candle experi-
ment that at least some halogenated organics
are easy to make, and from this we can deduce
that failure to find significant quantities of
the halogenated hydrocarbons in nature is due
to tha relative absence of the elemental halo-
gens in nature by the time that carbon com-
pounds appeared on the scene. Once chlorine
became available and it was recognized that it
could react with organic compounds, chemists
set out to deliberately prepare chlorocarbons,
and many of these turned out to have very use-
ful and attractive properties. So much so
that today the largest fraction of chlorine
produced is used for the manufacture of
chlorocarbon compounds.
The preparation of industrially impor-
tant halogenated carbon compounds dates from
about the beginning of the 20th Century. An
acceleration occurred after World War I when
liquid chlorine became available in quantity
and a second acceleration occurred during
and after World War II, associated with the
rise of the chlorinated hydrocarbon pesti-
cides, polyvinyl chloride, and the fluoro-
carbon refrigerants and propellants.
In 1973, total U. S. production of
acyclic halogenated hydrocarbons was about
22.6 billion pounds. These would include the
chlorinated solvents like trichloroethylene,
dichloroethane, chloroform, vinyl chloride
monomer and the fluorocarbon propellents and
refrigerants.
The chlorinated benzenes accounted for
about half a billion pounds in 1973 and the
chlorinated pesticides about 0.2 billion
pounds in 1973. Preliminary U. S. Tariff
Commission figures for 1974 and 1975 show a
drop from 1973. Some of this is probably due
to economic recession and some to a reaction
away from the use of some halogenated organics
because of environmental considerations.
I have been careful not to say that
there are no halogenated organics found in
nature. TKere is reason to believe that some
may be formed in the ocean and some in the
atmosphere, but not in economically or indus-
trially important amounts. Two antibiotics
which contain chlorine, Chloromycetin and
chlortetracycline, were isolated from soil
microorganisms. Two plant species are known
to produce ou-fluoroacids.
An important reason for our current
interest in the halogenated organics is their
relatively recent appearance on the industrial
scene, the practical utility of many of these,
and the growth in production during this cen-
tury from almost zero to some twenty billion
pounds per year in the United States alone.
The reason for this rapid rise in the
production and use of the halogenated carbon
compounds lies in their physical and chemical
characteristics. They have a greater affinity
for fatty and oily materials tnan they have
for water. Some make excellent solvents, such
as dichloromethane, chloroform, carbon tetra-
chloride and some of the chlorobenzenes.
2-4
-------
The first of these solvents was pro-
bably prepared in 1795 by four Dutch chemists
who reacted ethylene, a gas, with the green-
ish gas later shown to be elemental chlorine
and got a liquid of good solvent properties
which they called "Dutch oil." It was later
shown to be dichloroethane, and is still one
of the most widely used. Trichloroethylene
and perchloroethylene found wide acceptance
for degreasing and drycleaning applications.
Chloroform, or trichloromethane, was one of
the early surgical anesthetics. Carbon
tetrachloride and tetrachloroethane are per-
haps the best chlorinated hydrocarbon sol-
vents, but are seldom used for this purpose
anymore because of their toxicity.
A second virtue of the chlorohydro-
carbons is their relative inflammability com-
pared with the hydrocarbons. For practical
purposes, trichloroethylene and perchloro-
ethylene are non-flammable as well as excellent
solvents. Carbon tetrachloride once enjoyed
wide use as a fire extinguishant but is sel-
dom used for that purpose anymore because of
toxicity. CBrF^ is used today as a rapid
acting fire extinguisher in certain special
environments. Unlike carbon tetrachloride,
it has low toxicity.
It has been noted that it is not too
difficult to put halogens into organic mole-
cules. In many cases, it has not been too
difficult to replace these halogens with
other desired groups. Hence a halogenation
step may be intermediate in many organic
syntheses.
Of course, solvent power, non-flam-
mability, and reactivity are not the only
reasons for the rapid rise in production of
the halogenated carbon compounds. Some mem-
bers of the class had unique properties which
made them particularly useful. For example,
prior to World War II, the number of avail-
able pesticides was quite limited. During
World War II, the pesticidal efficacy DDT
was discovered and it proved to be so valua-
ble for the protection of troops against
louse-borne disease that all information
about it was officially classified. Success
spurs further research and it was soon found
that the benzene hexachlorides (BHC) also
possessed potent pesticidal activity. Tech-
nical BHC consists of a number of isomers, of
which the gamma isomer, known as Lindane,
proved to be the most useful and it soon dis-
placed technical BHC as a commercial pesti-
cide. These two, DDT and Lindane, were
enthusiastically used as civilian pesticides
after World War II, and they effectively dis-
placed most of the older pre-war pesticides
like lead arsenate and nicotine sulfate.
Another important class of pesticides, the
anti-cholinesterase compounds, also emerged
in the post World War II period, but these
were not halogenated hydrocarbons. According
to the report of the President's Science
Advisory Committee, "Chemicals and Health"
(1973) some 10 million people are now alive
who would have been dead were it not for DDT.
Other chlorinated hydrocarbon pesticides,
principally chlordane, heptachlor, aldrin and
dieldrin rapidly achieved economic importance
also.
Another very important class of halo-
genated hydrocarbons which developed since
1930 is the chlorofluorocarbons. Two members
of this class, fluorocarbon 11 or trichloro-
fluoromethane, and fluorocarbon 12 or di-
chlorodifluoromethane, were introduced in the
1930's as safe, non-flammable and relatively
non-toxic refrigerants. The production of
these compounds grew with the increasing use
of refrigeration and air conditioning, but
accelerated rapidly when these compounds were
discovered to be useful propellants for
aerosol products. Today, about 50% of the
total production of fluorocarbons goes into
aerosol propellants, about 28% into refri-
gerants, and the rest into other uses, such
as fire extinguishants, solvents and blowing
agents for foamed plastic.
Finally, one should mention vinyl
chloride monomer, used for the production of
polyvinyl chldride and sometimes as an
aerosol propellant.
It is obvious to this audience that
each of the main uses of the halogenated car-
bon compounds which I have mentioned, i.e.,
solvents, pesticides, propellants and plas-
tics, has raised questions of environmental
safety. Ironically, the environmental con-
cerns have arisen from the same properties
which made the halogenated carbon compounds
useful in the first place, i.e., their rela-
tively high solubility in and for fatty
materials, their relatively low solubility in
water, and their relatively good chemical
stability. I could add the word "toxicity,"
but I should like to speak to that a little
later. Indeed we should probably not be
greatly concerned about the toxicity of the
halogenated carbon compounds if it were not
for these other properties.
Because of high fat:low water solu-
bility the halogenated organics which enter
soil or water are not readily dissolved - at
least until they come in contact with fat.
And the most likely source of the fat is some
kind of living organism. If, for example,
lake water contains a small concentration of
a chlorinated hydrocarbon pesticide, and this
lake contains fish and the smaller organisms
which constitutes their food, the pesticide
will tend to concentrate in the fatty tissues
of these organisms and of the fish. As the
larger organisms eat the smaller, their own
fatty tissues receive a further increment of
the pesticide. The larger fish may then
become dangerous as a steady diet for birds
or mammals, including man.
That this phenomenon of bio-magnifica-
tion is predictable, did not prevent it from
happening. Once recognized, it was easily
explained. Again, one of the virtues of the
chlorinated hydrocarbon pesticides is their
persistence. If you wish to keep termites
out of your house, you would like to use a
pesticide that would stay active for, a long
period of time. Under other circumstances,
such_persistence could be a liability, result-
ing in an undesirable build-up in the environ-
ment, be it soil or water.
2
2-4
-------
When one considers the toxicity of the
halogenated carbon compounds, one finds that
they are all toxic to some extent, as are all
chemicals, but that individual halocarbon
compounds vary from practically non-toxic to
extremely toxic and very few general rules
apply. I shall discuss toxicity of the halo-
genated organics first from the point of view
of acute toxicity, and then from the point of
view of chronic or cumulative toxicity.
The acute toxicity of the volatile
chlorocarbon solvents is usually manifested
by the inhalation route. The vapors share
with those of other volatile organic sol-
vents the property of depressing the central
nervous system, i.e., narcosis. Early mani-
festations of this effect are dizziness,
headache, nausea, or sometimes just a plea-
sant feeling resembling alcoholic intoxica-
tion. A more severe exposure produces
anesthesia and unconsciousness. Actually,
chloroform was one of the first general
anesthetics used in surgery, and trichloro-
ethylene is still used in some places for
this purpose.
Chloroform was a good surgical anes-
thetic but for one thing. Patients sometimes
came out of anesthesia with a residual liver
damage, and it is characteristic of the acute
toxic action of the chlorohydrocarbons that
they can damage the liver and kidneys. The
extent of"damage is dependent on the indi-
vidual compound, on the dose, and on the
nutritional state of the exposed person.
If we consider the simple saturated
aliphatic chlorhydrocarbons, we find that
solvent power and narcotic potency increase
with increasing substitution of chlorine for
hydrogen. Carbon tetrachloride is the best
solvent of the methane series but also the
most toxic. Solvent power of the chloro-
ethanes seems to peak, at tetrachloroethane,
sometimes called acetylene tetrachloride, and
it is even more toxic than carbon tetra-
chloride. Hexachloroethane is a solid at
room temperature, and pentachloroethane has
too high a boiling point to make it very
attractive as a solvent.
In general, the unsaturated aliphatic
chlorohydrocarbons are less toxic than the
corresponding saturated members. Thus,
tetrachloroethylene is much less toxic than
tetrachloroethane, arid 1,2,2-trichloro-
ethylene is less toxic than 1,2,2-trichloro-
ethane. But 1,1,1-trichloroethane, or
methyl chloroform, is less toxic than
1,2,2-trichloroethylene. The distribution of
the chlorines in the molecule, as well as the
presence of double bonds, modifies the
toxicity.
A person who has been on a low protein
diet, perhaps through illness, tends to have
more fat in the liver than one who has been
on an adequate protein diet. The higher the
fat in the liver, the more it tends to
attratt and hold fat solvents. Hence sur-
gical patients in a poor state of nutrition,
were more apt to suffer liver damage after
anesthesia than well nourished patients.
Alcoholism also results in liver damage and
fatty livers. Hence alcoholics are more sus-
ceptible to liver damage following chloro-
hydrocarbon anesthesia than well nourished
patients.
Generally speaking the aliphatic
chlorofluorocarbons are less toxic than the
corresponding chlorocarbons and also show
less narcotic potency. Strangely, the un-
saturated aliphatic fluorocarbons, like
tetrafluoroethylene are more toxic than the
corresponding saturated members.
The halogenated hydrocarbon vapors
share with non-halogenated narcotics vapors,
the potential for causing cardiac arrhyth-
mias, which can become life-threatening in
the presence of excess epinephrine. Glue
sniffers and aerosol propellant sniffers
deliberately inhale high concentrations of
the vapors to reach a state of intoxication.
If, during induction of that state, they go
through an excitement stage which results in
the secretion of epinephrine, the heart may
go into ventricular fibrillation and death
ensues.
The chlorinated hydrocarbon pesticides,
e.g. DDT, produce, like the solvents, effects
on the central nervous system and on the
liver. They are probably attracted to these
sites for the same reason - lipoid solubility.
But whereas the chlorohydrocarbon solvents
depress the central nervous system, the
pesticides stimulate it, producing increased
activity and convulsions. The effects of the
chlorinated hydrocarbon pesticides on the
liver is similar to that produced by the sol-
vent but less marked. The pesticides kill
primarily through their GNS effects. On the
other hand, the polychlorinated biphenyls
(PCB's) and polychlorinated naphthalenes are
potent liver toxins but produce relatively
little overt effects on the central nervous
system.
I should like to mention vinyl chlorida
monomer only to the extent that its acute
toxicity is low, and its anesthetic potency
is comparable to that of ethyl chloride.
The chronic toxicity of the halogena-
ted hydrocarbons has caused more concern
than their acute toxicity. Because of their
high lipoid:low water solubility, they tend
to concentrate in the high lipoid organs,
brain, liver and depot fat. They will also
leave these organs at a certain rate and be
detoxified or eliminated from the body. If
the rate of daily intake exceeds the rate of
daily elimination, the amount retained in the
body will increase until toxic levels are
attained. The most usual chronic toxic
effect has been liver damage, and it has not
been very common in man in recent years be-
cause the chlorinated carbon compounds have
been studied, characterized with respect to
toxicity, and safe levels of exposure have
been established and have been generally
pretty well observed in industry.
With respect to the environment, the
situation has not been so well controlled.
Chlorinated carbon compounds have accumulated
in various places and at various times to the
extent that damage to wildlife was apparent.
-------
In some instances environmental damage
has resulted from uses which were inappropri-
ate. An example would be the use of a per-
sistent chlorinated hydrocarbon pesticide,
because it was cheaper, when a non-persistent
pesticide would have accomplished the task.
In other instances damage to non-target
species has occurred through failure to
anticipate biomagnification through the food
chain.
Above all, however, the specter of
cancer has arisen. It has been shown that
several of the halogenated hydrocarbon sol-
vents, including carbon tetrachloride,
chloroform, trichloroethylene, 1,2-dichloro-
ethane and 1,2-dibromoethane, can cause
liver cancer when fed to rats or mice at
relatively high dosage levels, i.e., dosage
levels Which cause liver damage aside from
cancer. The persistent chlorinated hydro-
carbon pesticides, DDT, aldrin, dieldrin,
chlordane and heptachlor, also in high doses
and particularly in mice, have been shown to
cause liver cancer. The PCB1s have similarly
been shown to cause liver cancer in mice
Vinyl chloride monomer was a surprise
in that it caused a very rare form of liver
cancer in rats and in man, i.e., angiosar-
coma, at quite low dosage levels which were
insufficient to cause the usual liver damage
produced by the chlorinated hydrocarbons.
As a result of these findings, a great
deal of public concern has arisen over the
detection of low concentrations of chlori-
nated hydrocarbon solvents, PCB's, and pesti-
cide residues in the environment. The ques-
tion being asked is: "Are these materials
in these low concentrations a cancer threat
to the public?"
There are those who believe that there
is no no-effect level of exposure for car-
cinogens. The EPA seems currently to be
holding this position. There are others,
including myself, who believe that there is
a no-effect level for carcinogens. Among
those who believe that a no-effect level
exists for carcinogens are the Department of
Labor, at least two expert committees of the
World Health Organization, and various can-
cer experts including Dr. Saffiotti of the
National Cancer Institute.
Considering that the chlorinated
hydrocarbons have been produced and used in
massive quantities, at least since World
War II, it seems unlikely that we should not
have noticed a sharp rise in human liver
cancer, i.e., primary hepatomas, as we have
noticed a sharp rise in human lung cancer
associated with the increase in cigarette
smoking since World War II. Yet, as cancers
go, liver cancer is considered uncommon. The
employees at the one plant in the U. S. still
manufacturing DDT have been thoroughly
studied. Compared to the population at
large, they carry a high body burden of DDT
but show no excess of cancer. A British
study of DDT applicators with heavy exposure
in India likewise showed no excess of cancer.
It is known that persons with cirrho-
sis of the liver have a higher than normal
incidence of liver cancer. Like alcohol, the
chlorinated hydrocarbons can, at high doses,
cause cirrhosis of the liver in man. Hence,
at high doses they could cause an excess of
liver cancer. The FDA has concluded that
alcohol and selenium do not cause liver cancer
at doses too low to cause cirrhosis. It may
be, therefore, that the cancer arises from an
aberration of the repair process in some
individuals. If so, we would have little to
fear from small residues of the chlorinated
hydrocarbons in the environment. This will
be an interesting field for continued
investigation.
In summary, the chlorinated carbon
compounds have been remarkably useful in a
number of areas in our century. The dictum
of Judge Learned Hand, that there is no value
to be obtained without some sacrifice to be
assumed, holds here as elsewhere. There are
risks as well as benefits. It would seem to
me that the public interest would best be
served by weighing benefits against risk in
each area of use, making sure that the bene-
fits and risks are real and not speculative,
and then using the chlorinated hydrocarbons
to our benefit, soberly and advisedly, with
due regard to what they can do for us as well
as against us.
4
2-4
-------
PREPARATION AND EVALUATION OF VINYL CHLORIDE STANDARDS
E. E. Hughes, W. D. Dorko, S. M. Freund and D. M. Sweger
National Bureau of Standards
Washington, DC 20234
Summary
An investigation has been completed of the
feasibility of preparation of gaseous stan-
dards of vinyl chloride in air. Both the
accuracy with which such standards can be
prepared and the long term stability have been
evaluated. The purpose of the investigation
relates to the possible future issuance of
such mixtures as National Bureau of Standards
Standard Reference Materials.
Introduction
There are two very important factors which
must be evaluated when preparing a gas stan-
dard of any sort. The first is the stability
of the mixture and the second is the accuracy
with which the components of the mixtures are
known. Obviously, a gas mixture in which the
component of interest does not remain at a
constant concentration is not suitable as a
standards The concentration of such a
component may change for several reasons. It
may react with the walls of the container or
it may adsorb on the walls. A reaction may
occur between the component of interest and
the other components of the mixture. In any
case an alteration of the composition may
occur. If reactions occur the concentration
will decrease. If adsorption occurs an initial
decrease may be observed followed by an
increase as the total pressure in the cylinder
decreases. The existence of such effects can
only be determined by experimental obser-
vations over a long period of time.
The accuracy with which a component of a
sample may be determined depends either on the
existence of an absolute method by which the
component can be measured or on the existence
of standards of known accuracy with which the
sample can be compared.
Vinyl chloride is not a very reactive mole-
cule and extensive reaction of vinyl chloride
mixtures contained in cylinders is not
expected. Further, the vapor pressure of
vinyl chloride is relatively high and as a
first approximation the extent of adsorption
would be rather slight. In terms of gas
mixtures in general, vinyl chloride would be
expected to behave in such a way that stable
standards could be prepared as gas mixtures
contained in cylinders.
If a gas mixture is stable primary standards
may be prepared by gravimetry, or alter-
natively, by pressure but with a greater
degree of uncertainty. The choice of either
of the methods depends on the accuracy
required of the final standard which in turn
is determined by the purpose of the analysis
and the precision with which the measurement
can be made. Gravimeteric standards can be
prepared with an uncertainty of less than
0.5 percent relative while pressure standards
generally can be prepared with an uncertainty
of about 0.5 to 1 percent relative. The
precision with which a sample can be compared
to a primary standard using gas chromatography
is about 0.2 to 0.3 percent relative. There-
fore, the total uncertainty assigned to the
analysis of a sample can be as low as *1
percent. Seldom are there analytical situa-
tions where such a high degree of accuracy is
necessary. Measurements of vinyl chloride in
the ambient atmosphere, for instance, need
not be performed with a high degree of
accuracy in order the assess the total atmo-
spheric burden of vinyl chloride. This results
from a situation where it is impossible to
make measurements at more than a few sites
and the uncertainties introduced into a calcu-
lation of total vinyl chloride because of lack
of knowledge of total air volume, air move-
ment, and site bias due to location, far
exceed the uncertainty of the standard. This
situation exists, unfortunately, in most
measurements whether in the ambient atmosphere
or within a plant manufacturing or using vinyl
chloride. Thus, while the preparation of
standards of the highest accuracy may represent
an intriguing analytical challenge it does not
represent the most effective utilization of
manpower nor does it recognize the very
real economic situation which exists.
For these reasons, primary standards prepared
by pressure are considered quite adequate for
the preparation of Standard Reference Materials
intended for analysis of ambient, work place,
and source levels of vinyl chloride.
The validity of a primary standard is not easy
to assess. If the stability of the standard
has been determined by experiment and poten-
tial sources of systematic errors have been
evaluated then a high degree of confidence
may be placed upon the. value calculated from
the pressure data. However, some confirmation
of the validity of the pressure data must be
obtained to eliminate the possibility of
blunders or similar sources of error. Inter-
comparison of standards covering a wide range
of concentration by an analytical method in
which the signal is linearly proportional to
concentration will establish the degree of
accuracy with which the standards were
prepared. Unfortunately, it is difficult to
establish instrument linearity without a
series of standards of known accuracy. If
two independent methods depending on different
principles are available both of which have
theoretical linearity then an observed linear
relationship between calculated concentration
and the signal produced will lend considerable
weight to the validity of the calculated
\
2-5
-------
concentration. However, it must also be
established that no other process or processes
are present which would alter the concen-
tration in some systematic manner that would
result in an unrecognized error.
Experimental
The experiments described below were intended
to evaluate the long term stability of mixtures
of vinyl chloride in air and to define the
accuracy which could be achieved in preparation
of mixtures by the relatively simple technique
of pressure blending. Long term stability was
assessed by observation of the concentration
of a number of samples over a period of time
of fourteen months. The accuracy was evaluated
by comparing the samples one with another by
two instrumental techniques both of which were
predicted to respond in a linear manner.
Samples were prepared by admitting a measured
pressure of analyzed vinyl chloride into a
small cylinder followed by addition of a
measured pressure of air. The manifold system
shown in figure 1 allows the pressure to be
measured to better than 0 . 3 percent. The
concentration of vinyl chloride is the ratio
of the measured pressure of vinyl chloride to
the total pressure corrected for deviations
from ideality of both components. The initial
mixtures are prepared from analyzed samples
of the pure components at concentrations of
about 1 percent and mixtures at lower concen-
trations are subsequently prepared by dilution
of this high concentration. A portion of this
mixture is admitted at a measured pressure
into a second cylinder followed by additon of
a measured pressure of air. Mixtures were
prepared to concentrations of 2 ppmv in the
manner shown in figure 2.
o-zooo PSI
0-1000 PSI
0-100 PSI
0-15 PSI
TO VACUUM
PUMP
Figure 1.
Gas manifold arrangement for
preparation of accurate gas
mixtures
VCL
I6.000PPMV
I
1000PPMV
50PPMV I00PPMV
I
2PPMV
I0PPMV
Figure 2. Stepwise dilution of vinyl chloride
and air.
Two separate sets of vinyl chloride pressure
standards at similar concentrations were
prepared. In addition, a set of propane in
air standards was prepared in which the
pressures of propane and air were almost
identical to the pressures of vinyl chloride
in air resulting in identical final concen-
trations. These were prepared for two
reasons. First, propane in air standards
can be prepared by the pressure technique
with a predicted accuracy of better than ±1
percent. This is illustrated in table 1,
The analyses on which the observed values
are based were performed against National
Bureau of Standards Standard Reference
Materials and with other standards of defined
accuracy. Thus, the effect of systematic
errors in the pressure measurement systems
may be recognized and if present a correction
to the vinyl choride standards may be applied.
Second, the long term stability character-
istics of propane in air mixtures are known
and comparison of vinyl chloride to propane
mixtures can serve as a means for determining
changes in concentration with time of the
vinyl chloride standards.
Table 1. Accuracy with which propane stan-
dards can be prepared by pressure
Sample Cone, in ppmv Observed conc.
number by pr^isure in ppmv
1
3.30 x 10"
3.25 x lO"
2
2.24 x 10"
2.24 x 10"
3
3312 .
3316.
4
522 .
524.
S
251.
250.
6
98.2
97.7
7
50.4
50.8
8
30. 3
30.5
9
24.8
24.8
10
19.4
19. 3
11
12.8
12.6
12
8.45
8.51
13
7.54
7.55
14
5.03
4.99
2
2-5
-------
Two methods of analysis were employed. The
first was conventional gas chromatography
while the second was a laser technique 1
described in detail elsewhere. The method
utilizes Stark modulated absorption of carbon
monoxide and carbon dioxide infrared laser
radiation by vinyl chloride as a sensitive
and selective detetor.
Experimental
Gas Chromatography
Both the vinyl chloride and propane standards
were analyzed by gas chromatography using a
6' x 1/8" Poropak Q column operated at 100 °C
and employing a flame ionization detector.
The stability of the vinyl chloride standards
was assessed by observing the signal generated
by vinyl chloride and comparing it to the
signal generated by a similar concentration of
propane. Accuracy was assessed by comparing
the magnitude of the signal produced by vinyl
chloride mixture at one concentration to that
produced at all other concentrations.
Assessment,of the accuracy requires a know-
ledge of the linearity of the particular
detector system ^hile assessment of stability
requires a knowledge of the relative sensi-
tivity of the detector system for the two
species, vinyl chloride and propane.
The samples, were analyzed over a period of
fourteen months with extensive comparison
performed at essentially zero time, four months
and fourteen months. In addition, a series of
commercially prepared mixtures was analyzed
shortly after preparation and again 10-11 months
later by comparison to the standards prepared
in this laboratory.
It is difficult to make a simple visual com-
parison of the data in order to arrive at any
conclusion concerning stability. However,
comparison of the peak area due to each species
at similar concentration levels will reveal
the presence or absence of trends with time
of the concentration of vinyl chloride. Table
3 is a comparison of the data, obtained simply
by dividing the signal due to vinyl chloride
by that generated by propane after adjusting
the signal for slight difference in concen-
tration.
Table 3. Comparison of response of vinyl
chloride to propane
Ratio of peak areas
Cone, of Propane/VCl
VC1 in ppmv 6/21/74 10/8/74 8/22/75
Results
Gas Chromatography
The results of analysis of the laboratory
standards are summarized in table 2. The
areas due to propane have been multiplied by
0.666 simply to facilitate comparison.
Table 2. Results of analysis of vinyl chlo-
ride and propane mixtures over a
14 month period
Cone, in ppmv
by pressure
2
Peak areas in divisions
6/21/74 10/8/74 8/22/75
Vinyl chloride
16680.
974.
965.
96.0
47.9
9. 38
1.98
16650
993.
98.2
48.2
9.52
1.95
8446.
501. S
8730.
511.9
47.50 49.49
23.31 24. 29
4.439 4.628
0.9371 0.9770
8426.
509.5
50. 33
24. 72
4.725
0.9968
Propane
8753.*
509.9
49.82
24.08
4. 649
0.9520
9027.
529. 3
52.25
25. 35
4.899
1.003
8800.
532.9
53. 32
25.89
4. 980
1. 002
all propane areas multiplied by 0.666
16680.
1.04
1. 04
1.04
974.
1.00
1. 01
96S.
1.02
96.0
1.02
1.03
1. 04
47.9
1.03
1.04
1.03
9. 38
1. 03
1.04
1.04
1.98
1.03
1. 04
1. 04
An obvious conclusion is that the ratio of
vinyl chloride to propane has not changed
signicantly over a period of fourteen months.
The minor variations at the different times of
analysis are undoubtably due to variations in
chromatographic conditions particularly to
difference in the rate of flow of hydrogen to
the burner of the flame ionization detector.
An example of this effect is shown in table 4.
Table 4. Effect of hydrogen pressure on
relative sensitivities
Ratio of peak areas
Propane/VCl
Cone, of
VC1 in ppmv
16680.
965.
96.0
47.9
9. 38
1.98
8.Opsig H2 9.5psig H2
1.04
1. 03
1. 04
1. 06
1.06
1.08
1.04
1.02
1. 04
1.03
1. 04
1. 04
The hydrogen flow is controlled by adjusting
the total pressure on the inlet side of the
system. All of the values shown in table 2
were obtained at a hydrogen pressure of 9.S
psig. The two sets of data shown in table 4
were obtained within a few days of each other
at hydrogen pressure of 8.5 and 9.5 psig,
respectively. The effect is obvious and
emphasizes the difficulty of performing accu-
rate quantitative analyses of multi-component
mixtures.
If the data of tables 2, and 3 can be inter-
preted to indicate inherent stability of vinyl
chloride mixtures then the concentrations
defined by the measurement of pressure during
compounding of the mixture should yield results
of predictable accuracy. If the measurement
system is linear, then the ratio of the signal
to concentration, often referred to as "sensi-
tivity" should be constant over the entire
range of concentration. The results in table
3 might be interpreted to indicate either an
error in preparation w:\th a maximum deviation
3
2-5
-------
in the vicinity of 1000 ppmv of vinyl chloride
or a variation in the response characteristic
of the particular chromatograph to vinyl chlo-
ride and propane. An error in preparation is
unlikely because the second set of vinyl chlo-
ride standards agree quite well with the first
set used to compile table 2. The sensitivities
based on measured areas and calculated concen-
tration agreed in all cases to better than
±0.4 percent of the amount of vinyl chloride
when a sample of one set was compared to the
same concentration level in the second set.
Thus, the only conclusion is that the instru-
ment does not respond linearily and that it
responds in a slightly different manner for
propane as compared to vinyl chloride. A gas
proportioning pump was adjusted to produce
several mixtures of propane in air ranging in
concentration from 5000 to 400 ppmv of propane
and the signal at each concentration was
measured. Mixture of vinyl chloride in air
were then prepared using the identical pump
settings as were used for propane. If the
instrument response characteristic is identical
for each substance then the ratio of the peak
area of propane to vinyl chloride will be a
constant. The results are shown in table 5.
Table S. Ratio of peak areas for four iden-
tical concentrations of vinyl chlo-
ride and propane produced with a
proportioning pump
Predicted conc. Ratio of peak areas
in ppmv Propane/VCl
5146. 1.040
1782. 1.034
900. 1.014
471. 1.022
The measurements, unfortunately, do not cover
the entire range of interest but they do con-
firm an obvious change in sensitivity for
either or both species and offer a reasonable
explanation for the differences in the ratios
at different concentration shown in table 3.
If the concentration of a gas sample in a
cylinder does not change over a long period of
time it is unlikely that any significant re-
action or adsorption has occurred on the
cylinder walls or in the vapor phase. If
neither of these processes occur then it is
even less likely that very rapid reactions or
adsorptions would occur when the sample is
first transferred into a cylinder. If neither
fast nor slow processes occur to alter the
sample then samples may be prepared by pressure
and the calculated concentration will closely
approximate the true concentration. The data
previously presented confirms the absence of
long term processes but does not exclude the
possibility of short term processes although
the consistency of the data would seem to
exclude any such processes.
Two samples of vinyl chloride in air at about
1000 and 50 ppmv were transferred from large
cylinders to small cylinders identical in size
and materials to those in which the vinyl
chloride-air standards had been prepared. The
samples were later compared and no difference
was observed between the original samples and
the transferred samples thereby eliminating
th^ possibility of any rapid reaction or
adsorption in the cylinders.
The negative information concerning instability
together with the agreement between the re-
sponse of the instrument to different concen-
trations of vinyl chloride indicates agreement
between the analysis and the concentration
based on the pressure mixing data. In spite
of the apparent non-linearity of the detecting
system it is obvious that the response is pro-
portional to the vinyl chloride concentration
and that samples prepared by this method are
comparable in accuracy to sample of propane
prepared by the same method.
Laser Detection
The recently developed laser technique for the
detection of vinyl chloride was evaluated in
part by comparing the infrared absorption by
various samples containing vinyl chloride.
The samples were prepared by pressure in the
manner described earlier. The concentration
range studied was from about 1000 to 2 ppmv.
The prinicple conclusion drawn from this phase
of the work concerns the linearity of the
relationship between calculated concentration
and signal. This is shown in figure 3 wfiere
the relative signal amplitude is plotted
against the concentration calculated from the
pressure data obtained during preparation of
the samples. The linearity, especially in the
concentration region above 10 ppmv and up to
1000 ppmv is excellent. There is no theoret-
ical basis for deviation from Beer's Law
behavior of the system in this region and ttie.
obvious conclusion must be that the concen-
tration assigned to the standards is correct.
This further confirms the absence of any
significant reaction or adsorption of the
mixture during or after preparation.
10-
0.1-
Figure 3. Analysis of vinyl chloride mixtures;
Log of the concentration vs. log
of the signal amplitude (^and +
sign signify CO and C02 lasers
respectively)
There is some evidence of deviation from
linearity of those mixtures at concentrations
below 10 ppmv. Later measurements with an im-
proved signal to noise ratio decreased this
4
2-5
-------
and is illustrated by figure 4. The phenom-
ena has been ascribed to memory effects, ad-
sorption of vinyl chloride on components of
the spectrometer, and is not unexpected. The
effect, moreover, is opposite in direction to
what would occur if the samples had degraded
in any way.
u
ic
9
8
e>
~ 7
<
4> &
>
o 4
«>
ce
3
2
I
Figure K. Analysis of vinyl chloride mixture:
Concentration vs. signal amplitude
to the low values and interpolation between
standards at the high values was necessary but
in spite of this good agreement between the
two sets was obtained indicating that the
samples were stable for the ten to eleven
month period.
~ Conclusions
Vinyl chloride standards in the range from
1.6 percent by volume to 2 parts per million
by volume can be prepared by the simple tech-
nique of pressure blending with a predictable
accuracy of about +1 percent of the amount
of vinyl chloride. The stability of such
standards is excellent over a period of at
least fourteen months.
Reference
1. S. M. Freund and D. M. Sweger, "Vinyl
Chloride Detection Using Carbon Monoxide
and Carbon Dioxide Infrared Lasers" Anal,
Chem. 47 930-32 (1975).
60
6
60
8
100
10
2
Conccnlrolipr ppm
The observed linearity of these results
support the conclusion concerning the chroma-
tographic results namely that the lower results
observed in the region of 1000 ppmv are due
to instrumental effects and not to errors in
the calculated concentration of vinyl chloride.
Analysis of Commercial Samples
Fifteen samples of vinyl chloride in nitrogen
were analyzed shortly after preparation and
again after a period of approximately eleven
months. The results are shown in table 6.
Table 6. Analysis of commercial mixtures of
vinyl chloride
Sample Initial conc.
number of VC1 in ppmv
1
874.
2
878.
3
40.4
4
40.6
5
40. 7
6
40. 2
7
40. 2
8
43. 5
9
0.88
10
0.88
11
0. 89
12
0. 88
13
0. 88
14
0. 89
15
0.91
Cone, of VC1 in ppmv
after 10-11 months
881.
883.
40. 7
40.8
40. 8
40.4
40.0
43.4
0. 89
0. 89
0. 89
0.88
0. 88
0.91
0.87
The analyses were performed by comparing the
samples with the standards using gas chroma-
tography only. Extrapolation of the standards
5
2-5
-------
IMPROVED METHODS OF SAMPLING
AND ANALYSIS OF VINYL CHLORIDE AND
OTHER GASEOUS CARCINOGENS
R. C. Lao, R. S. Thomas and J. L. Monkman
Technology Development Branch
Air Pollution Control Directorate
Environmental Protection Service
Environment Canada
Ottawa, Canada K1A 0H3
Summary
Attention has been focussed on the association be-
tween vinyl chloride (VC) exposure and human angiosar-
coma of the liver since December, 1973 when a diagnosed
case of this rare tumor was discovered during the au-
topsy of a VC worker from a plant in the U.S.A. manu-
facturing both polymer and monomer. Since then, inten-
sive epidemiological studies have been initiated in
many countries to ascertain the VC risk to human beings
occupationally exposed to VC or PVC, and to residents
in urban areas adjacent to production plants.
VC in air is sampled basically by two methods:
(1) grab, and (2) integrated. Grab sampling is useful
for the identification of point sources over short time
intervals and in real time analysis. For ambient air
sampling, the integrated method"is commonly used to
provide a time weighted average concentration of VC.
Investigations have been carried out in this laboratory
to provide a sampling system which is applicable to am-
bient and in-plant atmosphere. Computerized gas chro-
matographs equipped with automatic sampling system and
flow rate control are carefully calibrated with a re-
producibility in the sub-ppm range. Samples from vari-
ous sources have been analyzed.
Introduction
Because of increased public health and environ-
mental concern about the effects of vinyl chloride on
human health1' 2' 3, regulations have been promulgated
by several countries, ordering reduction in the concen-
trations which had been present in industrial and ambi-
ent atmospheres prior to such promulgation11. Compli-
ance with these various regulations demands straight-
forward, reliable analytical methods, which can be ap-
plied to a broad range of air concentrations and sam-
pling techniques. In addition, unit analytical time
should be very short because of the large number of
samples which have to be processed. Gas chromatogra-
phy, because of its fast turn around time, wide concen-
tration linearity and high sensitivity for lightweight
hydrocarbon molecules, has been selected as the analy-
tical finish for a variety of sample types, including
air, water, polymer resins, and solid polymer stocks,
filled or unfilled, plasticized or unplasticized5.
Although technology may well be available which
can reduce "fugitive" vinyl chloride emissions from
monomer manufacturing plants by as much as 90%, the
reduction of emissions from polyvinyl chloride (PVC)
plants and secondary fabricating industries Is a more
difficult problem which may entail complex process
changes6.
Since vinyl chloride is a gas at normal ambient
temperatures and pressures, it is ordinarily transport-
ed and stored as a liquid under pressure. The analysis
of vinyl chloride generally shows the presence of con-
taminants such as acetylene, methyl chloride, 1,3-buta-
diene and vinylidene in trace amounts or more6. The
raw polymer resins regularly contain entrapped or ad-
sorbed VCM in the ppm range, which may be released at
ambient temperatures and may be driven off at fabrica-
tion temperatures.
Apparatus
Two basic systems were used to produce the data
for this work. For qualitative-quantitative VCM deter-
minations, a Perkin-Elmer 3920 gas chromatograph (GC),
fitted with dual FID detectors and MS-41 solid sample
Injection system, was used. The output signal from the
GC was fed into a Perkin-Elmer PEP I data processor
which recorded retention times and peak areas. For
qualitative confirmation, the samples were assayed on a
Finnigan Model 1015D GC/MS quadrupole, equipped with a
series 6000 data system.
Both GC systems used 6 feet 1/8" stainless steel
columns, packed with Chromasorb 102, 80/100 mesh condi-
tioned to 200°C and a helium carrier gas flowrate of 30
ml/min. Injection and detector blocks were maintained
at 250°C to prevent condensation of heavy molecular
weight components. In addition, the MS-41 was main-
tained at 250°C to approximate industrial fabrication
conditions for calendering, extrusion and coating. The
dual flame GC carrier flows and detection systems were
adjusted and calibrated to give equivalent VCM response
through both channels. Flows were maintained using a
calibrated Perkin-Elmer Dial-a-Flow digital flow con-
troller. Careful attention to the instrumentation set
up allows calibration of the column coupled to the sol-
id probe by means of cross calibration with the second
chromatographic column using simple Injections of gas-
eous reference compounds. Figure 1 shows the response
in concentration curves for the two injection-column-
detector combinations.
Experimental
The 6-foot chromosorb packed column was evaluated
for VCM resolution on blends of compounds which might
potentially provide interferences when sampling indus-
trial atmospheres. Table I lists a numbar of fugitive
emission components, present in industrial atmospheres
with the retention times given under the analytical
conditions stated. In no case is there any interfering
coincidence between the elution time or retention vol-
ume data for these compounds and vinyl chloride.
In situations where categorical qualitative con-
firmation is required, the sample Is analyzed using the
Finnigan GC-MS system. Vinyl chloride is indicated by
three prominent ions appearing at m/e?=27, the molecular
ion at m/e=62 and the chlorine isotope ion at m/e""=64.
The ratio of the ion abundances for m/e™62 and m/e*=64
must be about 3:1 in accordance with the natural abun-
dance ratio of chloride isotopes.
The presence of VCM in carbon disulfide extracts
is easily determined using the same analytical system.
Good spectra may be obtained by injecting 2-3 micro-
litres of the CS2 extract onto the column and scanning
the mass spectra of the elution peaks, either in the
normal scanning or mass fragmentography mode. It is
important to subtract background contributions from
-------
column bleeding or residual air from the resultant
spectra of VCM before calculating the ion abundance
ratio for m/e*"62 and 64.
To evaluate the sensitivity of the GC/MS system, a
CS2 extract of "a reference compound was successively
diluted to lower concentrations and analyzed. Useful
analytical spectra were produced with 0.10 nanograms of
VCM injected into the GC. Using mass fragmentography,
this value could be lowered to about 0.025 nanograms,
with some deterioration of the signal to -noise ratio.
The collection and desorption efficiency of the
carbon tube CS2 extraction was evaluated to establish
the optimum reaction temperature providing highest
recovery. Ten gram aliquots of 20x50 carbon were used
to adsorb reference quantities of vinyl chloride in a
25 ml reactiflask fitted with mini-inert valves. Fif-
teen ml of CS2, refrigerated to -15°C, was added to the
flask and the system was allowed to reach equilibrium
for 15 minutes in a +15°C bath. Two to three micro-
litre injections of the CS2 containing VCM were
analyzed and comparative results for these desorption
conditions are listed in Table XI and shown graphically
in Figure 2. In no case did the head space concentra-
tion of VCM in the flasks exceed 2% of the concentra-
tion in the liquid. Other losses, at the lowest con-
centrations, may be attributed to syringe manipulation
and wall adsorption phenomena still undefined. Figure
3 shows a typical gas chromatogram of VCM in a CSz
extract.
Results and Discussion
The selection of Chromosorb 102 was based on its
ability to resolve vinyl chloride from among the com-
pounds prevalent in industrial atmospheres. A variety
of other column packings including 10% FFAP on 80/100
acid washed OCMS Chromosorb W and 10% SE-30 on 80/100
Chromosorb W were evaluated prior to the selection of
Chromosorb 102. The single phase column provided a
simple uniform packing with high reproducibility of
retention times, column after column. This fact
becomes increasingly important as columns become
plugged by high molecular weight degradation products
through continuous use of the solid sample probe. A
short, replaceable stripping column, packed with
Chromosorb 102 and placed ahead of the analytical
column prevents delays between analyses.
The recovery data, for carbon disulfide extraction
of VCM from activated carbon, demonstrate that with
care and correct temperature selection, recoveries in
excess of 90% may be expected for samples containing
more than 10 nanograms of vinyl chloride. Below this
concentration, factors such as polymerization on the
activated carbon or adsorption on the walls of the
reaction vessel may well account for the drop in
recovery efficiency.
Entrapped vinyl chloride monomer in both PVC resin
and fabricated stocks has been analyzed although the
VCM is evidently not distributed homogeneously through-
out the polymer. An appropriate sample injection
system is required. The correlation between the
concentration of the entrapped VCM and head space VCM
is shown in Table III. For higher values of entrapped
monomer, the spread in the concentration distribution
is more pronounced than for the lower values. In
addition, an equilibrium phenomenon appears to be
established with entrapped monomer concentrations below
15 micrograms per gram. Below this concentration
value, the head space concentration remained constant
while the entrapped VCM concentration is reduced. This
phenomenon is only noticed for the raw resins in
granular form.
Bias - 100 *
Difference = 200 x
Analyses were performed on charcoal tubea supplied
by the EPA vinyl chloride monitoring unit, as a part of
a series of interlaboratory studies designed to compare
different methods of analysis. These tubes included
quality control unknowns prepared by the NBS, as well
as field duplicate samples. The vinyl chloride concen-
tration in the tubes ranged from about 8 to 3500 micro-
grains. The results, in terms of a comparison between
different methods, are given in Table IV. The mean
values for the quality control unknowns were calculated
using the NBS value as the reference value. Thus,
(Lab result - NBS)
(NBS)
The duplicate field samples were treated similarly; in
this case, Method A was the reference, since it alone
was used to analyze all tubes. Thus,
(Method A - Lab)
(Method A + Lab)
It appears, then, that neither the calibration gases
nor the NBS charcoal tubes deteriorate appreciably with
time.
Vinyl chloride was also analyzed on-site with
Tedlar grab bag samples. These samples were collected
in a survey of the air in the vicinity of vinyl chlor-
ide plants whose estimated current vinyl chloride
emission was about 3000 pounds per day. It quickly
became obvious that the area air contained airborne
contaminants which had not been so far encountered in
the laboratory. In particular, one compound, which is
omnipresent in this atmosphere, has a chromatographic
retention time relative to the vinyl chloride monomer
o£ 0.94. This compound could not be Identified without
the use of mass spectrometry, which was not available
in the field survey. At-a chromatographic column
temperature of about Htf'c, the chance of interference
from this unknown compound, was sufficiently high to
warrant reducing the operational temperature to 80°C,
thus permitting deeoavolution of the vinyl chloride
and "interference" chromatographic peaks. Since
ambient concentrations of vinyl chloride were in the
range of 0.0 to 0.3 ppm, this interference is signifi-
cant. When concentrations of vinyl chloride are 0.3
ppm or greater, the interference, when recognized, may
be eliminated by judicious Blope sensitivity adjust-
ments to the computer algorithm.
Conclusions
(1) A simple uncoated stationary phase provided
the most stable, reproducible and reliable analytical
gas chromatographic column.
(2) The desorption of vinyl chloride from acti-
vated carbon, with care of standard conditions,
especially temperature, will produce recoveries of
tetter than 90% for absorbed VCM concentrations in
excess of 10 nanograms. For definitive identification,
the combination of GC-MS should be used, preferably
with background subtraction capacity.
(3) More attention should be given to the defini-
tion and identification of hazards related to the
storage and fabrication of PVC resins. Until specifi-
cations for PVC stocks require the raw resin to
contain only negligible amounts of entrapped monomer,
storage and fabrication facilities.must be considered
potentially important sources of fugitive emissions.
(4) The grab sample provided a useful means for
plume chasing and sampling site selection, although
weather played a major factor in the quality and
quantity of data provided by this analysis.
2
2-6
-------
Table I
Table XXI
Retention Times of Some Potential Interferences
to Vinyl Chloride Analysis on Chromosorb 102
80/100 mesh, 6 foot x 1/8" packed column
145 °C isothermal
Vinyl Chloride Retention Time is 1.65
Comparison Head Space VCM and VCM Entrapped
for Contained Raw Resins
min.
(1)
Hexafluoroethane
0.44
(2)
Trifluoromethane
0.45
(3)
Tetrafluoromethane
0.46
(4)
Chlorotrifluoromethane ....
0.52
(5)
Ethylene
0.55
(6)
1,1-Difluoroethylene
0.55
(7)
Vinyl fluoride
0.62
(8)
Chloropentafluoroethane ...
0.74
(9)
Bromotrifluoromethane
0.74
(10)
1,1-Difluoroethane
0.86
(11)
Octafluorocyclobutane
0.87
(12)
Chlorodifluoromethane
0.88
(13)
Propylene
1.03
(14)
Freon 12
1.15
(15)
l-Chloro-l,l-difluoroethane
1.51
(16)
Freon 114
1.99
(17)
Isobutylene
2.06
(18)
1,3-Butadlene
2.09
(19)
Butene-1
2.14
(20)
Vinyl-methyl-ether
2.24
(21)
Trans-butene
2.27
(22)
Cis-butene
2.42
(23)
Vinyl bromide
3.06
Table II
VCM Recovery from Activated Carbon
ng VCM
Injection of
Standard Gas
Desorbed VCM
from CS2
Recovery
126
0.2388
0.2383
0.2380
0.2371
0.2357
0.2300
0.2351
0.2314
0.2279
0.2277
97%
Mean
area
0.2276
0.2S04
54
0.1013
0.1006
0.1018
0.1014
0.1007
0.1001
0.0986
0.0981
0.0993
0.0995
98%
Mean
area
0.1012
0.09912
14.1
0.0271
0.0269
0.0285
0.0278
0.0270
0.0255
0.0261
0.0258
0.0/63
0.0267
95%
Mean
area
0. 0275
0. 0261
2.7
0.0060
0.0058
0.0062
0.0058
0.0058
0.0059
0.0057
0.0052
0.0052
0.0048
0.0049
0.0053
0.0052
0.0048
0.0055
0.0051
88%
Mean
area
0.0057
0. 0051
Sample
Number
Run
Entrapped
VCM in yg/gm
Run
Head Space
VCM in ppm
1
416
2
428
3
429
4
379
1
1133.04
A
5
540
2
695.81
6
528
3
517.22
7
572
8
513
9
426
Mean
470
1
94.88
2
89.85
3
75.07
1
99.21
B
4
84.68
2
80.67
5
86.39
3
70.70
6
81.38
7
79.79
Mean
84.58
1
4
10.05
15.91
15.49
13.54
1
2
3
16.76
11.84
10.80
Mean
13.75
1 4.64
2 6.57
3 4.53
4 4.50
Mean 5.06
15.71
13.93
9.97
Table IV
Results of Different Methods
Applied to NBS Tubes
Environment Canada
(Method A) (Method B) (Method C)
Average
Bias
-8 %
-1 %
29 %
Average
Standard
Deviation
8 %
15 %
15 Z
Results of Field Duplicates
Relative to (A) Results
Environment Canada
(A) (B) (C)
Mean
Difference
-4 %
15 %
-21 %
Standard
Deviation
45 %
40 %
33 %
2-6
-------
50-
Col. A
45-
Col. B
40-
35-
30-
0. 25-
20-
1S-
10-
0.12
o.oa
l.1»
Area (Amplifier X 10|
0.04
0.20
Figure 1: Comparative responses for two separate
combinations ofinjection, column and detector.
iVinyl Chloride
1
e
i
I
0
4
1
Retention Time (min)
Figure 3: Representative gas chromatogram of
vinyl chloride in carbon disulfide.
0.25
V. C. heomf Inxn GS2
Std.
<
10
30 40
Conc'n PPM
•0
References
1. Kramer, C.G., and J.E. Mutchler. Am. Ind. Byg.
Aseoo. J., 33, 19 (1972).
2. Creech, J.L., and M.N. Johnson. J. Oooup. Med.,
16, 150 (1974) .
3. Selikoff, I., et al. Saw England J. Med., 292,
17 (1975).
4. "Report of a Working Group on Vinyl Chloride",
Internal technical Report, No. 74/005, Internation-
al Agency for Research on Cancer, Lyon, France.
June (1974).
5. Sassu, G.M., F. Zilio-Grandi, and A. Coute. J.
Ckpomatog., 34, 394 (1968).
6. Brighton, C.A., in "Encyclopedia of Polymer Sci-
ence and Technology". Ed. Kirk-Othermer, John Wi-
ley, New York, 14, 305 (1971).
Figure 2: Recovery of known amounts of vinyl
chloride desorbed from activated carbon
with carbon disulfide as compared with
known amounts Injected directly into gas
chromatograph.
4
2-6
-------
BASIC MEEDS IN AIR MONITORING IN A. EEVELOPING COPNTRY
J.M. DAVE
Deputy Director
National Ehviromantal Engineering Eeaearch Institute
Nagpur, India
1. INTRODUCTION
In India, air pollution is becoming an increasing-
ly Important aspect of environmental pollution in the
wake of rapid industrialisation. Because of Invisibi-
lity of low levala of air pollution, preventive action
on this has often been neglected and, as a result, it
has reached a stage of concern in many cities. lack
of adequate data, of technical personnel, equipment
and laboratory facilities has always left much to be
desired in adopting environmental pollution control
measures. National fiivironmental Engineering Research
Institute, Nagpur is engaged to develop appropriate
techniques for assessment of air pollution status in
urban and industrial areas to suit the local situat-
ions like availability of skilled man power, equipment
and other facilities, legal status of the authorities/
agencies, meteorological factors and the nature of
pollution. The pattern of the monitoring sys tem is
discussed in the paper together with the results that
have been achieved.
2. HROBUM OF AIR POLLUTION IN INDIA & IIS STATUS
India is a rapidly industrialising country add-
ing consequently to its air pollution sources. Pri-
ncipal growth is observed in the fields of power,
petro-chemicals, fertilizers, synthetic fibres like
rayon, metallurgy, sulphuric acid, nitric acid, phar-
maceutical industries and various other chemical and
ceramic industries. In addition to the industries,
the other important source of air pollution in the
Indian environment is the domestic consumption of low
grade fuels, resulting in intensely smoky a-taao sphere.
Contribution of the fine dust by the deserts and
other 6j>endryfields and the unpaved streets is none
leas to the ataoaphere which has resulted in a pro-
portionally higher pollution due to dust. Early
Indian citizens have been probably suffering from
the sooklness and dust in the air during summer and
winter months and reference to this affect is avail-
able in the ancient literature in descriptions of
the villages and the towns. These natural factors
combined with the human activities have caused heavy
pollution of the urban air with the suspended particu-
late matter.
This aspect has also been confirmed by surveys
conducted by NEERI in the large cities in India. It
has been observed that the main pollution is rather
concentrated in a few urban industrial pockets of
the country and the major pollutant is the particu-
late matter. In contrast to the other cities of the
world, Indian urban atmosphere presents a completely
different picture. The ratio of particulates to
sulphur dioxide ranges from 4 to 5 times as compared
to one to one and half times in the developed coun-
tries. As per the available data, concentration of
suspended particulate matter in large cities range
from 250 to 500 ag/m for 24 hour-annual mean with
the maximum levfi. above 1000 ug/nr under specific
conditions of dust storms and so on. The other com-
mon pollutants such as oxides of nitrogen, oxidants
and photochemical reactions record surprisingly low
levels under Indian conditions. The other special
problems pertinent to Indian conditions are emissions
of substances like fluoride, arsenic, cadmium etc
from the coal. Mast of the thermal power stations are
using coal from deposit belts of the Central and the
Eastern India which contain arsenic from 2 to 10 go/
ton and fluoride about 100 to 200 go/ton.
Auto-exhaust is yet another contributor to the
environmental pollution in the cities. ELghty per
cent of the Indian automobiles are more than five
years old and because of the age and improper main-
tenance mainly due to lack of the spare parts, they
emit large quantities of carbon monoxide, hydrocar-
bons, oxides of nitrogen and other pollutants. Stu-
dies for the two major cities show that automobiles
emissions alone account for 70 per cent of carbon
monoxide, 50 per cent of hydrocarbons, 30 to 40 per
cent of particulates and the equal percentage of
oxides of nitrogen in the atmosphere.
The air pollution problem in India is a complex
one due to natural dust, domestic smoke combined
with automobiles and industrial emissions.
3. PRINCIPAL FACTORS
Besides the characteristics of air pollution
there are various other important factors which have
a considerable impact on the monitoring techniques.
The objectives and the pattern of data collection
may entirely depend on these factors. The principal
ones ares
3.1 Man-power
There is a dearth of the trained people in the
field of air pollution chemistiy and analytical pro-
cesses, of course, well qualified chemists, electro-
nic specialists and science graduates in physics,
inorganic chemistxy etc can be quite capable when
trained specially to deal with the analytical techni-
ques and instrumentation etc as required in environ-
mental monitoring. But, this is an expensive propo-
sition. Plenty of personnel at a lower -cadre invol-
ving salaries are available. Though they have limi-
ted experience, they have reasonable skills and can
maintain and operate almost all the equipments ava-
ilable in this country.
3.2 Methods
The sampling and analytical procedures establi-
shed elsewhere had to be modified to make it suitable
under Indian conditions. The principal factors of
considerations are weather and meteorology.
3.2.1 Weather
India has a dry weather of 9 months w^tb a very
hot temperature ranging an average from 25 C at night
to 45 C during the day time with a humidity as low as
15 per cent. Also India experiences heavy dust storms
during the summer months when almost the entire
1
3-1
-------
northern India is covered with dust raised by the
wind blowing from the Rajasthan desert sometimes to
the height of £000 meters and above the ground level
with a high concentration of JOOGyag/m .
3.2.2 Meteorology
During winter months situation is very similar
to European countries except for the inversions which
start developing at about 6 pm and is lifted in the
morning at about 9-30 or 10 am with gn average maxi-
mum temperature difference of 6 to 8 C at an altitude
of 100 - 150 meters from the ground level. During
monsoon months heavy rainfall occurs during the short
period of 2 to 3 months giving a high humidity with
high temperatures and practically zero level of the
pollutants, as all of it is washed away with the rains.
In order to meet with such a situation the sampl-
ing procedures had to be modified. High volume sampl-
er sometimes could not be operated for more than 6
hours due to choke up by the heavy dust concentration.
Gaseous samplers also could not be operated with the
standard flow rate of one liter or half liter a minu-
te as extreme low humidity of 10 to 15 pec cent and
high temperatures upto 4^> C, which may dry up the bub-
blers leaving no yeagents for analysis. Similarly,
very high intensity of sunlight affect the stability
of reagents for S0_ and oxides of nitrogen techniques.
In order "to solve Such a situation, new sampling
methods had to be developed. These are short runs
for the high volume samplers and also 2 to 4 hour
period sampling with the change of bubblers for gases.
3.3 Equipment
Air pollution, being a new field, the equipment
involved has a limited market potential. Plenty of
common chemical analytical units are available in the
market, but specialised ones for air pollution purpose
are not available. Manufacturers from outside India
have also not started any units in India as there is
virtually a very limited demand for it to justify.
Due to foreign exchange difficulties, it is also not
possible to import any of the monitoring eqy&pment
in large numbers from other countries. When imported
it is very expensive and it is also difficult to
obtain spare parts and therefore many a times they
are not operative for long tine. Under these circum-
stances, indigenous instruments had to be developed.
National Environmental Engineering Research Institute
has constructed simple and cheap manually operated
gas samplers, high volume samplers and other equip-
ments for determination of meteorological parameters
which can be manufactured, operated and maintained
by the local personnel. Also extensive use of simple
sulphation candles in a coordinated manner has been
made. They are observed to be quite suitable for
the field operation.
4. NATIONAL SET-tJP
The political set-up in India is a Federal Stru-
cture with 22 states having autonomous internal autho-
rities. Many legislative measures with respect to
environmental pollution control are within their pre-
view while the location and licencing of the industry
is under the overall supervision of the Central Gover-
nment. This has resulted in a split of the responsi-
bility in the Centre and the States creating difficu-
lty in solving environmental pollution problems parti-
cularly relating to air pollution. Besides this, the
authority is again divided into the district local
authorities called ZLlla Parishad, Municipal Councils
and Committees and Corporations. Unfortunately,
clear demarcation of authority with respect to envi-
ronmental monitoring is not provided in the laws.
With increasing industrialisation, the enterpreneurs
sometimes select sites on all other considerations
except for air pollution.
Under the circumstances, the authority to main-
tain air quality and to undertake monitoring for that
purpose is not clearly defined. There are other con-
siderations as well of national impor twice which have
an impact on the monitoring activity. These are:
4.1 Priorities
Industrial development for consumer goods and
social needs have a priority over the others. Ferti-
lizers, textiles, increase in food production through
irrigation otc are given high priority whereas some-
times the associated environmental problems are
neglected.
4.2 Finance
Most vital issue in determining air monitoring
needs is the availability of funds. With the resour-
ces available, national priorities are many times
dictated in the developing countries for the demand
of basic needs such as food, clothing, shelter etc
which have a bearing on higher production of fertili-
zers, building materials etc. The allotment for
research is about 0,03 "per cent of national budget.
This has resulted ir. lesser funds available many
times for the environmental measures. Air monitor-
ing activity has also to compete with the other
environmental demands for the funds and its share
is too small at O.OOO175C, Obviously, therefore,
there has to be a system which can be easily operated
within this restraint.
4.3 Legal Provisions
Legal support is the most essential requirement
for maintaining a systematic monitoring activity.
India has at present limited laws enforced for air
pollution purposes. Four cities in India have Smoke
Nuisance Act for a considerable time such as Calcutta
in 1906, Kanpur in 1912, Bombay in 1920 and Ahmedabad
in 1960, Besides this, few Acts are provided in the
Corporation Acts to curb the nuisance as also some
provisions in the factory act for the prevention of
environmental pollution. None of than provides for
the establishment of a regular agency for the monitor-
ing activity. This probably is also true of megy
developing countries where they may not have a local
authority to set up the monitoring agency,
tJnder the circumstances all the monitoring work
done in India is either voluntary by the research
organisations like NEEB1 or stray measurements during
some post-graduate studies in the Universities.
Another type of monitoring done by HEE8I is the spon-
sored programme of air quality survey financed by
municipal authorities through NEERI for a period of
3 or 4 years.'
All these programmes do not have any legal sup-
port and are make-shift arrangements on a voluntary
basis. The Programme undertaken in India so far ir.
the air pollution monitoring activity is given in
2
3-1
-------
this paper,
4.4 Agencies
Presently in the national level, research organi-
sations and Universities are the only active agencies
in the field of air pollution monitoring. NEHRI has
a national air monitoring network and in Universities
sod© isolated work is done at one or two stations by
post-graduate students in few cities like Kanpur,
Bangalore, Surat, Ahmedabad and Baroda. The other
agencies are the Labour Institute, the Meteorological
Department, Bhabha Atomic Research Centre, where the
interests are limited in the specific field of their
activities and the sampling is generally undertaken
for limited purpose for a short period for the assess-
ment of the particular situation.
NEERI, is the national agency actively engaged
in the environmental pollution and has voluntarily
undertaken monitoring of national air quality trend
in the select nine centres of the country. NEERI is
also collaborating with the WHO as a Reference Centre
in Air Quality Monitoring. Three major types of acti-
vities such as residential., industrial and commercial
ore considered in air quality monitoring programme.
5. MOKITCKIMG NETWORK
The objectives for monitoring should be such that
the activity is directly effective and useful in air
pollution control and air quality improvement program-
mes. Also it should be within the reach of the capa-
bility and resources available. The basic objectives
of air monitoring programme c an be broadly termed as
(1) determination of facts, (2) diagnosis, and (3) Pre-
diction. The objectives and the scope of monitoring
within the above-aentioned broad outline, may cover,
information on local, regional or global air quality,
evaluation of the impact of the industries, study of
the effects of natural factors such as meteorology,
topograph and the fate of the pollutants after dis-
charge by dispersion, chemical reactions etc under the
£yjiical cliaatological conditions and forecast of the
air quality. The other important purpose of study can
also be evaluation of the effect of pollution on human
health, plants, animals and property -where intimate
correlation is involved. The air monitoring techni-
ques adopted may very accordingly.
In a developing country like India with various
constraints as described above, it is very expensive
to adopt most sophisticated equipment on which short
interval (10 minutes, 5 minutes) concentrations are
recorded continuously for 24 hours automatically.
The type of equipment, sampling period and frequency
etc should be within the available resources end
also to obtain data which by itself will be most
useful for air quality improvement. WIS) Ebqpert Com-
mittee in their Report No. 506 (1972) has suggested
that the health effect standards for air quality-
should be baaed on 24 hour annual mean basis rather
than en hour-to-hour basis. To meet this recommenda-
tion it is not necessary to have an automatic sampl-
ing equipment, if data can be obtained with the rea-
sonable confidence limit. With due consideration to
this, in India, a sampling pattern has been evolved
covering two aspects of the air pollution.
5»1 National Network
The purpose of the network is to assess the trend
of air pollution over years in the principal nine
cities of India. Parameters selected are particu-
lates, sulphur dioxide, oxides of nitrogen, total
oxidants and hydrogen sulphide.
The equipment consists of locally developed
vacuum pump from which flow is divided equally
through a manifold into four different air sampl-
ing bubblers. The gases collected are sulphur
dioxide, oxides of nitrogen, total oxidants and
hydrogen sulphide. The flow rate is adjusted
manually to 500 ml/min. The bubblers are changed
after every four hour sampling. This is attended
to for 24 hours to have a 24 hour evaluation.
The particulates are collected by an indige-
neusly developed high volume sampler which draws
air at a rate of 1.8 to 2 cu r of air per hour
through a fibre glass filter paper of 8" x 10"
size. The particulates are analysed gravimetrica*-
lly for the total mass and also for various other
components like metals, organics, inorganics, ben-
zene solubles etc.
The meteorological equipment consists of record-
ing type wind instruments to measure direction an$
speed. These are designed by NEERI which operate on
a clock mechanism. The site is generally selected
to give a representative air sample for the whole
city as far as possible. Generally, it is selected
between the height of 4 to 6 meters from the ground
taking into consideration such other aspects as
power supply, safety of the equipment etc. The work
is carried out by the zonal staff of NEERI. The
samples are collected once in 15 days round the year./
5.2 City Air Monitoring Network
The City of Calcutta and Bombay have wide moni-
toring networks established by NEERI. They consist
of 6 to 10 major sampling stations called type 'A'
stations where parameters like suspended particulates,
sulphur dioxide, oxides of nitrogen, hydrogen sulphidq
sulphation rates and dustfall rates are collected
simultaneously. Additional stations known as type *B*
are also located to measure sulphatlon rate and dust-
fall. The stations are selected to give a represen-
tative picture of the air quality from the area as
far as possible. These cities are divided into
varioios zones for location of the sampling stations.
In each zone, one 'A* type and two 1B' -type stations
are set up.
Meteorological data is also simultaneously col-
lected at these stations. Additional information is
also obtained from the respective local observatory
of the Meteorological Departments.
Gas sampling kit is used for collectingggaseous
samples with 24 hour attendance. The bubbling rates
used for this purpose are 250 and 500 ml per minute
with bubbler capacity of 35 ml with 20 ml of absorb-
ing reagent. The high volume sampler Is also a simi-
lar unit identical in all respects. The sulphatlon
candle consists of lOOsq. cm on 25 cm dla x 200 mm
PVC cylinders mounted inside a standard louvered
box.
5«3 Frequency
The sampling is carried out once a week on dif-
ferent days of a week in Bombay and once in three days
3
3-1
-------
for Calcutta with, the staggering of the sampling days.
The sulphation rates are carried out continuously for
30 days. The correlation of sulphation rate with con-
tinuous sulphur dioxide measurements is also made for
different seasons*
Mobile vans are used for field work for the spot
sampling.
6. ETFECTTVEHESS OF THE SYSTEM & DATA ANAIS5IS
the -typical data obtained for this network are
shown in Table Nos. 1 and 2. The initial studies
Table No.li Pollution features of Greater Bombay
(Annual 24 hour mean and maximal in Mg/m )
Sta-
tions
S02
no2
V
SFM
Mean
Max.
Mean
Max.
Mean Max.
Mean
Max.
Bandra
28
118
38
245
5
21
223
486
Bhandup
56
155
28
63
5
26
317
594
Chembur
65
231
66
182
6
35
341
698
Dadar
61
205
49
297
6
57
329
860
Dongri
85
231
45
133
4
29
305
638
Goregaon
58
247
35
140
5
19
359
793
Ghatkopar
Ifi
145
35
140
7
63
319
702
Lalbag
128
372
52
105
8
51
454
2900
Sion
* 46
127
38
154
7
59
309
636
Average
61
43
6
326
Table No.2t SO-, data for Calcutta
^expressed in pg/nr')
1TO W" w
Mission Road
16
355
33
Chittaranjan
335
32
29
44
Cossipore
395
37.2
41
54
tiyde Road
314
62
35
43
HIS Hospital
357
44
26
Basgur
364
44
37
27
Maniktola
381
23
25
27
Paster Lab
-
35
39
Howrah
507
64
47
44
Botanical Garden
-
15
14
Dun Dum
-
24
16
Nareoodrapur
-
6
11
-
Tiljala
267
-
33
34
were carried out to compare the values of the continu-
ous sampling with these frequencies and it has been
observed that the 15 days sampling period gave the
confidence limit (also reported by EPA in their report}
of 80 per cent for the national network for obtaining
the annual 24 hour mean concentration.
In the city survey, the weekly samples resulted
in confidence limits of over 91 per cent for annual 24
hour mean concentration. But hourly maximum concen-
tration confidence limit was less than /fi per cent.
7. COST
For the trend studies each station costs
approximately Ji.20,000 (US $ Z/fiO a year) Of this
US $ 800 accounted for the capital equipment cost for
the station and about US $ 1600 as the operational
cost). For -the city monitoring, cost of each station
was Rs.50,000 (US $ 1200 for the capital and US $ 3600
for operational costs for each station). This cost
included expenses of man power, equipment, chemical
glasswares etc,
8. DISCUSSIOM AND CONCLUSION
The data obtained has bean evaluated by statis-
tical analysis. The typical data obtained from the
national network for 24 hours was within the 80 per
cent confidence limits and was comparable with the
international standards of the WHO Expert Committee
recommendations.
The data for the city monitoring had a confi-
dence limit of more than 90 per cent for the 24 hour
annual average mean.
Though there might be a lacuna in accurate pre-
cision with respect to the exact half a hour or one
hour levels but for practical purposes this sytem
served the purpose of comparing air quality from one
cily to another and also to a limited extent with
the international standards.
With the limited finance and legal provisions
for establishing regular monitoring network, this
probably is the best possible that can be achieved
in a monitoring system which has been used as a
simple and inexpensive tool for primary assessment
of air quality data.
8.1 Advantages
(1) It is reliably a good system for very low
eoqpense.
(2) Moat of the procedures used are primary methods
like (TCM) for S0„, and NO- and therefore error
will be very limited.
(3) The equipment could be used except for power
failure. Every thing is simple and made to
operate indigenously.
(4) Skilled maintenance or operation is not required.
(5) The equipment suits the climatological end other
conditions in the country and gives troubla-free
operation.
(6) Visual observations of the air pollution situa-
tion can be made which are very useful in inter-
preting the air quality data later.
8.2 Shor-twywii-lppa
(1) Human factor plays a major role in the system
which is very difficult to assess precisely.
Different analytical procedures may involve
possibility of human error particularly while
comparing the data from different cities.
(2) Man power requirement is considerably greater
for a large scale operation.
(3) Limited sensitivity of the method since it can
only obtain concentration for one hour
only.
4
3-1
-------
(4.) The data for half an hour peaks if required will
require higher capital expenditure on the equip-
ment.
(5) This system may not be suitable for precise air
quality measurements for health affect studies.
8.3 Possible Improvements
Based on the above assessment of the system the
possible areas of improvement that can be considered
for the system ares
(1) The system has been designed on the basis of the
local situations and therefore data obtained may
not be comparable to other standards. It may be
advisable to undertake comparative studies with
the other standard methods adopted in developed
countries.
(2) Certain manual operations in analysis can be
avoided by utilising some of the simple labora-
tory automatic analytical procedures which may
reduce the errors due to the human factor.
(3) The development or modification of the precise
automatic measuring instruments with reduced
accuracy may suit the local conditions with the
decreased sampling error.
9. RECOMMEKDAII QMS
Based on the experience obtained in India by the
author and also in view of the activity as a Reference
Centre for WHO, the following recommendations may be
useful.
9.1 General
(1) There is an imperative need to standardise the
sampling procedures and state the requirements
in terms of achieving the accuracy for compara-
bility of data from various countries.
(2) There should be some sort of a standardised
system for laboratory procedures and sampling
equipment to meet the above mentioned precision
requirements suitable for both developed and
developing countries so that data obtained is
comparable.
(3) Technical guidance should be recommended by the
authorities as a basic need of the network moni-
toring programme.
(4-) Training facilities should be established where
found necessary to standardise techniques and
procedures.
9.2 Role of International Agencies
International agencies like WHO, VJMO and UNDP *
can play very important role In collaboration with
other national agencies like EPA etc who have coopera-
ted and pooled all their efforts to bring about this
seminar. It would be in the fitness of things if the
deliberations of the seminar come out to be conclusi-
vely beneficial to the problem of monitoring as a who-
le to all the nations. For this purpose, the inter-
national agencies might consider the following course
of actions
*UNEP, UNESCO etc
(1) Establish priority areas of monitoring for
international cooperation.
(2) Rationalise the criteria for the purpose of
monitoring so that all monitoring pattern are
designed to meet the stipulated requirements
on a minimum basis. These minimum needs should
however be matched with the maximum benefit,
especially in view of the limited resources
available in the developing countries such as
equipment, man power, finance etc. It should
not be just collection of data but a basis of
interaction beneficial to human beings.
(3) Standardise the equipment and laboratory proce-
dures to make the results comparable for such
an international monitoring network.
(4.) Only a few nations have basic facilities in the
field of air pollution. They should extend all
the training and other such assistance to the
needy countries for developing a minimum and
basic monitoring system.
(5) Establish channels for exchange of information
and techniques.
(6) Evaluation of such data by single agency so
that assessment is made uniformly.
10. ACKWOWiaDGEMEMTS
The author would like to express his grateful
thanks to the Director, NEERI for his kind permission
to prepare this paper. The data used ore taken from
the NEERI work. The author also would like to take
this opportunity to express his sincere thanks to
Mr F.K. Yennawar, Scientist for his able assistance
in collecting information, Mr R.K. Saraf, Scientist
for editing the paper and Mr B. Damodaran, for the
secretarial assistance.
5
3-1
-------
A PRELIMINARY DESIGN FOR A NATIONAL
ENVIRONMENTAL CENSUS
RICHARD H. ROSEN, CHIEF SCIENTIST
•RONALD P. BECK, MANAGER ENVIRONMENTAL MONITORING
Energy Resources Co. Inc.
185 Alewife Brook Parkway
Cambridge, MA 02138
Abstract
This paper first examines the present
level and future requirements for pollution
control expenditures in the United States.
The explicit and implicit mandates of the
Water Pollution Control Act (PL 92-500) and
the Air Pollution Control Act for monitoring
and evaluation of environmental quality are
then considered. In order to summarize the
formal and informal requirements for monitor-
ing, the existing system of environmental
monitoring is critiqued. An analysis of the
difference between the existing system and
that which has been mandated by a rational
environmental monitoring policy is then
presented. Finally, specifications for a
national environmental census designed to
deal with the requirements of the law and
the inadequacies of the present system are
provided.
Investment for Pollution Control
The requirements for investment in water
pollution control are substantial by any
standard. Capital investment alone is ex-
pected to exceed $80 billion between now and
1982. Total pollution control costs are
expected to exceed $40 billion annually by
1982"'" — over three times the present level
of expenditure. The total level of environ-
mental expenditures is comparable to other
large public programs, such as national
defense procurement.
A variety of studies by the Council on
Environmental Quality and the EPA have shown
that the country at large can afford these
expenditures, but the country cannot afford
to see these expenditures misapplied so that
a massive opportunity cost is suffered by the
American population. Unfortunately, there is
growing evidence that the existing investments
for pollution control have not resulted in
improvements in environmental quality propor-
tionate to their cost. Indeed, some invest-
ments for environmental quality control
appear to have actually worsened the quality
of the environment in areas where the intent
was to improve it. A classic example of this
may be found in the results of the municipal
treatment program — whose very operation can
frequently do more damage to the environment
than the maintenance of even a primitive
existing water pollution control system in a
2
particular area. If, as some recent studies
suggest, most of the apparent benefits in
water quality improvement are attributable to
industrial pollution control (which has had
dramatically smaller expenditure than has
existed in municipal treatment), one may well
ask the question, "Why should we be following
the present path of management, technology
and investment in municipal waste treatment?"
Similar questions can be raised about the
nation's air quality program, where recent
studies have shown little or no change in
morbidity or mortality resulting from sub-
stantial investment in air quality improve-
ment. Several investigators have suggested
the reason for this can be found in the
politics surrounding the establishment of
criteria pollutants."* Fine particulates and
toxic agents, including carcinogens such as
benzapyrine, have proliferated in urban areas
with no presently existing formal basis for
their control. Equally troublesome is the
lack of control of a variety of pollutants
attributable to the operation of motor vehi-
cles. In both cases, one observes zealous
attempts on the part of administrators in
both Federal and State governmental agencies
to reduce quantities of certain pollutants
to very low levels without regard to the
impacts of the control procedures on the
generation of other pollutants which may well
be more virulent. This kind of environmental
quality management is clearly irrational and
must riot be allowed to continue.
At the present time it is practically
impossible to rigorously evaluate the nation's
environmental quality regulations and invest-
ment decisions. This is made difficult
because an adequate system of monitoring
which allows for complete trend analysis along
relevant dimensions does not exist — in a
comprehensive enough sense — to provide
policy makers and the Congress with the infor-
mation that they need to sensibly revise the
Clean Air and the Federal Water Pollution
Control Acts. Proper observation and evalua-
tion of environmental control policy provides
the first major stimulus to a program for
national environmental data gathering.
Legislative Requirements
The second major influence comes from
the legislation itself, including the many
requirements for the generation of environ-
mental impact studies and environmental
assessments. The need for historical data to
properly perform environmental impact analyses
and environmental assessments is obvious. It
is hard to examine the differential effect of
a particular facility if there is little or
no understanding about the historical behavior
of the affected area.
The legislation has explicit and implicit
requirements for monitoring beyond the need
for data for environmental impact analyses
1
3-2
-------
and environmental assessments. Seven basic
monitoring needs are common to both pieces
of legislation. These are:
1. The review of the effectiveness
of existing programs
2. The determination of substances
which should be controlled by
either regulation or statute
3. The establishment of standards
for control
4. The establishment of land use
planning implications
5. The design of a useful R&D
program
6. The evaluation of the utility of
existing pollution control tech-
nologies
7. The enforcement of regulations
In addition, the Water Pollution Control
Amendments provide for several other specific
monitoring requirements, including State
305 b reports, in-place pollution studies,
special studies of acid mine drainage and
the assessment of the need to control toxic
and hazardous materials.
Levels of Disaggregation
One of the major weaknesses of the pre-
sent pollution control monitoring activities
is that they rarely relate data in a reliable
way to the needs of relevant decision makers.
This paper examines specifically, monitoring
requirements for three levels of data dis-
aggregation: The nation as a whole, indivi-
dual states and basinwide or environmental
planning district. The latter conforms quite
closely to environmental planning districts
in several states and to those isolated by
the 208 program. For national evaluation,
approximately 1600 monitoring sites are re-
quired. For statewide estimates, the number
increases to 8250. When one examines
individual basins, the number increases to
16,500.4
The Existing System
It is now necessary to examine the exist-
ing system. One can look at how useful the
existing data is by taking some simple policy
problems and attempting to provide informa-
tion on them. We have analyzed the program
to monitor the limitation of industrial
discharges of mercury into waters and of SOx
into the air. To evaluate the effectiveness
of these regulatory programs, several years
of data on ambient concentrations of these
chemicals in the water and in the air are
required. Data is needed on a nationally
representative set of stations and the dis-
tribution of data points must be representa-
tive of the population distribution; since
the effectiveness of control can be measured
by the reduction of the population-at-risk
associated with each of these chemicals.
We have collected all of the available
nationally tabulated data for ambient levels
of both sulfates and mercury. The following
summary tables provide an assessment of the
cost effectiveness of these monitoring pro-
grams. By dividing the U.S. into air quality
regions or water quality regions, the
proportion of monitoring effort to population
can be examined. One would expect that the
ratio of these proportions would be close to
unity. In fact, the analysis shows that many
of the observations are zero, indicating no
monitoring effort; while in several cases the
ratio exceeds four, and in one case it exceeds
one hundred. Moreover, when in addition it is
required that data satisfy minimum statisti-
cal constraints for historical analysis,
five of the ten EPA regions are completely
unrepresented with data.
FREQUENCY DISTRIBUTION OF MR QUALITY
REGIONS BY SAMPLING INTENSITY FOR PARTICULATES AND SO
SAMPLING INTENSITY PANGE
PARTICULATES
S°*
0
104
169
>0
<
1
87
76
>1
<
2
68
31
>2
<
ss
3
29
24
>3
<
4
8
8
>4
26
14
There are 322 air quality regions. Sampling
intensity for region i is determined as the ratio
of the percent of the nation's monitoring stations
in region i/percentage of nation's population in
region i.
FREQUENCY DISTRIBUTION OF STATES BY SAMPLING
INTENSITY FOR WATER QUALITY
SAMPLING INTENSITY RANGE
WATER QUALITY STATIONS
>0 * 0.5
10
>0.5 i 1.0
12
<
>1.0 * 1.5
9
>1.5
19
There are 50 states. Sampling intensity for
state i is determined as the ratio of the percent
of the nation's monitoring stations in state i/
percentage of nation's population in state i.
For water monitoring, ten of the fifty
states have less than half of the number of
monitoring stations required to represent the
population, and nineteen of fifty states are
overrepresented by over fifty percent. For
air monitoring of sulfates, 169 of 322 air
quality regions are unrepresented. Forty-six
of 322 air quality regions are more than 100%
overrepresented with respect to the popula-
tion in that region.
It is clear from the aforementioned
analyses that no national perspective has
been provided to the design of water and air
monitoring networks. Part of this problem
2
3-2
-------
relates to the extent of bureaucratic atomi-
zation which has characterized environmental
monitoring. No one organization is in charge
of the effort. Federal and State agencies
collect data to satisfy their parochial
interests. „This has led to a condition that
must clearly be viewed as bizarre from the
perspective of efficiency. Over 5,453 water
monitoring and 1,343 air monitoring stations
exist. In the case of water, this is more
than enough to meet the need for national
estimates and over sixty percent of the
requirement for state numbers, provided the
resources were rationally distributed.
Unfortunately, they are not. The same ob-
servation can be made for air monitoring.
A firm force must be exerted to reprogram
these resources so that they may be employed
efficiently.
This force should be a national environ-
mental census. In the past, environmental
censuses have been conducted only to be
abandoned as too costly by the cognizant
agency because they greatly overestimated
data requirements. It is quite possible to
establish guidelines to satisfy one or more
monitoring requirements alluded to above.
The issue is how much will it cost.
WATER AND AIR MONITORING COSTS
NO. OF
STATIONS
COST/STATION
TOTAL
COST <$)
WATER
national
1,600
10,000
16 x
106
State
8,250
5,100
42 x
106
District
16,500
2,800
46 x
10 6
AIR
National
1,600
32,000
51 x
106
State
8,250
24,000
198 x
106
District
16,500
14,000
231 x
10s
Source: "Evaluation of Monitoring Costs''^
For water monitoring, the census would
cost between 68 and 104 million dollars per
yearj while for air, the cost would be be-
tween 304 and 480 million. Adding fifteen
percent to these costs for administration and
analysis, the total cost of an environmental
census would be between 428 and 584 million
dollars per year.
At the present time, this appears to be
less than the amount spent by Federal, State
and local governments and industry for
environmental monitoring. When this amount
is compared to the annual cost of pollution
control, it is found to be within the typical
range of two to four percent of the amount
expended for control by most well-run organi-
zations.6 In 1980, two percent of annual
costs will be 920 million dollars.
References
1. Council on Environmental Quality.
Environmental Quality 1974. The
Fifth Annual Report of the Council
on Environmental Quality. U.S.
Government, Washington, D.C.
December 1974.
2. Energy Resources Co. Inc. Cost-
Effectiveness of the Municipal
Wastewater Treatment Program.
U.S. Environmental Protection Agency,
Washington, D.C.
3. Stephen Marcus has reviewed some of these
issues for a variety of environmental
publications. A recent summary of his
findings appears in an article in The
Christian Science Monitor. Others,
including J. Speyer of EPA, have also
made this observation.
4. These calculations were based on the pro-
cedures found in Sampling Techniques,
W.G. Cochran, Wiley, New York, 1963.
5. These numbers were obtained as a part
of a research effort presently underway
for EPA. The Final Report from this
project will be available in January
1976 from either Energy Resources or
the EPA.
6. R. Anthony, Hazard Business School.
3
3-2
-------
DESIGN OF NATIONWIDE WATER-QUALITY MONITORING NETWORKS
by
R. J. Pickering and John F. Ficke
Hydrologists, U.S. Geological Survey, National Center, Reston, Va.
Abstract
The many facets of "water quality," which include
physical, chemical, and biological characteristics of
water, make it impossible to design an all-purpose
water-quality monitoring network. The specific
objectives of the monitoring program must be clearly
defined before a proper mix of sampling sites,
characteristics to be measured, and frequency of
measurement can be selected. Design of ground-water-
quality networks, for example, must be based on a
knowledge of direction and rate of water movement,
while proper sampling in streams requires a knowledge
of velocity distribution in the cross section and
expected variation in streamflow and water quality
with time. Funding and manpower are commonly the
primary constraints for network design, rather than
acceptable error or ideal coverage of water-quality
characteristics.
The U.S. Geological Survey, as the agency responsible
for describing and evaluating the water resources of
the United States, has designed several networks for
monitoring water quality on a national and regional
scale. The largest of these is the National Stream
Quality Accounting Network (NASQAN), which is designed
to monitor quantity and quality of streamflow as it
moves from one hydrographic "accounting unit" to
another or to the oceans. Reports from NASQAN will
describe geographic and year-to-year variations in
streamflow and water quality, their probable causes,
and detectable long-term trends. Organic and radio-
active substances and biological characteristics are
measured at selected network stations that constitute
subnetworks of NASQAN.
Introduction
The subject of this paper, national networks for
monitoring quality of water, is an extremely broad
one. Its full scope could be the subject of many
papers. I will address the subject in general terms
at first, and then turn to a specific example of such
a network that is being operated by the U.S. Geologi-
cal Survey.
First, some definitions would seem to be in order. By
"monitoring" I refer to successive measurements over a
period of time for the purpose of detecting change, or
lack of it. "Network," in the broadest sense, refers
to an organized system for collecting a specific kind
of information. In the narrower sense, as used in
this paper, "network" refers to a series of points at
which one or more measurements are made, and which
have been selected to satisfy a specific objective.
For example, an objective might be to detect changes
in the quality of the Nation's large rivers, or to
measure the beneficial effects of our Nation's
pollution-abatement efforts, or to provide an early
warning of intrusion of salt water into a fresh-water
aquifer.
Water quality can be defined as "the wide variety of
physical, chemical, and biological characteristics of
water that make it fit or unfit for a particular use."
The several hundred inorganic chemical constituents
and tens of thousands of organic compounds that can
occur in water serve to demonstrate that not all
aspects of water quality can be addressed by any one
monitoring network. When reference is made to a
"water-quality monitoring network," we must immediate-
ly ask "what kind of network?--What is the objective?—
What water quality characteristics are of interest?"
Once the objective of the water-quality monitoring
network is defined, decisions must be made on (1]
water-quality characteristics to be measured, (2) sites
at which measurements are to be made (objectives may
be 2-dimensional or 3-dimensional with depth the
third dimension], (3) frequency of measurement for
each characteristic and at each point, and (4) the
statistical parameter that will be used to report each
characteristic.
Almost all networks rely on point measurements, and
interpolation and extrapolation therefrom -- both in
space and in time. Exceptions include certain measure-
ments that can be made by aircraft or satellite, such
as temperature, or for which sensors have been develop-
ed that can be used to record variations with time at
a point at streamside, or variations in space by being
towed in a boat -- for example, specific conductance,
an indirect measure of dissolved solids; dissolved*
oxygen; pH; turbidity; radioactivity; and a few other
water-quality characteristics.
This paper will review some of the principles of design
and operation of water-quality monitoring networks,
and in doing so will concentrate on the techniques and
experience of the U.S. Geological Survey.
Design of Networks
In designing any kind of water-data network, one must
initially decide CI) the function of the network, and
(2) the geographic scope and network must cover. It
is clear that a pollution-surveillance network will not
call for measurement of the same water-quality charac-
teristics as a network designed to define the "natural"
quality of water. It is clear also that selection of
monitoring sites for a network that can be used to
describe streamflow quality on a nationwide or broad
regional basis will be on a different scale than for
a network aimed at quantifying non-point pollution
sources in a small river basin.
The U.S. Department of the Interior's Office of Water
Data Coordination has developed a function-and-level
concept of network design which it has been using in
its efforts to coordinate water-data collection or
acquisition in the United States (figure 1), and to
develop a National System for Water Data\ The
horizontal axis on the diagram corresponds to
functional categories, that is, general objectives
for gathering data. The vertical axis refers to the
geographic scope of the data-gathering effort, and
thereby indicates the level of detail of the
4
information needed.
Within this generalized conceptual model, there is
much room for differences in approach to network
design. For example, a "systems concept," or
hierarchical approach, could be used for selecting
data-collection sites and water-quality characteris-
tics to be measured. A carefully-designed random
sampling procedure might be employed in other circum-
stances, such as to describe water-quality in an
aquifer for which variability is poorly known. Or a
hydrographic basis, calling for measurement of water
quality at successive downstream points, could be used
to illustrate incremental changes in water quality as
surface water moves toward the sea.
1
3-3
-------
f
LOCAL OPERATION AND MANAGEMENT OBJECTIVES
(Not uniform* not subject to design)
SUBREGIONAL PLANNING OBJECTIVES
(Non-uniform; varying in response to
local conditions, but subject to design)
NATIONAL AND REGIONAL PLANNING OBJECTIVES
(Uniform nationwide)
STREAMFLOW
STREAM QUALITY
WATER USE
SURFACE RESERVOIRS
GROUND-WATER
RESERVOIRS
ACCOUNTING
WATER QUALITY
GROUND-WATER LEVELS
SURVEILLANCE
DATA COLLECTION AND
INTERPRETATIVE STUDIES
OF QUANTITY AND QUALITY
OF GROUND AND/OR SURFACE
WATERS, OFTEN INVOLVING
SEVERAL PHASES OF THE
HYDROLOGIC CYCLE.
AREAl SYNTHESIS
FUNCTIONAL CATEGORIES
Figure I.— Relation of levels of information to
functional categories in the National System for.
Water Data.
Design of surface-water quality monitoring networks
differs from that vof ground-water networks in several
respects. Because surface flow is confined to a stream
channel, there is usually no question about the
direction'of movement of the water, and techniques are
available by which to measure or estimate the length
of time it takes the water to move from one point to
another under different stTeamflow conditions.
Nevertheless, a representative sample of water in the
stream cross-section at a particular time may require
a series of depth-integrated samples along the cross-
section because of non-homogeneous distribution of
the water-quality characteristics of interest. The
sampling or measurement pattern should be determined
through a site evaluation study made at the time the
measurement site is first selected.
Ground water, on the other hand, is much less
accessible than is surface water, sampling points
necessarily are fewer in number, and the hydrologic
characteristics of the aquifer are seldom known well
enough to predict in detail either the exact direction
or the rate of movement of the water. A general
knowledge of the hydrology of the aquifer must be
acquired before the significance or cause of changes
in water quality with time can be assessed.
Monitoring of pollution sources and other factors
affecting water quality, such as land use and natural
basin or aquifer characteristics, will provide a
basis for postulating causes of water-quality changes
and identify tools for predicting and modifying such
changes if necessary. Funding and manpower are
commonly the primary constraints to network design,
rather than acceptable error or ideal coverage of
water-quality characteristics.
The U.S. Geological Survey has utilized all the above
concepts in designing its water-data collection
networks. Its Benchmark Network, which is aimed at
defining the range of "natural" streamflow and water
quality and the factors controlling them, consists of
data-collection sites in small stream basins that are
1 2
unaffected by man. ' The basins in the network were
selected to cover the Nation geographically and to
represent a good mix of basin characteristics. Our
National Stream Quality Accounting Network (NASQAN)
consists of data-collection sites at the mouth of each
of 325 hydrographic "accounting units," and is
designed to provide a measure of the quantity and
quality of water moving from one accounting unit to
3
the next. NASQAN data thus provide a measure of the
quality of the Nation's major rivers.
Let us look at the design of NASQAN in more detail.
As stated above, the United States has been subdivided
into 325 hydrologic accounting units. For units with
regular, well-integrated drainage, one station was
placed as near to the downstream end of the unit as
was practical. Our goal was to measure and have access
for sampling of 90 percent of the streamflow moving
from one accounting unit to the next. This is what we
mean by accounting. However, because of parallel
drainage patterns along coastal areas, and complex
hydrologic situations elsewhere caused by dams and
reservoirs, it was necessary to have more than one
station in a number of accounting units. Indeed, in
coastal regions it has been necessary for us to use a
"representative station" concept. NASQAN stations
have been located to provide a sampling of from 30 to
50 percent of the outflow of each coastal accounting
unit. This water-quality information, together with
more extensive streamflow data, will be used to
estimate the quality of the remainder of the outflow
of a particular accounting unit.
To allow for the above considerations, the design of
NASQAN calls for 525 stations when the network is
completed. To date, 345 stations have been establish-
ed at the locations shown in figure 2. Note that a
large share of the stations are along coastlines,
along international boundaries, or near the mouths of
tributaries to major rivers where they monitor the
drainage from tributary basins.
A modified accounting approach has been used in the
design of NASQAN subnetworks for monitoring radio-
chemicals and pesticides. Because money was not
available to operate other full-scale networks the
size of NASQAN, 153 NASQAN stations were randomly
selected for monitoring of pesticides, and 51 stations
were selected for monitoring of radiochemicals. The
pesticide subnetwork is operated in cooperation with
the U.S. Environmental Protection Agency (EPA) and
will be described in more detail by another speaker
in this symposium. The Geological Survey also operates
the bulk of the stations in EPA's National Water
Quality Surveillance System, which provides a sampling
of selected pollution problem areas throughout the
Nation.
Operation of Networks
Once station locations have been selected, the next
issue that must be considered is what to measure and
how often. There are many possible courses of action,
and again the objective of the network must be the
controlling factor. There usually is a strong
temptation to select a suite of measurements that will
attempt to address all needs but, as noted before,
this is clearly impossible. The water-quality
characteristics listed by the U.S. National Academy of
Sciences as controlling the suitability of water for
the six broad uses that the Academy has addressed,
number close to a hundred.
The often-asked question, "What do you measure, and
why?" should be turned around to first approach the
subject of why one is measuring water quality. Is the
purpose to appraise the water's suitability for drink-
ing, for support of fish life, for irrigation or
industrial use; or is it to monitor for suspected
pollution? If the purpose is to evaluate suitability
for a particular use, the National Academy of Sciences'
compilation is probably the best general source of
information that is available. Determinations
required by recent Federal legislation must be
2
3-3
-------
Figure 2.—Locations of stations in the National Stream Quality Accounting Network in operation as of
January 1, 1975.
included also if the intended water use falls within
their purview. Thus, in most cases a list of
standards or criteria is already available.
Clearly, high costs and limited amounts of money often
will limit the length of the list of chemical
constituents and physical characteristics to be
monitored. A selection must be made of the water use
or uses of primary interest, and the water-quality
characteristics that are most critical to the suitabil-
ity of water for those particular uses.
In some situations it is practical to use broad
"indicator" measurements or to estimate the concen-
tration of certain chemical constituents from measure-
ments of other water-quality characteristics, such as
electrical conductance. Studies of stream quality
records in the United States have revealed many
situations where it is possible to closely estimate
the concentration of dissolved solids and of many
major inorganic constituents from specific conductance
data. In some situations, data on total and dissolved
organic carbon can be used to determine changes in
concentration of a particular organic compound which
is present in relatively high concentrations.
Biological information, such as amounts of or changes
in numbers of cells, biomass, diversity, or other
factors may be used as estimators of water quality,
although in many cases such data do not directly
relate to standards or criteria for use. The U.S.
Geological Survey presently is collecting several
types of biological data in our national networks in
order to provide information that can be used to
evaluate the potential application of biological
measurements for judging the suitability of stream
waters for specific uses.
Frequency of measurement often is decided by a
compromise involving two factors: (1) the need to
define the variations over time of a particular
water-quality characteristic, and (2) limitations
imposed by the amount of money available to make the
necessary measurements. It is necessary, for example,
to make frequent measurements and estimate controlling
factors if one hopes to define the range of concen-
tration of a constituent. If the load, or total
weight of a particular constituent is desired, measure-
ments must be made over a range of streamflow
conditions; and a basis for estimation or indirect
measurement must be developed, because continuous
measurements are practical only for a few water-quality
constituents, such as dissolved oxygen and electrical
conductance.
It is important to consider network objectives in
selecting sampling frequency. Referring back to the
classification of water-data collection (fig. 1),
it may be possible to greatly reduce measurement
frequency for a Level I accounting network once the
variability of a water-quality characteristic has been
defined over the expected range of streamflow, or
after the nature of the variations has been defined—
over time, with water discharge, and with variations
in other characteristics.
On the other hand, measurement frequency cannot so
easily be relaxed for surveillance, which includes
monitoring for hazards and collecting data on the
nature and extent of pollution. Surveillance must,
by definition, be designed to detect the unexpected,
if and when it occurs. Surveillance monitoring usually
involves automatic monitors or samplers, frequent
visits to the monitoring station and, commonly,
immediate, automatic data transmission.
The state of the art in automatic monitoring is
changing every day; but there is one aspect of it that
has not changed in recent years — it is expensive.
Rugged, dependable field monitors are available for
recording temperature and specific conductance. Good
equipment also exists for monitoring dissolved oxygen,
pH, turbidity, and chloride, and for collecting
3
3-3
-------
representative sediment samples; but the equipment
is expensive, requires frequent servicing, and even
when working well commonly results in loss of a
significant fraction of the record. Sophisticated
equipment also exists for monitoring selected ions
and certain organic compounds; but again, cost is
high and reliability varies, especially in the case
of field installations.
Telemetering equipment has been used for many years
in hydrologic monitoring, and it will see far greater
use in critical situations in the future. Telemetry
is expensive, but the cost often can be justified if
there is a real need for immediate information. In
many cases, however, the weak link in a system for
telemetering data is in the sensors and equipment
which supply the basic measurement signal. Without
adequate servicing, the information obtained will be
equivocal, and unusable for regulatory or management
agencies. However, an incidental benefit of
immediate, automatic data-reporting is the advantage
of being immediately aware of malfunctions of field
equipment.
In operating any monitoring network, it is desirable
to strive for uniformity—in space and in time--in
equipment, in methods, and in operating standards.
So far this paper has dealt with field considerations,
but laboratory performance must also be considered.
In the Geological Survey, we have attempted to gain
uniform quality in laboratory data by standardizing
methods; by generally adopting a system of fewer,
larger, and more highly automated laboratories; and
by constantly checking and evaluating analytical
performance through carefully-designed quality control
programs. It has been estimated that quality control
accounts for from 10 to 20 percent of the cost of
operating the Survey's three central laboratories,
which annually process more than 90,000 samples and
make more than 1,000,000 individual physical, chemical
and biological determinations. Beyond this, we are
working with other data-collecting agencies and
technical organizations to develop a National Handbook
of Recommended Methods for Water Data Acquisitions
which will provide for improved comparability,
reliability, and usability of the resultant data.
Reporting Results
When it comes to reporting results of monitoring
networks, the objectives of network operation again
must be considered. The commonly-used tables of
numbers or computer printout often do not meet the
needs of the busy planners and decision makers who
must know the quality of water within their juris-
dictions; or the statistical parameters used for
reporting the water quality information may not
provide the perspecitve the water manager may need
to make his decision.
During the past several months, the Geological
Survey has devoted considerable thought and effort to
determining how to report information on water
quality. An interagency research project, conducted
in cooperation with the U.S. Environmental Protection
Agency (EPA} and the Council on Environmental Quality
(CEQ) is providing comparative evaluations of water-
quality indices for use reporting water-quality
data. Ways of presenting water-quality information
in relation to water-quality criteria or standards
are being studied also. The search for simpler,
more practical and useful ways of reporting water-
quality information will remain a continuing challenge
in the years ahead.
Summary
The many facets of "water-quality," which include
physical, chemical, and biological characteristics
of water, make it impractical and too costly to
design an all-purpose water-quality monitoring
network. Specific objectives of the monitoring
program must be clearly defined before a proper mix
of sampling sites, characteristics to be measured,
and frequency of measurement can be selected. In
addition, one must have some knowledge of the
functioning of the part of the hydrologic system to
be monitored. Presentation of the information
collected through network operations in the form that
is most suited to the needs of water managers and
decision makers is a major challenge for technicians
in the field of environmental sensing and assessment.
References
1. Biesecker, J. E., and Leifeste, D. K., 1971, Water
Quality of Hydrologic Bench Marks - An
Indicator of Water Quality in the Natural
Environment: U.S. Geological Survey Circular
460-E, 21 p.
2. Cobb, E. D., and Biesecker, J. E., 1971, The
National Hydrologic Bench-Mark Network: U.S.
Geological Survey Circular 460-D, 38 p.
3. Ficke, J. F,, and Hawkinson, R. 0., 1975, The
National Stream Quality Accounting Network
(NASQAN) - Some Questions and Answers: U.S.
Geological Survey Circular 719, 23 p.
4. Langford, R. H., and Davis, G. H., 1970, National
System for Water pata: Am. Soc. Civil
Engineers Proc., "Hydraulics Div. Jour., v. 96,
no. HY 7, paper 7392, p. 1391-1401.
5. U.S. National Academy of Sciences and National
Academy of Engineering, Environmental Studies
Board, 1972, Water Quality Criteria 1972:
Washington, D.C., U.S. Environmental Protection
Agency rept. EPA.R3.73.033, Mar. 1973, 594 p.
4
-3
-------
SOME INTERNATIONAL ACTIVITIES IN ENVIRONMENTAL
MONITORING
Guntis Ozolins
World Health Organization
Geneva, Switerland
During the past several years, a number of inter-
national organizations - both inside and outside the
UN system - have become increasingly involved in
issues concerning the environment and with the moni-
toring of environmental quality. This increasing
interest stems from a number of factors. As the
awareness of the hazards associated with environmental
pollution grew at the national level and national
pollution control programmes came into being,
there was a demand for similar development to take
place on the international scene as well - to
establish programmes through which assistance could
be provided in strengthening national pollution
control efforts.
The realization that many of our environmental
pollution problems have regional and even global
aspects further accentuated the need for inter-
national collaboration in the study and control of
the pollution problem. The increasing levels of
carbon dioxide and fine particulates in the atmos-
phere and the accumulation of persistent pollutants
in the environment are some of the examples for
global concern. The transport of pollutants from
one country to the next through common waterways and
through long-distance atmospheric transport are
further examples of the need for international
concern and involvement.
A notable milepost in the development of inter-
national monitoring programmes was the 1972
Stockholm Conference on the Human Environment and
the formation of the United Nations Environment
Programme (UNEP). This is not to say that the
various agencies did not have some monitoring
activities underway before, but only that the
formation of UNEP intensified these efforts
and provided a central coordinating point for them.
The full development of UNEP's Global Environmental
Monitoring System (GEMS) will bring together and
expand the existing monitoring programmes in air,
water, etc. to furnish a more balanced and complete
view of the global problem. The objective of GEMS
Is to provide a collection of measured data on
various parameters characterizing the environment,
including not only pollutants in all media, but
parameters such as those concerned with arable
land, desertification, water quality, forest area
and disease, etc.
The purpose of this presentation is to briefly
outline the primary obje^tives of the international
environmental monitoring programmes and to give an
overview of the major activities that are ongoing
or are being implemented. The discussion is mostly
limited to the UN agencies. It must be recognized,
however, that other international bodies outside
the UN system are also very active in this field
and that close collaboration with these organizations
is maintained. Examples are governmental agencies
such as the Council of European Communities (CEC),
the Council for Mutual Economic Assistance (CMEA.) ,
and non-governmental bodies such as the International
Council of Scientific Unions (ICSU) and its
Scientific Committee on the Pollution of the
Environment (SCOPE).
Objectives
The objectives of international environmental
monitoring programmes, whether they are concerned with
the monitoring of air, water, soil, food or the working
environment, are quite similar. These objectives,
simply stated, are:
1. to assist national governments in developing
and strengthening their own national or local
programmes;
2. to provide the coordinating mechanism for
increasing the uniformity and comparability
of measurements that are made;
3. to collect, analyze and disseminate data of
international or regional and global signi-
ficance.
Direct assistance to Member States in developing
monitoring systems has been and remains to be a top-
priority objective for most of the international organ-
izations. Assistance is normally rendered through
internationally supported country projects and through
short-term aid. Under these programmes, training is
provided to the national staffs, assistance is given
in obtaining instrumentation and equipment, and expert
consultants are utilized in solving specific problems.
In recent years a good number of environmental pollu-
tion monitoring programmes have thus been established
or stengthened.
The need to harmonize measurement methodology is
becoming more and more Important, and many interna-
tional organizations now have projects underway to
develop more uniform and compatible approaches to
environmental quality monitoring. This is particu-
larly necessary where groups of countries share a com-
mon water resource or where air pollution crosses
international boundaries. Many working groups and
committees have been established to deal with this
subject matter. Similarly, international seminars,
symposia and conferences are being organized with the
aim of information exchange and reaching consensus on
approaches and methodologies.
The third primary objective is to provide inter-
national syntheses of information on environmental
quality, where such information is needed. For the
most part, the collection of data by the international
organizations depends on already existing national
data gathering network and only seldom are moni-
toring systems installed directly by these organiza-
tions. Thus, when mention is made of an "international
network", it does not mean that an international agency
is physically involved in monitoring. In most in-
stances, these "networks" are composed of already
existing monitoring stations which contribute their
data for compilation and synthesis.
These three objectives provide the basic frame-
work under which the international monitoring pro-
grammes generally operate. Some of the programmes
are well underway and have made substantial contri-
butions while others are in various stages of imple-
mentation.
3-4
-------
Air Quality Monitoring
The principal UN agencies engaged in the moni-
toring of air quality are the World Meteorological
Organization (WMO) and the World Health Organization
(WHO). Whereas WMO's primary Interest is to observe
global changes in the atmosphere.the WHO programme is
directed towards the monitoring of air quality at the
impact level to assess the human exposure to air pol-
lution.
The WMO programme, which dates back to 1970, in-
cludes the operation of a global network of stations
as well as projects dealing with the development and
standardization of the related measurement methods.
The WMO network, which is currently being implemented
and expanded, contains two types of stations - base-
line and regional. The baseline stations are located
in remote areas and are operated to provide informa-
tion on long-term trends in atmospheric composition.
At these points, measurements are minimally made of
turbidity, carbon dioxide and chemical composition of
precipitation. Where possible, other contaminants
vill be measured as well. A total of 20 such stations
are planned, with 9 being operational now. The re-
gional network is operated to provide information on
atmospheric changes due to regional land use prac-
tices or other activities. This network also measures
precipitation parameters and now comprises some 100
stations in 61 countries. It is planned to increase
the network by 50 stations.
MHO, on the other hand, has concentrated its
efforts on monitoring air quality in highly populated
areas in order to measure and assess risks of human
exposure. An air monitoring study has been in opera-
tion since 1972. In this study, data on sulfur
dioxide and suspended particulate are befng collected
from about AO stations in 14 countries. An inter-
laboratory comparison study is also currently under-
way. Plans are now being made to expand this opera-
tion to include other countries and other pollutants
as well. In addition, the Pan American Health
Organization operates an air pollution monitoring
network which encompasses some 100 stations in 14
countries.
Both organizations have also been engaged in the
preparation of manuals and notes on measurement
methodology. These manuals aim at increasing the
comparability of the measurements that are made,
Mention should also be made in this regard of the
International Standards Organization (ISO) where a
working group on measurement methods for air quality
was established in 1972.
Water Quality Monitoring
Because of the multitude of uses of water
resources, it is not surprising that many agencies,
both at the national and the international level, are
involved in water quality monitoring activities. The
Protection of water for human consumption, recreation,
food production, fisheries and industrial uses spans
the responsibilities of many agencies. On the inter-
national scene, the Food and Agriculture Organization
Of the United Nations (FAO), the United Nations
Educational, Scientific and Cultural Organization
(UNESCO), WMO, WHO and the UN regional economic com-
missions and, of course, UNEP, are all involved to
varying degrees in water quality monitoring and re-
lated activities.
The monitoring of drinking water, although rou-
tinely conducted in the more industrialized countries,
is still not sufficiently developed elsewhere. There
2
are still large populations, particularly in the rural
areas, which use water of uncontrolled bacteriological
and chemical quality. WHO statistics for 1970 show
that out of 91 developing countries, only 30 provide
regular bacteriological examination of their urban
supplies. The situation is much worse in rural areas
where 20 countries have no surveillance at all. WHO
has been trying to stimulate the national programmes
by providing direct assistance in establishing sur-
veillance programmes and in training monitoring per-
sonnel . To this end, a training manual on surveillance
of drinking water quality is being prepared. In addi-
tion, a pilot programme for the measurement of selected
contaminants on a worldwide basis has been started.
The aim is to develop an internationally accepted
methodology for establishing baseline levels of hazard-
ous substances in drinking water.
The international organizations have also been
active in the monitoring of surface water quality.
The FAO is deeply involved in providing technical
assistance and guidance on the development of sur-
veillance programmes for water quality in relation to
commercial fisheries. In this effort, FAO has pub-
lished a series of guidelines dealing with water qual-
ity criteria for various parameters of importance to
freshwater fisheries as well as the development and
harmonization of measurement methodologies. FAO is
also pursuing projects in water quality monitoring
associated with agriculture and particularly the use
of irrigation waters.
The interests of WMO in water monitoring activi-
ties are primarily directed towards operational hydro-
logy. It is mainly water quantity and flows rat her
than water quality which are in the foreground of their
efforts. They have, however, also addressed soma
aspects of water pollution monitoring. Design of water
quality networks in relation to existing hydrometric
stations and methods for water quality measurements
with respect to physical and chemical parameters are
the two major components of the WMO programme in this
field.
UNESCO's efforts have been concentrated primarily
on research aspects of the problem and on the coordin-
ation of the International Hydrological Decade (IHD)
and more recently on their involvement in the Inter-
national Hydrological Programme (IHP). Under IHD,
many countries selected hydrometric stations which
now routinely supply discharge data to UNESCO. Under
this same programme, UNESCO together with WHO has pre-
pared a guidebook for water quality surveys. Under the
IHP, the influence of diffusion, dispersion and self-
purification processes on the selection of monitoring
sites in surface waters will be investigated. Futher-
more, UNESCO's programme,Man and the Biosphere (MAB)y
also has some monitoring components. Routine moni-
toring will not be conducted under this programme,
but investigations will he made on identifying the
critical variables which need to be monitored.
WHO has also undertaken some activities in the
monitoring of surface water quality. Considerable
assistance has been given to Member States in develop-
ing national water pollution monitoring programmes,
In many instances these activities are part of the
United Nations Development Programme country projects.
Work has also been started in the harmonization of
measurement methods with a manual on analytical
methods to be published shortly. A project to har-
monize monitoring programmes on a siajor international
river basin - the Danube - has also been starred.
Considerable attention has been given by the
national and international authorities in recent years
3-U
-------
to the pollution of the seas and oceans and the need
to develop adequate monitoring programmes. The
Intergovernmental Oceanographic Commission (IOC), the
Intergovernmental Maritime Consultative Organization
(IMCO) and others already mentioned are involved to
varying degrees in studying the pollution of oceans
and the seas. A number of working groups and scien-
tific committees have been established by these organ-
izations to address various parts of the marine pol-
lution problem, including its monitoring needs. The
Joint Group of Experts on Scientific Aspects of
Marine Pollution (GESAMP) is one such committee. In
addition, the Global Investigation of Pollution of
Marine Environment (GIPME) and the Integrated Global
Ocean Station System (1G0SS) are concerned with the
monitoring of the marine environment. A pilot project
on marine pollution monitoring was started in 1975
which includes observing oil slicks, clumps of oil
in the open seas and on the beaches, and dissolved
hydrocarbons.
In addition, a number of projects are being
handled by the FA0 in relation to deep-sea fisheries
and agriculture, and by UNESCO with regard to pollu-
tion contributed by large rivers to the marine envi-
ronment. This registry of rivers will represent
about 80% of all the discharges by rivers into oceans
and seas. The pollutants to be measured will be
selected according to the Oslo, London and Paris
conventions. Implementation will take place on a
tegion-by-region basis starting with the Mediterranean
and the Baltic Sea.
WHO's concern has been primarily in the field of
coastal water quality as regards bathing and recre-
actional activities, as well as shellfish growing and
processing.
Food Monitoring
The question of food contaimination is being
addressed by FAO and WHO. These agencies are col-
laborating in establishing an internationally coordi-
nated programme for monitoring those contaminants in
foods that are known to present a hazard, to human
health. Initially, the programme will be concerned
with mercury, cadmium, lead, organochlorine compounds,
mycotoxins and biological contaminants. This pro-
gramme will be based on data provided by countries in
which national monitoring and food survey programmes
exist. Assistance will also be provided in estab-
lishing such programmes in other countries. The data
will be submitted to FAO/WHO for collation and evalu-
ation by specially convened meetings of experts. The
conclusions drawn will be used by the Codex
Alimentarius Commission to elaborate limits for con-
taminants for inclusion in international food stan-
dards. It should also be mentioned that FAO and WHO
have made substantial progress in establishing accep-
table methods of analysis for some chemical contami-
nants in food.
Monitoring of Radiation and Other Physical Hazards
The monitoring of environmental radioactivity
has been of interest to International organizations
for quite a while. The United Nations Scientific
Committee on the Effects of Atomic Radiation (UNSCEAR),
which was established in 1955, has been responsible
among other things for collecting, evaluating and
disseminating information on the observed levels of
ionizing radiation and radioactivity in the environ-
ment, and for recommending standards for sample col-
lection and Instrumentation.
The International Atomic Energy Agency (IAEA) has
been actively Involved for some time in promoting the
development of environmental monitoring programmes for
radioactive contaminants. In this regard, the IAEA
has published a series of guideline manuals dealing
with the design and operation of environmental moni-
toring programmes.
Since 1960, the IAEA in cooperation with WMO,
has operated a global network of stations for moni-
toring selected radionuclides in precipitation. At
present, 122 such stations are engaged in this data
collection network all over the world. Measurements
are typically made of tritium, deuterium and oxygen-18.
Some monitoring of environmental levels of radio-
nuclides has also been carried out by WHO in collabora-
tion with its collaborating center. The main activi-
ties include an interlaboratory comparison programme
and the collection of information on levels of
strontium813 ant^ cesium 137 and some trace elements
in milk.
The monitoring of noise is receiving increased
attention. WHO has undertaken a project to establish
guidelines for monitoring noise in urban areas, and
several projects are being supported at the national
level. Similarly, some work is being pursued by WHO
in monitoring of the working environment. Guidelines
for preliminary surveys and control of the working
environment factors have been prepared.
Monitoring of Human Exposure and Health Effects
The assessment of human exposure and the associ-
ated health effects has not received much attention
until very recently. Integrated monitoring systems
to provide information on the relative contributions
of various pathways to the total exposure need to be
developed. WHO has started activities in this direc-
tion as well as In the measurement of contaminants in
human tissue and fluids. The methodology for moni-
toring of environmental carcinogens is similarly
being discussed and developed under the auspices of
WHO and the International Agency for Research on
Cancer, as well as a number of national institutions.
Conclusion
From the foregoing, it is evident that-a consid-
erable amount of work is going on in the monitoring
field. Many agencies are Involved and many diverse
subject areas are being addressed. Substantial pro-
gress has been made, particularly in the area of
measurement methodology, where a series of manuals
have been put togethet by scientific groups. Their
application will no doubt move us forward to the use
of more uniform methods which will facilitate the
exchange of information.
The development of national monitoring programmes
is progressing at an extremely variable pace. In some
countries, such programmes are well established, while
in many others measurements of environmental quality
are not routinely made. The level of economic develop-
ment, the severity of the pollution problem and the
national priorities dictate the rate at which these
programmes progress. The international organizations
have a very instrumental role to play in providing
assistance as needed.
In closing, the point should be made that the
sources of international programmes depends almost en-
tirely on the willingness of national governments, of
institutions and of individual experts to cake part in
3
-------
bringing together the information and expertise which
is needed to further pollution control in all parts
of the world.
4
3-4
-------
EARTHUATCH-SENTINEL FOR THE FUTURE
Clayton E. Jensen and Dail W. Brown
National Oceanic and Atmospheric Administration
Rockville, Maryland
This paper describes the United Nations Environ-
ment Program (UNEP) that had its inception at the 1972
Stockholm Conference on the Human Environment. The
paper focuses upon the global environmental assessment
program—EARTHWATCH—which is a major element of UNEP.
Guidelines are presented for implementing EARTHWATCH.
These include suggested actions for global monitoring,
research, evaluation and information exchange.
A Beginning in Stockholm
The United Nations Conference on the Human Environ-
ment was held in Stockholm, Sweden, in June 1972. One
hundred and thirteen nations met to translate their
environmental concerns to a plan for action to preserve
and enhance the human environment. The highlights of
the action proposals, since implemented, include (a) a
United Nations Environment Program (UNEP) headed by an
Executive Director; (b) a voluntary UNEP fund of $100
million over 5 years; (c) a UNEP Governing Council of
58 nations; and (d) a United Nations interagency
coordination board for UNEP.
The Stockholm Conference envisioned that UNEP
would carrv out its 109 specific recommendations that
were adopted after careful deliberation. These recom-
mendations were cast into a UNEP framework composed of
three major parts: environmental assessment, environ-
mental management, and supporting measures. The desig-
nation of EARTHWATCH was adopted for the environmental
assessment part of the program that would provide the
basis for responsible environmental management. EARTH-
WATCH is designed as a four-part program involving
monitoring, research, evaluation, and information
exchange.
Environmental management consists of the setting
of goals and criteria and the consummation of inter-
national agreements and conventions.. The UNEP frame-
work for action was completed with a third element con-
sisting of education and training, technical assistance,
and public information which supports both environ-
mental assessment and management activities. In
addition to these functional activities, seven priority
program areas were recognized that addressed critical
global environmental needs: (a) Human Settlements and
Habitat; , (b) Health of People and of the Environment;
(c) Terrestrial Ecosystems; (d) Environment and Develop-
ment; (e) Oceans; (f) Energy; and (g) Natural Disasters.
Since the time of the United Nations General Assem-
bly endorsement of the UNEP Action Plan, there have been
three meetings (June 1973 in Geneva, and March 1974 and
April 1975 in Nairobi) of the UNEP Governing Council
with intervening intergovernmental meetings on various
topics including one on monitoring in Nairobi, Kenya,
February 11-20, 1974. This latter meeting took great
strides toward the design and implementation of the
Global Environmental Monitoring System (GEMS)—a major
step in the evolution of EARTHWATCH, Among the accom-
plishments of the meeting were the specification of
principles, the identification of program goals, pri-
ority pollutants and related environmental factors, and
the recommendation of a program of future work and
institutional arrangements.
The principles that were adopted governing inter-
governmental cooperation in monitoring include (a)
building upon existing national and international
systems to the maximum possible extent; (b) providing
assistance where necessary, especially in the field of
training and equipment, to ensure effective involvement
of the developing countries; and (c) sharing the respon-
sibility for implementing international monitoring
systems in areas outside national jurisdiction, such as
oceans and space.
Program goals provide the focus for GEMS so that
it can be responsive to priority subject areas of the
United Nations Environment Program. These program
goals are:
o An expanded human health warning system;
o An assessment of global atmospheric pollution
and its impact on climate;
o An assessment of the extent and distribution of
contaminants in biological systems, particularly food
chains;
o An assessment of critical environmental problems
relating to agriculture and land and water use;
o An assessment of the response of terrestrial
ecosystems to pressures exerted on the environment;
o An assessment of the state of ocean pollution
and its impact on marine ecosystems; and
o An improved international system allowing the
monitoring of the factors necessary for the under-
standing and forecasting of disasters and the imple-
mentation of an efficient warning system.
A priority list of pollutants was established.
The design of GEMS must take selected elements of this
list into account along with appropriate related
environmental factors so that the respective program
goals may be achieved. For.example, in approaching
the assessment of the impact of atmospheric pollution
on climate, GEMS must not only provide a system for
monitoring relevant pollutants, such as SO2, suspended
particulates, O3, N0X, and C02> but also for monitoring
climate indicators, such as areal extent of sea-ice,
the advance and recession of glaciers, sea level
change, drought, desertification and changes in fresh-
water bodies. In this way, cause and effect relation-
ships may be critically investigated.
EARTHWATCH Monitoring
The essential elements of the monitoring portion
of EARTHWATCH are global environmental observations,
data processing and analyses, and communications.
Observations
Observational programs now underway provide the
foundation for international warning, prediction and
assessment of health hazards, natural disasters and
other potential environmental concerns. The World
Health Organization (WHO), for example, coordinates
international programs for the surveillance and
issuance of warnings relating to communicable diseases
and the adverse effects of drugs. It also coordinates
programs for the monitoring of air quality, environ-
mental radiation, and community water supplies in
cities and Industrialized areas. These programs help
to identify and forecast the levels and trends of
specific air pollutants, such as sulphur dioxide,
-------
suspended particulates, ozone, and oxides of nitrogen,
as well as their effects on the health of the people.
The World Heather Watch (WWW) of the World Meteo-
rological Organization (WHO) provides a comprehensive
means for detecting, locating, and tracking weather
systems and for issuing timely warnings and predictions
of potential natural disasters, such as floods and
tropical cyclones. The WWW includes some several
thousand land stations and merchant ships, numerous
aircraft, special ocean weather ships, an increasing
number of automatic weather stations, ocean buoys, and
environmental satellites. The operation of this system
exemplifies the benefits to be derived from an inte-
grated global observing program. This concept is
currently being extended to include the observing of
oceanic conditions through the development of the Inte-
grated Global Ocean Station System (IGOSS) by the Inter-
governmental Oceanographic Commission (IOC) which will
include the common use of facilities, sensors, and
platforms such as ocean buoys, ships, and satellites.
Currently, several baseline stations are being
operated by tke United States to monitor background
atmospheric constituents including carbon dioxide,
ozone and particulate matter. Additional stations are
planned by many countries as part of a proposed WflO
global network. A systematic program to record base-
line conditions for other aspects of the environment
has not yet been initiated. The Man and the Biosphere
program, however, provides an appropriate framework
for the global monitoring of terrestrial ecosystems.
Several observing programs of regional and
national scope now underway can provide further support
to a global network. These include:
o FAO's monitoring of agricultural and fisheries
resource utilization;
o WMO's regional air quality monitoring network;
o The International Hydrological Program (IHP) of
the United Nations Educational, Scientific, and
Cultural Organization (UNESCO);
o The International Tsunami Warning System of the
IOC;
o The Worldwide Standardized Seismic Network; and
o The WMO Tropical Cyclone Project.
Global surveys provide the most comprehensive
assessment of the condition and availability of natural
resources and the impact of management practices. The
high resolution, multi-spectral sensors of research
satellites provide global information on such environ-
mental processes as land use patterns, the status of
terrestrial ecosystems, hydrological conditions, and
oceanic phenomena.
Pata processing and analysis
Under the World Weather Program, the WMO coordi-
nates a global network of world, regional, and national
meteorological processing centers. These centers can
provide the basis for an interdisciplinary data pro-
cessing service. A beginning has been made in this
direction with the processing of oceanographic infor-
mation collected by the IGOSS pilot project. The role
of the centers can be expanded further to process a
broader range of environmental data. Examples of other
international centers for the processing and analysis
of data and information include the International
'Tsunami Warning Center, and the system of national and
regional oceanographic sorting centers that provide
service to the international community.
¦Communications
Communications for the exchange of data must be
shared by multidisciplinary activities for efficiency
and economy. A likely candidate for this is the global
telecommunications system (GTS) of the WWW. The GTS
links 25 Regional Telecommunications Network Hubs that
serve three World Meteorological Centers in Washington,
Q.C., Moscow (U.S.S.R.), and Melbourne (Australia), 25
Regional Meteorological Centers, and approximately 150
national meteorological centers. This system could be
considered as a prototype for the global environmental
communications network required for EARTHWATCH.
The Geostationary Operational Environmental
Satellite (GOES) system will provide a new capability
to the international communications system. Signals
from remote monitoring equipment such as automated
hydrological, atmospheric, oceanographic, and terres-
trial facilities can be received by the satellite and
relayed to processing centers via the GTS.
EARTHWATCH Research and Information Exchange
A number of major research programs has been initi-
ated or are being planned by the United Nations agen-
cies. The Global Atmospheric Research Program (GARP),
jointly developed by the WHO and ICSU is aimed at pro-
viding scientific knowledge needed to improve the time
range, scope, and accuracy of weather forecasts, and to
gain a better understanding,'of the physical basis of
climate.
The Long-Term and Expanded Program of Oceanic
Exploration and Research is a comprehensive program
coordinated by the IOC to study the processes of the
world's oceans. The acceleration phase of this
research has been designated the International Decade
of Ocean Exploration. Another major part of the pro-
gram—the Global Investigation of Pollution in the
Marine Environment—is expected to provide needed
information on pollution sources, pathways, and sinks
within the marine environment and the impact of pollu-
tants on biological systems.
The Han and the Biosphere (MAB) Program of UNESCO
is designed to identify and evaluate the impact of
man's activities on the biosphere, to study and compare
the structure, function, and -dynamics of natural, modi-
fled, and managed ecosystems, and to develop methods to
monitor significant trends in order to assess manage-
ment alternatives.
The International Hydrological Program under the
coordination of UNESCO is designed to develop a better
understanding of processes that influence global water
resources. Other research activities include the
environmental aspects of disease prevention and cure
programs of the WHO, and investigations of agricul-
tural, forestry, and fisheries processes coordinated by
FAO.
A major existing activity is the World Data Center
system of ICSU which provides the international commu-
nity with data storage and dissemination services for
the following: meteorology, geophysics, geochemistry,
aurora and airglow, ionospheric physics, solar activity,
cosmic rays, longitude and latitude, glaciology,
oceanography, seismology including tsunamis, and gra-
vimetry. In addition, the FAO archives and dissemi-
nates data and information relating to the production
and utilization of agricultural, forestry and fisheries
resources.
Implementation of EARTHWATCH
A programmatic approach comprising tha areaa
previously identified at the Intergovernmental Meeting
on Monitoring (and noted above) is propoaad as a
rational scheme for the evolution of an «££ectiv«
2 3-3
-------
EARTtlVIATCH. Program. The framework for this program-
matic approach calls for the designation or establish-
ment of three levels of assessment activities—world,
regional and national—to which governments, United
Nations agencies and non-governmental organizations are
expected Co contribute. An implementation strategy is
proposed that consists of two parallel action streams.
The first stream should utilize proven technology
and accepted scientific procedures and techniques for
the immediate improvement and integration of the capa-
bility of existing international systems to observe,
communicate, process, and analyze information on world-
wide environmental conditions. The second stream
should develop new technology, procedures, techniques
and facilities so that the unmet needs for global
environmental assessment may be adequately met.
The proposed framework for EARTHWATCH calls for
the designation of World Environmental Assessment
Cetvtets for each of the seven program areas that have
been identified. Regional Environmental Assessment
Centers are required to support the World Centers; the
Regional Centers, in turn, are supported by national
facilities. Assignments would need to be coordinated
and acceptance of respective responsibilities would
have to be gained from nations to operate the World and
Regional Environmental Assessment Centers and the
supporting national facilities. To facilitate partici-
pation by developing nations in this effort, assistance
could be sought from UNEP or from other sources for
funding and technical support.
The function of each World Environmental Assess-
ment Center would be to provide the focus and capa-
bility for intensive assessment of the worldwide
environmental conditions relevant to its assigned pro-
gram area of responsibility. Along with the associated
Regional Environmental Assessment Centers that would be
concerned with environmental problems having regional
impact, the World Center would share the responsibility
for maintaining quality control, developing standard
reference materials, insuring full intercalibration of
analytic procedures, and instrumentation, and serving as
focal points for training and education programs.
The World Centers would be expected to issue
periodic assessment reports and special alerts, as
appropriate, on the basis of which UNEP may wish to
organize international actions, or individual nations
may wish to respond in some manner.
National facilities would be expected to concen-
trate upon the analysis of environmental samples and
data collected as a part of the monitoring programs at
reference sites (long-term baseline and trend infor-
mation) , impact stations (immediate and near-term
information and trends), and indicator programs.
Proper management of data and information is
essential to Insure its availability both currently and
in the future for analyses and the preparation of
environmental assessments. Immediate steps are
necessary to designate a network of national, regional,
and world centers for data and information management
based upon existing facilities and to identify where
gaps occur. Further, efforts should be taken to use
existing or planned communications circuits or methods
such as the VIorld Weather Watch global telecommunica-
tions system, and the international geostationary
operational environmental satellite system.
EARTHWATCH Reference Sites would be expected to
provide a coherent, integrated base of benchmark data
and information otl physical, chemical and biological
conditions and trends in environmental processes.
These sites should include selections from ongoing and
planned activities, such as (a) WMO baseline, and
upper atmospheric programs; (b) 1HP hydrology stations;
(c) lake biome programs; (d) MAB bioroe programs; (e)
open ocean baseline sites; and (f) river outflow
stations.
EARTHWATCH Impact Sites would be expected to pro-
vide the basis for a continuing, systematic appraisal
of immediate and near-term impacts of natural and
anthropogenic activities on the environment and to
allow timely warnings of environmental conditions that
threaten man and the biosphere. These sites should
include selections from ongoing and planned programs,
such as (a) regional marine monitoring activities; (b)
climatic index stations; (c) health effects and expo-
sure programs; (d) land use monitoring activities; (e)
natural disaster warning and preparedness programs; and
(f) regional atmospheric monitoring stations.
EARTHWATCH Indicator Programs would be expected to
begin with the designation of a series of target sub- *
stances or organisms to be included where feasible in
all reference and impact monitoring activities. Three
pilot projects are proposed that would focus on fish,
lichens and human hair.
Effective environmental management to mitigate the
impact of pollution on the environment and man will
require a detailed understanding of sources, pathways,
and sinks. EARTHWATCH can provide the data and infor-
mation that will allow tracing a pollutant from its
sources through the atmosphere, oceans, soils, fresh
water, and biological systems to its ultimate fate. A
pilot program is proposed that would focus on a single
pollutant; e.g., lead, and track it through the
environment. Particular emphasis should be placed on
data acquired at the EARTHWATCH Reference Sites and
within the impact monitoring and indicator programs.
This pilot project would be useful for developing the
techniques that could be applied in future programs
involving other critical substances.
The future growth of global environmental assess-
ment activities will call for full support to relevant
research programs and proper scientific advice from
non-governmental bodies. This will require the timely
planning of both regional and global environmental
assessment research Involving the cooperative^partici-
pation of all concerned nations.
The implementation of EARTHWATCH will call for the
broadest international participation and cooperation.
The months ahead will be critical to the future via-
bility of EARTHWATCH; and the viability of EARTHWATCH
may be critical to the future well-being of the earth
and its inhabitants.
3-5
-------
NETWORK DESIGN CONSIDERATIONS FOR
THE GLOBAL ENVIRONMENTAL MONITORING SYSTEM (GEMS)
OF THE UNITED NATIONS
By Robert Citron
Smithsonian Institution
Cambridge, Massachusetts
INTRODUCTION
The United Nations Environment Program has a man-
date to develop a Global Environmental Monitoring
System as part of its EARTHWATCH Program. One of the
first priorities of GEMS is to coordinate and support
existing and planned monitoring programs that can pro-
vide data and information that may be used to assess
the states and trends of global pollution.
The first step toward developing global networks
to monitor priority pollutants was to undertake a
worldwide survey of existing, long-term pollution pro-
grams. During the past two years the Smithsonian
Institution, under contract to the United Nations En-
vironment Program, undertook such a survey and compu-
terized the'results to facilitate evaluation of exist-
ing programs and to develop techniques for network
design.
The resultant data base includes information on
the administrative and operational characteristics of
more than 700 individual pollution monitoring programs
operating over 66,000 pollution monitoring sites in
78 countries and all of the oceans of the world.
The purposes of this paper are to summarize the
kinds of data and information that are contained in the
Smithsonian Pollution Monitoring Program Data Base, to
outline the requirements for the establishment of
global pollution monitoring networks, and to begin to
use the data contained in the base to assist in the
design of a hypothetical global network that can be
used to monitor toxic heavy metals and chlorinated
hydrocarbons in specified media.
POLLUTION MONITORING PROGRAM DATA BASE
The data base is designed to be used as a basic
reference guide in the planning of integrated networks
of pollution monitoring sites that would systematically
measure pollutant levels throughout the world. It con-
tains information on the purpose of each program, how
the programs are administered, and on the pollutants
they monitor. ' In addition, information is provided on
the physical medium monitored (including air, soil,
marine water, fresh water, drinking water, plants,
animals, food, and man), the geographical areas covered,
the number of sites in each program, their sampling
frequencies, their means of sample or data acquisition,
their measurement techniques, their methods of sample
analysis, precision of analysis, mode of data storage,
and the date the monitoring programs began. The lati-
tudes and longitudes of each monitoring or sampling
site location in each program is contained, when
available, in the data base.
The data base can also be used to identify gaps
in existing networks for any one of the priority
Pollutants in any specified medium. It can supply
answers to specific questions about the data. For
example, one may request a list of programs monitoring
a specified pollutant in a given medium in a certain
group of countries. The data base also can supply
statistics on the number of programs sharing the same
monitoring characteristics such as monitoring methodo-
logy. sampling frequency, medium monitored, analysis
technique, etc. Furthermore, a computerized
Plotting machine can draw world or regional maps
indicating the location of sites selected
according to common characteristics. Plot maps and
computer printouts will greatly simplify the design
and implementation of planned global pollutant moni-
toring networks.
Table I summarizes the total number of programs
and the total number of sites that are currently mon-
itoring 25 priority pollutants in all of the 78 coun-
tries contained in the data base.
Table II summarizes the total number of programs
and the total number of sites that are currently mon-
itoring in 11 media in all countries in the data base.
Table III summarizes the total number of programs
and the total number of sites currently monitoring the
25 priority pollutants for each country contained in
the data base.
While the data base does not contain information
on all pollution monitoring programs in all countries
known to have major pollution monitoring networks, it
does represent a first approximation of the world's
existing pollution monitoring resources. The data and
information in the base should be continuously updated,
expanded, and verified in order to obtain a more com-
prehensive picture of the pollution monitoring now
going on throughout the world.
A SAMPLE GLOBAL NETWORK DESIGN FOR MONITORING
MERCURV. lEAb, cAEMuM. m CHLORINATES HYftROCARSONS
AT EXISTING FACILITIES
The data base indicates that there are at least
160 programs operating over 10,000 sites in 44 coun-
tries that are currently monitoring toxic heavy metals
and chlorinated hydrocarbons in air, fresh water,
marine water, drinking water, soil, plants, food,
animals, and man. Table IV tabulates the countries
now monitoring these pollutants and summarizes the
number of programs and the number of sites monitoring
each pollutant by country. The data indicate that
there are 140 programs in.37 countries operating
5,414 sites to monitor levels of mercury in all media,
there are 166 programs in 37 countries operating
11,096 sites to monitor levels of lead; there are
136 programs in 30 countries operating 10,083 sites
to monitor cadmium, and there are at least 96 programs
in 30 countries that monitor chlorinated hydrocarbons
at 3,855 sites.
Table V outlines the total number of pollution
monitoring programs and the total number of sites they
operate by pollutant and by geographical area. The
data indicate that in Europe there are at least 17
countries that operate over 2,000 sites in nearly 100
programs to monitor toxic heavy metals and chlorinated
hydrocarbons 1n all media. In Asia at least 13 coun-
tries operate over 5,000 sites in several dozen pol-
lution monitoring programs and in North America three
countries operate over 3,500 sites 1n more than 50
programs to monitor these pollutants. Data and infor-
mation are sparse from Africa, South America, and
other developing areas.
A SINGLE COUNTRY
The data base can supply descriptions of a
country's monitoring resources for any group of prior-
ity pollutants. Table VI summarizes the toxic heavy
metal and chlorinated hydrocarbon monitoring programs
of one country, Canada, as an example. Canada operates
1
3-6
-------
TABLE I
TABLE III
WORLDWIDE POLLUTION MONITORING PROGRAMS
AND~SITES BY PARAMETERS
WORLDWIDE POLLUTION MONITORING PROGRAMS
And sites by country
STATISTICAL BREAKDOWN BY parameter
STATISTICAL BREAKDOWN BY COUNTRY
15887 SITES
2702* SITES
3020 SITES
1106 I SITES
6981 SITES
AIR-bORNE SULPHUR DIOXIDE AND SULPHATES
167 PROGRAMS
SUSPENDED PARTICULATE MATTER
240 PROGRAMS
CARBON MONOXIDE
71 PROGRAMS
CARBON DIOXIDE
74 PRGGRAM5
AIR-EORNE OXIDES OF NITROGEN
80 PROGRAMS ji.wj
CZONE. PHOTOCHEMICAL OXIDANTS AND REACT IVt HYDROCARBONS
109 programs nils sites
ker'cury
145 PROGRAMS 17661 SITES
lead
130 PROGRAMS 25725 SITES
CAOTIUM
89 PROGRAMS 22541 SITES
HALOGENATED ORGANIC COMPOUNDS. ESPECIALLY DDT AND PCB.S
113 PROGRAMS 14475 SITES
ASBESTOS
7 programs 134 sites
PETROLEUM HYDROCARBONS
71 PROGRAMS 20548 SITES
TCXIN5 OF BIOLOGICAL ORIGIN (FROM ALGAE, FUNGI. 6ACTERIA)
33 programs 11279 sites
NITRATES AND NITRITES
150 PROGRAMS 19076 SITES
AMMONIA
102 PROGRAMS 16834 SITES
BIOLOGICAL OXYGEN DEMAND
144 programs 18374 sites
dissolved OXYGEN
186 PROGRAMS 20401 SITES
PH
214 PROGRAMS 30542 SITES
col I form bacteria
122 PROGRAMS 22898 SITES
selected radionuclides
180 PROGRAMS 14417 SITES
soluble salts of alkali metals and alkaline earth metals
185 PROGRAMS 24521 SITES
EUTRCPHICATORS (NITRATES and PHOSPHATES)
129 PROGRAMS 19427 SITES
OTHER SUBSTANCES,E.g. ARSENIC,BORON,SELENIUM,FLUORIDE
144 PROGRAMS 31830 SITES
NOISE
8 PROGRAMS 470 SITES
WASTE HEAT
5b PROGRAMS 6897 SITES
TABLE II
WORLDWIDE POLLUTION MONITQRIN PROGRAMS
AND SITES BY MEDIUM
AIR
326
PROGRAMS
16103
SITES
50R
90
PROGRAMS
9662
SITES
MARINE WATER
145
PROGRAMS
15837
SITES
FRESH WATER
278
PROGRAMS
32246
SITES
OR INK INS WATER
46
PROGRAMS
6440
SITES
PLANTS
126
PROGRAMS
8267
SITES
ANIMALS
143
PROGRAMS
$432
SITES
FOOD
56
PROGRAMS
3034
SITES
MAN
2)
PROGRAMS
257
SITES
FRESH WATER (RAIN) 9
PROGRAMS
232
SITES
SEWAGE
2
PROGRAMS
18
SITES
ARGENTINA
2
PROGRAMS
2
SITES
AUSTRALIA
24
PROGRAMS
461
SITES
AUSTRIA
21
PROGRAMS
774
SITES
BELGIUM
15
PROGRAMS
794
SITES
BERMUDA
1
PROGRAMS
5
SITES
BOTSWANA
1
PROGRAMS
5
SITES
BRAZIL
3
programs
16
51TE5
CANADA
84
PROGRAMS
4871
SITES
CHAD
3
PROGRAMS
53
SITES
CHILE
9
PROGRAMS
82
SITES
COLOMBIA
2
PROGRAMS
20
SITES
congo. republic
OF I
PROGRAMS
20
SITES
COSTA RICA
2
PROGRAMS
53
SITES
CUBA
1
PROGRAMS
14
SITES
CZECHOSLOVAKIA
5
PROGRAMS
245
SITES
CENMARK
3
PROGRAMS
63
SITES
ECUADOR
4
PROGRAMS
43
SITES
EGYPT
I
PROGRAMS
2
SITES
EL SALVADOR
4
PROGRAMS
73
SITES
ENGLAND
32
PROGRAMS
2311
SITES
FINLAND
27
PROGRAMS
906
SITES
FRANCE
10
PROGRAMS
2305
SITES
GAMBIA
1
PROGRAMS
4
SITES
FED REP OF GERMANY 35
PROGRAMS
6481
SITES
GHANA
2
PROGRAMS
18
SITES
GREECE
1
PROGRAMS
40
SITES
GUYANA
2
PROGRAMS
15
SITES
HONG KONG
3
PROGRAMS
44
SITES
HUNGARY
3
PROGRAMS
26
SITES
ICELAND
3
PPOGRAMS
77
SITES
INDIA
17
PROGRAMS
59
SITES
IRAN
1
PROGRAMS
14
SITES
IRAQ
I
PROGRAMS
5
SITES
IRELAND
7
PR06RAMS
248
SITES
ISRAEL
17
PROGRAMS
3108
SITES
ITALY
26
PROGRAMS
906
SITES
IVORY COAST
1
PROGRAMS
10
SITtS
JAMAICA
2
PROGRAMS
30
SITES
JAPAN
25
PROGRAMS
5909
SITES
JORDAN
I
PROGRAMS
0
SITES
KENYA
3
PROGRAMS
47
SITES
3-6
-------
TABLE III (CONTINUED)
KUWAIT
2
PROGRAMS
16
SITES
MALAWI
i
PROGRAMS
2000
SITES
KEX1C0
3
PROGRAMS
50
SITES
MONACO
2
PROGRAMS
205
SITES
NETHERLANDS
22
PROGRAMS
2853
SITES
NEW GUINEA
2
PROGRAMS
4
SITES
NEW ZEALANO
34
PROGRAMS
1266
SITES
NICARAGUA
I
PROGRAMS
141
SITES
NORWAY
12
programs
159
SITES
PAKISTAN
4
PROGRAMS
79
SITES
PERU
1
PROGRAMS
5
SITES
PHILIPPINES
2
PROGRAMS
28
SITES
POLAND
5
PROGRAMS
1839
SITES
PORTUGAL
18
PROGRAMS
230
SITES
PUERTO RICO
2
PROGRAMS
19
SITES
RHODESIA
1
PROGRAMS
5
SITES
ROMANIA
I
PROGRAMS
25
SITES
SCOTLAND
26
PROGRAMS
629
SITES
SENEGAL
3
PROGRAMS
103
SITES
SIERRA LEONE
1
PROGRAMS
4
SITES
SOUTH AFRICA
«~
PROGRAMS
26
SITES
SPAIN
10
PROGRAMS
89
SITES
SWAZILAND
1
PROGRAMS
20
SITES
SWEDEN
26
PROGRAMS
1071
SITES
SWITZERLAND
12
PROGRAMS
81
SITES
TAIWAN
8
PROGRAMS
105
SITES
TANZANIA
I
PROGRAMS
5
SITES
THAILAND
2
PROGRAMS
15
SITES
TOGO
1
programs
10
SITES
TURKEY
1
PROGRAMS
e
SITES
UGANDA
1
programs
3
SITES
USSR
1
programs
5
SITES
USA
180
programs
22732
SITES
URUGUAY
i
programs
1
SITES
VENEZUELA
1
PROGRAMS
1
SITES
virgin ISLANDS
1
PROGRAMS
10
SITES
MIES
3
PROGRAMS
52
SITES
west indies
1
PROGRAMS
9
SITES
JUGOSLAVIA
2
PROGRAMS
SO
SITES
Zaire
1
PROGRAMS
3
SITES
total
SITES*
66115
TABLE IV
EXISTING MERCURY, LEAD, CADMIUM, AND
chlorinated hydrocarbon pollution MONITORING PROGRAMS
MERCIJ
RY
LEAD
CADM
UM
CHLOR. HYDRO.
COUNTRY
M. iff
PROGS.
NO. OF
SITES
NO. OF
PROGS.
NO. OF
SITES
NO. OF
PROGS.
NO. OF
SITES
no. or
PROGS.
NQ. OF
SITES
Australia
2
110
4
169
3
166
1
100
Azores
1
2
1
. Z
Belgium
3
89
3
. 89
3
89
1
75
Brazil
1
7
1
7
1
1
1
7
Canada
' 11
431
11
463
9
393
e
213
Columbia
1
9
1
9
1
9
i
9
Czechoslovakia
1
11
2
12
1
11
Oenmark
1
8
1
8
1
8
Egypt
i
a
Finland
1
1
1
4
1
France
2
4
156
2
150
3
Fed. Rep. Germany
4
9
104
8
62
3
Greece
1
30
1
30
1
30
1
10
Iceland
2
8
1
3
India
1
1
Iran
1
1
1
Ireland
1
4
\
Israel
2
8
2
•
Italy
84
82;;
38
49
Japan
232
4845
5004
12
Kenya
10
Korea
2
42
42
42
12
Kuwait
8
8
Malagasy Republic
Mexico
2
Morocco
20
Netherlands
4
4
New Guinea
1
3
New Zealand
46
49
2
47
46
Nigeria
5
1
S
1
5
5
Norway
IS
3
IS
3
15
Pakistan
1
Philippines
1
6
Poland
4
176
3
IB
2
18
2
170
Portugal
6
74
106
71
6
63
Puerto Rico
14
14
14
14
Singapore
24
24 ¦
2
24
Spain
4
37
7
4
25
Swaziland
5
Sweden
S
78
Switzerland
2V
14
4
26
3
14
4
Taiwan
2
3
2
1
United Kingdom
24
774
39
1657
31
822
23
620
United States
33
3141
38
3056
33
3049
20
2365
TOTALS 44
140
5414
11,096
136
10,083
96
3855
at least 17 programs that monitor mercury, lead
cadmium, and chlorinated hydrocarbons 1n air, soil,
marine water, fresh water, plants, animals, and man.
At least four programs sample for mercury, lead, and
chlorinated hydrocarbons in air dally; two programs
sample fresh water for mercury, lead, and cadmium
monthly; one program samples air, soil, marine water,
plants, and animals for levels of mercury, lead, cad-
mium and chlorinated hyrdocarbons on a monthly basis;
and a dozen other programs sample soil, fresh water,
food, plants, and animals for levels of all four pol-
lutants on a quarterly, seasonal, or annual basis.
NETWORK DESIGN CONSIDERATIONS
In designing a global network monitoring locations
-------
TABLE V
< NO OF PROGRAMS AND SITES MONITORING MERCURY, LEAD, CAC
BY GEOGRAPHICAL AREAS. (OCEAN SAMPLING PROGRAMS ARE INCLUDED
IMIUH, AND C
UNDER COUNT
HLORINATED HYDROCARBONS
RY ADMINISTERING THE PROGR
AMSJ
No. of
Countries
No. of
Programs
No. of
Sites
No, of
Programs
No. of
Sites
No. of
Programs
No. of
31 tes
lio. Of
Programs
No. of
Programs
Mercury
Mercury
lead
Lead
Cadmium
Cadmium
Chlorinated
Hydrocarbons
Chlorinated
Hydrocarbons
Europe
17
69
1317
86
2392
58
1332
46
1016
Africa
6
3
15
3
10
2
5
3
33
Asia
13
18
770
24
5143
21
5283
14
173
North America
3
45
3586
50
3533
43
3447
29
2612
South America
3
2
16
3
18
2
16
2
16
Atlantic1
2
3
10
-
-
-
2
5
TOTALS
44
140
5414
166
11096
136
10083
96
3855
should be selected carefully to Insure that analytical
results answer specific questions concerning current
pollution levels and trends. Initially, monitoring
site selection should be based upon best estimates of
pollution levels in order to provide baseline informa-
tion that can be used to adjust sampling site locations
and sampling frequencies at a later time.
Valid conclusions about the general levels of
pollutants in the environment can be obtained by de-
signing monitoring networks that Include two approaches:
one that establishes sites which will take into account
the pollution transport mechanisms such as global wind
and water systems in relation to the major sources of
pollution, and another in which sites are selected in
areas remote from pollutant sources (Baseline Stations).
The global networks should attempt to obtain
systematic sampling of all significant air, water, land,
and biological systems. To begin with, a modest net-
work of several thousand sampling sites strategically
located on all continents and oceans, including land
sites, river sites, lake sites, estuary sites, glacier
sites, continental shelf sites, and
deep ocean sites would be employed.
Site locations should be selected to
provide preliminary baseline informa-
tion on the current state of the glo-
bal pollution burden of marine, at-
mospheric, and terrestrial systems
and keep track of the rates of change
and variability of thier pollution
levels.
Program priorities and monitor-
ing requirements (including site dis-
tribution and sampling media and
frequencies) should be continuously
re-evaluated to insure that the most
critically Important data are obtained
from the networks.
Utilizing the data base of exist-
ing pollution monitoring programs, it
is possible to design global networks
to monitor specific pollutants in
specific media with specific method-
'ology at specified frequencies.
APPENDIX I
Appendix I contains a statistical summary of the
numbers of programs and monitoring sites by media,
coverage, sampling frequency, means of data or sample
acquisition, and method of data storage for the toxic
heavy metals, mercury, lead, and cadmium and the
chlorinated hydrocarbons in 44 countries.
MERCURY
MEDIUM
PROGRAMS
SITES
AIR
11
691
SOIL
16
1110
MARINE WATER
33
1669
FRESH WATER
51
6673
DRINKING WATER
4
983
PLANTS
10
1461
ANIHALS
46
2S16
FOOD
11
40
TABI.E VI
TYPICAL EXAMPLE OF A COUNTRY'S KRCURr, LEAO. CAOHIUH, MID CHLORINATED HYDROCARBON MONITORING PROGRAMS
CANADA
Proarin
Hwcurv
I
•id
Cad
m1um
Chlorinated Hydro
carbons
NURMr
MMiua
sites
Frequency
Medium
Sites
Frequency
Medium
Sites
frequency
Medium
>1tes
Frequency
1
Fresh water
16
monthly
Fresh water
16
monthly
Fresh water
16
monthly
2
Fresh water
250
quarterly
Fresh water
2S0
quarterly
Fresh water
250
quarterly
Fresh water
100
annually
3.
Soil, fresh
water,
plants* and
animals
50
season-
ally
soil, fresh
water,
plants, and
animals
50
season-
ally
«.
Air
25
monthly
Mr
2S
monthly
S.
soil, food,
plants, man,
animals
%
annually
6.
animals
5
1.
Air
13
dally
8.
Air
57
dally
9.
Air
3
dally
10.
Food
1
annually
Food
1
annually
Food
1
annually
food
1
annually
11.
Fresh water
20
nonthly
Fresh water
20
monthly
Fresh water
20
monthly
12,
Fresh water
S3
annually
Fresh water
53
annually
Fresh water
S3
annually
13.
Atr
SO
dally
14.
Food
Food
Food
food
15.
Animals
10
annually
Animals
7
annual1y
It.
Fresh water
28
quarterly
Fresh water
28
quarterly
Fresh water
28
quarterly
17.
Air, soil
marine
water,
plants,
anlMls
monthly
Air, soil,
marine
water,
plants
animals
Monthly
Air, soil,
marine
water
plants
animals
monthly
4
-------
MERCURY
LEAD
COVERAGE
LOCAL
REGIONAL
national
INTERNATIONAL
GLOBAL
SAMPLING FREQUENCY
CONTINUOUSLY
DAILY
WEEKLY
MONTHLY
QUARTERLY
seasonally
ANNUALLY
SPORADICALLY
OTHER
DURING RAJN RUNOFF
WEATHER DEPENDENT
NOVEMBER AND DECEMBER 1973
BI-WEEKLY
EVERY 2 YEARS
EVERY 4 YEARS
PERIODICALLY .
BIMONTHLY
MEANS OF DATA OR SAMPLE
QROUlIb SITES ACQUISIT
SHIPS
BUOYS •
AIRCRAFT
OTHER
FOOD
DIVERS
VARIOUS
SMALL BOAT
5NOWMO0ILE
DATA STORAGE
PUBLISHED
NOT PUBLISHED
PUNCHCAROS
MAGNETtC TAPE
UNPUBLISHED FILE
UNPUBLISHED REPORTS
UNPUBLISHED (NODC)
TO BE PUBLISHED
DISK
UNPUBLISHEO < NOQC1
LEAD
MEDIUM
AIR
SOIL
MARINE WATER
FRESH WATER
DRINKING WATER
PLANTS
ANIMALS
FOOD
MAN
SEWAGE
COVERAGE
LOCAL
REGIONAL
NATIONAL
INTERNATIONAL
GLOBAL
SAMPLING FREQUENCY
CONTINUOUSLY
DAILY
WEEKLY
MONTHLY
QUARTERLY
seasonally
ANNUALLY
SPORAOICALLY
OTHER
EVERY 3 WEEKS
DURING RAIN RUNOFF
WEATHER DEPENDENT
NOVEMBER AND DECEMBER 1*73
BI-WEEKLY
EVERY 2 YEARS
EVERY * YEARS
PERIODICALLY
BIMONTHLY
PROGRAMS
SITES
51
178*
*0
2525
ST
6497
7
196
3
9
PROGRAMS
SITES
26
963
I
0
2
46
19
2406
15
3666
9
630
16
1441
16
2244
4
330
1
4
1
35
1
5
1
1
2
51
2
0
30
3746
1
35
PROGRAMS
SlTtS
93
7532
*4
2465
7
925
1
6
7
710
2
1
2
t>
3
125
1
11
1
11
PROGRAMS
SITES
32
35 80
29
*94
U
1060
16
3625
2
B
I
14
1
14
3
6
2
511
I
6
PROGRAMS
SITES
*2
762
12
4925
23
1150
52
10290
5
791
2*
5036
32
1671
9
5
1
6
1
17
PROGRAMS
SITES
48 .
1395
35
2246
46
9896
T
163
2
9
PR0GRAM5
SITES
17
253
3
3
8
142
27
2654
1$
3694
6
539
16
5402
16
2263
3
225
1
3
1
4
1
35
1
5
1
1
2
51
1
0
16
2848
1
35
MEANS OF DATA OR SAMPLE ACQUISIT
GROUND SITES
SHIPS
BUOYS
AIRCRAFT
OTHER
FOOO PURCHASED FROM PRODUCER
DIVERS
VARIOUS
SMALL BOAT
SNOWMOBILE
DATA STORAGE
PUBLISHED
NOT PUBLISHED
PUNCHCAROS
MAGNETIC TAPE
OTHER
UNPUBLISHED REPORTS
UNPUBLISHED (NODC)
TO BE PUBLISHED
DISK
CADMIUM
MEDIUM
AIR
SOIL
MARINE WATER
FRESH WATER
DRINKING WATER
PLANTS
ANIMALS
FOOD
COVERAGE
LOCAL
REGIONAL
NATIONAL
INTERNATIONAL
GLOBAL
SAMPLING FREOUENCY
CONTINUOUSLY
DAILY
WEEKLY
MONTHLY
QUARTERLY
SEASONALLY
ANNUALLY
SPORADICALLY
OTHER
EVERY 3 WEEKS
DURING RAIN RUNOFF
WEATHER DEPENDENT
NOVEMBER AND DECEMBER 1973
BI-WEEKLY
EVERY 2 YEARS
EVERY 4 YEARS
PERIODICALLY
BIMONTHLY
3 PER WEEK
MEANS OF DATA OR SAMPLE ACQUISIT
GROUND SITES
SHIPS
BUOYS
AIRCRAFT
OTHER
FOOD
DIVERS
DATA STORAGE
PUBLISHED
NOT PUBLISHED
PUNCHCAROS
MAGNETIC TAPE
OTHER
UNPUBLISHED REPORTS
UNPUBLISHED INODC)
TO BE PUBLISHED
DISK
PROGRAMS
SITES
100
11129
34
1507
2
220
1
6
2
608
2
1
2
8
1
20
1
11
1
11
PROGRAMS
SITES
36
2998
31
4644
13
539
22
3222
4
30
i
16
1
1*
3
6
2
511
PROGRAMS
SITES
16
261
13
4959
24
1311
31
7403
4
786
19
4791
21
1*26
9
S
PROGRAMS
SITES
25
1261
33
2060
33
7202
6
146
3
9
PROGRAMS
SITES
7
150
I
2
4
12*
18
2*3*
13
3651
7
57#
16
5395
15
1853
3
145
1
3
1
4
I
35
I
5
I
I
2
51
I
0
7
560
I
35
1
1
PROGRAMS
SITES
62
81B2
32
1607
1
200
1
6
1
650
2
i
2
a
PROGRAMS
SITES
26
2350
27
4616
U
477
20
252*
1
3
2
14
L
1*
3
6
1
500
3-6
-------
halogenated organic compounds, especially dot and
MEDIUM
air
SOIL
MARINE WATER
FRE5H WATER
DRINKING WATER
PLANTS
ANIMALS
FOOD
MAN
COVERAGE
LOCAL
REGIONAL
NATIONAL
INTERNATIONAL
GLOBAL
sampling frequency
continuously
DAILY
WEEKLY
, MONTHLY
QUARTERLY
SEASONALLY
ANNUALLY
SPORADICALLY
OTHER
DURING MAIN RUNOFF
WEATHER DEPENDENT
NOVEMBER AND DECEMBER 1973
EVERY 2 YEARS
EVERY 4 YEARS
PERIODICALLY
BIMONTHLY
means OF DATA OR SAMPLE ACOUISIT
GROUND SITES
SHIPS
BUOYS
AIRCRAFT
OTHER
FOOD
DIVERS
Various
DATA STORAGE
PUBLISHED
NOT PUBLISHED
PUNCHCARD5
MAGNETIC TAPE
OTHER
UNPUBLISHED FILE
UNPUBLISHED REPORTS
UNPUBLISHED (NODCJ
TO 8E PUBLISHED
DISK
PROGRAMS
5
23
23
29
3
36
56
18
e
PROGRAMS
24
30
62
5
1
PROGRAMS
22
4
3
11
9
10
21
16
2
1
1
1
2
1
26
1
PROGRAMS
72
30
3
I
5
1
1
2
PROGRAMS
31
29
5
13'
2
pce.s
sites
163
2094
1535
2107
23$
1730
2161
66
89
SI TEi
1209
1929
3311
194
i
SITES
767
Si
46
1229
1564
621
10fc4
844
210
4
35
5
137
0
1675
3b
SITES
4267
1251
225
2
36
1
5
105
SITES
1546
738
545
1126
46
3
2
14
5
500
REFERENCES
Citron, Robert. A Plan for the Implementation of the
United Nations Global Environmental Monitoring
System (GEMS), 1975-1980, Smithsonian Institution,
June 1975.
Whitman, John. The Smithsonian Institution Survey of
National and International Pollution Monitoring
Programs 1973-1974. Smithsonian Institution,
June 1975.
Citron, R. International Environmental Monitoring
Programs - A Directory - prepared for the United
Nations Environment Program. Smithsonian
Institution, January 1974
Citron, Robert, and Whitman.John.Directory of National
and International Pollution Monitoring Programs" "
3 vols. Prepared for the United Nations Environment
Program, February 1974. Cambridge, Massachusetts:
Smithsonian Institution, 1974.
The Global Environmental Monitoring System (GEMS).
United Nations Environment Program, UNEP/GC/24
Corr. 1, 18 March 1974.
Whitman, John R. The United Nations Environment
Program Data Base on Global Environmental Monitor-
ing System (GEMS). Smithsonian Institution,
May 19/4.
WMO. Guidelines for Baseline and Regional Stations
with Expanded Programmes. WMO, Geneva, 1974.
Munn, R. E. Global Environmental Monitoring System
(GEMS), SCOPE III. United Nations Environment
Program, 1974.
Proposals Concerning the Nature of the Global Envir-
onmental Monitoring System. United Nations
Environment Program, UNEP/IG. 1/3, 26 December 1973.
Report of the Inter-Agency Working Group on Monitoring
ort Development of a Global Environmental Moni-
toring System (GEMS). United Nations Environment
Program, UNEP/IG.1 ft, 15 November 1973.
SCOR-ACMRR-UNESCO-IBP/PM. Monitoring Life in the Ocean
SCOR, London, England, 1973. 71 pp. ———
Goroshko, B. B. The Question of the Selection of the
Number of Sampling Stations and the Freguency of
Observations of Air Pollution"! In Air Pollution
and Atmospheric Diffusion] (M. E. Berlyand, Ed.)
English transl. by Israel Program for Sc. Transl.,
John Wiley and Sons, February 8, 1974, pp. 147-159.
Preliminary Report on GIPME and IGOSS Marine Pollution
Monitoring Programs. IQC-WMQ/ITECH-1/13, Inter-
national Oceanographic Commission, March 28, 1973,
22 pp.
Plan for Earthwatch and Implementation of Monitorin
United Nations Environment Program, 2 October 1973.
Jenkins, D. W. Biological Monitoring of the Global
Chemical Environment. Smithsonian Institution,
Washington, D. C., 1971, 54 pp.
Lundholm, B. "SCOPE I" A Plan to Establish a Global
Environmental Monitoring System. Special Committee
on Problems of the Environment, International
Council of Scientific Unions, 1971.
SMIC. Inadvertent Climate Modification. MIT Press,
Cambridge, Massachusetts, USA, 1971, 308 pp.
Citron, R. The Establishment of an International
Environmental Monitoring Program—A Plan for Action.
Prepared for the United Nations Conference on the
Human Environment, Stockholm, Sweden. Smithsonian
Institution, March 1971.
Citron, R. National and International Environmental
Monitoring Programs, Cambridge, Massachusetts",
Smithsonian Institution, 1970.
SCEP: Man's Impact on the Global Environment. MIT
Press, Cambridge, Massachusetts, USA, February 1,
1970.
Citron, R. MABNET: The Establishment of a Global
Network of Eco-stations for the Man and the 6'io-
sphere Program. Cambridge, Massachusetts, USA,
March 1 , 1970, 31 pp.
Citron, R. and Staff. National and International
Environmental Monitoring Activities, Smithsonian
Institution, October 19/0.
ACKNOWLEDGEMENTS
I would like to acknowledge the fine work done by
Mr. John Whitman, who has spent the better part of the
past two years, helping to plan, develop, and manage
the Pollution Monitoring Program Data Base.
I would also like to thank Mr. Wayne Mills for
his assistance in helping to prepare some of the Infor-
mation and data contained 1n the tables and the appen-
dices.
INFORMATION ON THE
POLLUTION MONITORING PROGRAM DATA BASE
Information on the Pollution Monitoring Data Base
can be obtained by writing to the Program Manager,
United Nations Environment Program Pollution Monitor- .
ing Program Data Base, Center for Short-Lived Pheno-
mena, 60 Garden Street, Cambridge, Massachusetts 02138.
3-6
-------
POLYCHLORINATED BIPHENYLS OFF
SOUTHERN CALIFORNIA
David R. Young
Southern California Coastal Water Research Project
1500 East Imperial Highway
El Segundo, California 90245
Deirdre J. McDermott
Southern California Coastal water Research Project
1500 East Imperial Highway
El Segundo, California 90245
Theadore C. Heesen
Southern California Coastal Water Research Project
1500 East Imperial Highway
El Segundo, California 90245
Summary
Polychlorinated biphenyls are ubiquitous con-
taminants of the marine ecosystem off southern
California. The greatest known source is the
submarine discharge of municipal wastewaters;
aerial fallout also appears to be an important
input route. In contrast, surface runoff is
only of secondary importance, and direct in-
dustrial discharge appears to be completely
insignificant.
Maximum concentrations of 1254 PCB found
in bottom sediments off Palos Verdes Peninsu-
la were 10 ppm, with values falling 100 fold
to baseline values over a depth of 20 cm.
Muscle tissue from benthic crabs (Cancer
anthonvi) and flatfish (Microstomus pacificus)
collected from discharge regions also exhib-
ited 100 fold increases in 1254 PCB levels
over control values. It appears that the
numerous harbors in the Bight also have rela-
tively high levels of PCB contamination.
Intertidal bay mussels (Mvtilus edulis) con-
tained up to 20 times more 1254 PCB than cor-
responding specimens collected nearby along
the open coast.
Finally, an offshore biomonitoring sys-
tem has been developed, utilizing the inter-
tidal coastal mussel (M. californianus), which
appears to offer good potential for indicating
zones of greatest biological availability of
synthetic pollutants released from submarine
outfalls.
Introduction
Chlorinated hydrocarbons are among the most
important pollutants yet identified in the
Southern California Bight.1"° The coastal
plain adjacent to this marine ecosystem
(Figure 1) is inhabited by approximately 11
million persons (approximately 5 percent of
the U.S. population), and large quantities of
waste materials are introduced both directly
and indirectly into the waters of the Bight.
Potentially significant sources and transport
routes include submarine discharges of muni-
cipal wastewaters, direct industrial dis-
charges, antifouling paints and other vessel
related materials, surface runoff, aerial
7
fallout, and ocean current advection.
Here we summarize our findings from
studies into the industrially important poly-
chlorinated biphenyls (PCB) conducted over the
last several years while investigating the
inputs and distributions of chlorinated hydro-
carbons off southern California.
N
St"
N
33°
N
32'
N
figure 1. The Southern California Bight.
Major outfall systems are (1) Oxnard City,
(2) Hyperion, Los Angeles City, (3) JWPCP,
Los Angeles County, (4) 0CSD, Orange County,
and (5) Point Loma, San- Diego City. Major
harbors are (6) San Pedro, (7) Newport, and
(8) San Diego.
Sampling
Five major treatment plants* located along th£
southern California coast contribute more than
9$ percent of the 4 x 10® liters of municipal
wastewater discharged daily into the Bight.
In order to obtain an estimate of the impor-
tance of this input route, we analyzed one-
week composites (usually collected twice a
year) of the final effluents obtained from
each of these plants since 1972.
in southern California, most industrial
effluents discharged directly into marine
waters (excluding power plant cooling waters)
flow into San Pedro Harbor or San Diego Bay.
* Joint Water Pollution Control Plant (JWPCP),
Los Angeles County; Hyperion Treatment
Plant, City of Los Angeles; Orange County
Sanitation District Treatment Plant (OCSD),
Orange County? Point Loma Treatment Plant,
City of San Diego? Oxnard Treatment Plant,
City of Oxnard.
CAUFQ«/V7A
Pt
Concep
Oxnard
\
Major mumcipil
harbors
i»W
120W
m'u
)>B°W
in'w
l
4-1
-------
Therefore, to estimate the level of PCB input
from this source during 1973 and 1974, half-
day effluent composites obtained for each of
the major types of industries discharging into
both areas were analyzed. In addition, during
1973 we conducted an intensive survey to deter-
mine the composition and quantities of major
brands of antifouling paints used on recreat-
ional, commercial, and naval vessles in the 14
harbors and marinas in southern California.
We estimated the contribution of surface
runoff by sampling major channels along the
southern California coast during both storm
and dry weather conditions between 1971 and
1973. To measure aerial inputs, replicate 1
week collections of dry aerial fallout were
analyzed from 13 coastal and 5 island stations
throughout the Bight during two 13 week peri-
ods in 1973 and 1974. Finally, to estimate
the magnitude of ocean current advection in-
puts during 1973, we analyzed results of PCB
measurements for replicate samples of surface
seawater which we had collected in the Santa
Barbara Channel at the edge of .the Bight.
During 1971, we collected ocean bottom
sediments near the JWPCP outfalls off Palos
Verdes Peninsula using a deliberate box corer.
Utilizing traps and benthic trawls, samples of
the benthic crab (Cancer anthonyi) and a flat-
fish, the Dover sole (Microstomus pacificus)
were collected from around the Hyperion, JWPCP
and OCSD outfalls during 1971 and 1972; the
Dover sole was subsequently resampled in these
regions in 1974 and 1975. In addition, during
1974 we sampled the intertidal bay mussel
(Mytilus edulis) from throughout San Pedro,
Newport and San Diego Harbors. Later that
year, to estimate uptake rates and steady
state concentrations, we transferred 2,000
specimen of the intertidal mussel (M. califor-
nianua) from an uncontaminated control station
to net bags suspended from a taut line buoy
system situated above highly contaminated
sediments near the JWPCP outfalls.
Methodology
Most of the various categories of samples were
processed and analyzed utilizing well estab-
lished laboratory techniques in conjunction
with electron capture gas chromatography. The
aerial fallout samples were analyzed using a
procedure developed by Dr. Vance McClure
(National Marine Fisheries Service, La Jolla,
California), and the seawater samples were
analyzed in collaboration with Brock de Lappe
and Dr. Robert Risebrough, utilizing a tech-
nique they developed at the University of
California at Berkeley.
Results and Discussion
Sources and Transport Routes
Table 1 summarizes the estimated annual
mass emission rates of 1242 PCB and 1254 PCB
(42 and 54 percent ehlorination, respectively)
to the Bight via six modes of input, and Table
2 presents the breakdown of 1974 inputs from
manor municipal submarine discharges. Approxi-
mately equal concentrations of 1242 and 1254
PCB were measured in surface runoff; thus,
although reliable measurements of 1242 PCB in
dry aerial fallout have not yet been made, we
estimated an upper limit for this input assum-
ing a 1 to 1 ratio of 1242 and 1254 PCB in
fallout.
Table 1. Estimated annual mass emission
rates (kg/yr) of polychlorinated biphenyls
to the Southern California Bight.
Route Year
1242 PCB
1254 PCB
Municipal
Wastewater 1974
4300
1100
Direct
Industrial 1973-74
-
60
Antifouling
Paints 1973
< 1
<1
Surface
Runoff 1973
4 500
300
Aerial
Fallout 1973-74
£ 1500*
1500
Total (Land)
4500-6000*
3000
Ocean
Currents 1973
-
4000
* Assuming a 1 to 1 ratio between
1254 PCB in aerial fallout
1242 and
Table 2. Estimated 1974 mass emission
rates (kg/yr) of polychlorinated biphenyls
to the Bight via major municipal waste-
waters.
Flow
Treatment Plant (mgd)
1242 PCB
1254 PCB
JWPCP 350
910
360
Hyperion (5-mi) 330
50
110
Hyperion (7-mi) 5
470
320
OCSD 150
1890
210
Pt. Loma 95
950
v 110
Oxnard 10
1
3
Total 940
4270
1110
These results indicate that municipal
wastewater is the dominant mode of input for
PCB to the coastal waters, followed by dry
aerial fallout. Surface runoff contributions
are an order of magnitude lower, while esti-
mated coastal inputs from direct industrial
discharges and antifouling paint application
are relatively insignificant.
In view of the fact that the quantity of
1254 PCB (1,100 kg) introduced during 1974
was more than one-fourth the estimated quan-
tity (4,000 kg) carried annually by surface
layer (0-50 m) currents flowing through the
entire Bight, one would expect that local con-
tamination of nearshore sediments and orga-
nisms by this synthetic material would be
clearly evident.
Marine Sediments
Figure 2 presents vertical profiles of
1254 PCB concentrations in bottom sediment
2
4-1
-------
II
r
.. Bib
0
Depth (cm)
Figure 2. Vertical profiles of 1254 PCB con-
centrations in sediments collected northwest
of the JWPCP outfalls, box cores, July 1971.
layers from box cores collected at three sites
within the monitoring zone and down current of
the JWPCP outfalls off palos Verdes Peninsula.*
Concentrations of 1254 PCB in nearsurface
sediments (10 ppsn, mg/dxy kg) collected at
Station B21, located 1.5 Xm from the nearest
outfall diffuser, exceeded the lowest levels
measured at the bottom of the three cores (0.1
ppm) by a factor of 100. Surface sediments
from the Santa Barbara Basin (180 km to the
northwest) also contain about 0.1 ppm 1254
PCB.9
We previously calculated that, in 1972,
the upper 30 eta of bottom sediments in a 50 gq
km area of this discharge zone contained
approximately 200 metric tons of DDT and its
residues^; we believe this estimate to be
accurate within 25 percent of the actual value.
Although we have much fewer data on the dis-
tribution of PCB in these sediments, extra-
polation of our present information using
observed PCB to DDT ratios suggests that the
corresponding load of 1254 PCB in these sedi-
ments is on the order of 6 metric tons.
Harine Animals
Benthic Organieans. samples of muscle
tissue from crabs (C. anthonyi) and Dover sole
(M. pacificus) collected in these discharge
regions also exhibited significant levels of
* Distance downcurrent? Station Bl8, 7 3cm;
Station B21, 1.5 km? and Station B20,
between the legs of the wye of the outfall.
PCB contamination above baseline concentra-
tions (Figures 3 and 4, respectively). Levels
as.
33°
SB'N
°J>,
100
i i t
KM
Depths in meters
l>0a3O'W
yob.
Figure 3. 1254 PCB concentrations (mg/wet kg)
in muscle tissue of the crab, Cancer anthonyi,
collected off southern California, f9?l-72.
Port *
-------
Bay Mussels. Our surveys of 5 cm bay
mussels (M. edulis) from three of the largest
southern California harbors revealed contamin-
ation levels up to 20 times those found in
specimens of the same species collected from
nearby coastal sites.'1'0 Highest levels were
found near regions of greatest vessel activity
as illustrated by the results for San Diego
Bay presented in Figure 5. Comparison of sea-
water levels (generally on the order of one
32"
50'N
Harbor Island.
ia0fc 0.18 J- Commercial Docks
orifctj
0.04
0>
t±. calLPornianus
Km
Figure 5. 12 54 PCB concentrations (mg/wet kg)
in whole soft tissues of the bay mussel,
Mytilus edulis, collected in San Diego Bav,
January 1974.
part per trillion, ng/1) and mussel concen-
trations of 1254 PCB found in San Diego and
the other harbors suggests that the whole soft
tissues of M. edulis accumulate PCB on the
order of 100,000 times above seawater values.
Although antifouling paints presently
applied to vessel bottoms in southern Cali-
fornia generally contain PCB concentrations of
less than 1 ppm (mg/dry Kg), a few samples of
old paint chips averaged about 10 percent PCB
(100,000 ppm).8 Thus it is possible that
thousands of kilograms of this synthetic ma-
terial could have been released annually to
the harbor and coastal marine ecosystems,
before the widespread use of non-recoverable
PCB was discontinued in the United States
during the early 1970's.
Offshore Biomonitorinq System
In June 1974, uncontaminated specimens of
open coast, intertidal mussels (M. californi-
anus), 5 cm in length, were transferred to net
bags fastened at five different levels between
the sea surface and bottom (35 m) on a taut
line buoy system installed over the highly
contaminated sediments near the JWPCP outfalls.
The rate and extent of PCB uptake was then
monitored for 3 months in specimens collected
at 1 to 2 week intervals. The mussels appeared
to survive well under these conditions; less
than 10 percent mortality at any level was
observed over the study period.
On the average, mussels living at Royal
Palm Beach inshore of the buoy contained 20
times as much PCB as the control specimens at
the time of their transfer to the buoy north-
west of the outfalls. Thus, exposure of these
mussels to PCB was greatly increased upon
their transfer to the discharge region.
Resultant uptake of 1254 PCB with time in
male specimens is presented in Figure 6;
(BOTTOM)
IS M
""'SURFACE
Week
Figure 6. 1254 PCB concentrations in whole
soft tissue composites of male intertidal
mussels, Mytilus californianus, translocated
from Point Sal to the taut line buoy system
off Palos Verdes, June through September 1974.
similar results were observed for female
specimens, and application of the Wilcoxon
signed-rank test revealed no significant
effect of sex on PCB uptake. These data show
that there is a direct relationship between
uptake of PCB and proximity of the bioindi-
cator to the contaminated bottom sediments and
to the wastewater plume, which is trapped
beneath the thermocline. The bottom specimens
became approximately 10 times as contaminated
as did the surface specimens. Projected
"steady state" concentrations for the two
levels of suspension appear to be on the order
of 0.4 and 0.04 ppm (mg/wet kg). To date, the
highest 1254 PCB concentrations measured in
the water above the JWPCP outfalls is 4 parts
4
4-1
-------
per trillion. This suggests a concentration
factor on the order of 100,000 for 1254 PCB
in the whole soft tissues of M. californianus,
in good agreement with the estimate for M.
edulis in the harbors.
Conclusions
Municipal wastewaters and aerial fallout dom-
inate the known inputs of PCB to the Southern
California Bight. More than 5 metric tons of
1242 and 1254 PCB were discharged during 1974
from the five major southern California treat-
ment plants. Such inputs have resulted in
localized zones of relatively high bottom
sediment contamination around submarine out-
fall systems.
Measured annual emissions of 1254 PCB
from the three major outfall systems agreed
within a factor of 2; similar agreements were
observed in the 1254 PCB levels in crabs
(C. anthonyi) and flatfish (M. pacificus) col-
lected from the three discharge regions. Thus
it is apparent that PCB compounds are wide-
spread contaminants of both coastal wastewater
inputs and""benthic organisms in southern
California.
Highest concentrations in the bay mussel
(M. edulis) were measured near locations of
greatest vessel activity, such as mooring and
bottom repainting facilities. The contamina-
tion of harbor organisms may be the result of
past usage of antifouling paints containing
high levels of PCB, although inputs from other
vessel-related materials (such as hydraulic
fluids) could also be important.
Although the bioindicator (M. californi-
anus) used in the offshore biomonitoring
system is an intertidal organism, less than 10
percent mortality was observed, even at a
depth of 35 m, over the 3 month study period.
Thus, this mussel's hardiness, its ubiquitous
distribution along many coast lines around the
world, its very high ability for concentrating
chlorinated hydrocarbons above seawater values
(approximately 100,000 fold for 1254 PCB), and
its apparent ability to rapidly respond to
changes in environmental levels of such con-
taminants makes it a very useful bioindicator
both in natural intertidal communities and on
offshore monitoring substrates.
Acknowledgements
Dr. Robert Risebrough and Brock de Lappe
(Bodega Marine Laboratory, University of Cal-
ifornia at Berkeley) provided seawater PCB con-
centrations discussed in this paper, and assis-
ted us in analyzing harbor water and municipal
effluents. Dr. Vance McClure (National Marine
Fisheries Service, La Jolla, California) devel-
oped the techniques used to collect and analyze
dry aerial fallout. Bottom sediments and
organisms were collected with the assistance
of Dr. Ronald Kolpack (University of Southern
California) and Douglas Hotchkiss and the crew
of the Sea-S-Dee (County Sanitation Districts
of Los Angeles County).
Municipal wastewaters were collected with
the assistance of personnel from corresponding
treatment plants. Members of this Project who
contributed to this research include Willard
Bascom (Director), Elliott Birkihiser, Rick
Gammon, Joseph Johnson, Jack Mardesich, Dr.
Alan Mearns, Michael Moore, Paul Smokier,
Harold stubbs, Xleana Szpila, and Michael
Westbrook.
This program was supported in part by the
U.S. Environmental Protection Agency (Grant
Number R801153), and in part by a contract
with the University of California at San Deigo
in connection with the Marine Research commit-
tee, Department of Fish and Game (State of
California Agreement M-ll). Equipment used to
sample chlorinated hydrocarbons in seawater
was developed with the support of the National
Science Foundation International Decade of
Ocean Exploration, (Grant Number ID072-06412
A02). Contribution number 50 of the Southern
California Coastal Water Research Project.
References
1. Duke, T.W. and A.J. Wilson, Jr. 1971.
Chlorinated hydrocarbons in livers of
fishes from the northeastern Pacific
Ocean. Pest. Mon. Jnl. 5:228-32.
2. Risebrough, R.W., F.C. Sibley, and M.N.
Kirven. 1971. Reproductive failure of the
brown pelican on Anacapa Island in 1969.
Aroer. Birds 25:8-9.
3. Southern California Coastal Water Research
Project. 1973. The ecology of the Southern
California Bight: Implications for water
quality management. Rept. TR104, So. Calif.
Coastal Water Res« Proj., El Segundo, Ca.
4. De Long, R.L., W.G. Gilmartin, and J.G.
Simpson. 1973. Premature births in Cali-
fornia sea lion: Association with high
organochlorine pollutant residue levels.
Science 181:1168-70.
5. MacGregor, J.S. 1974. Changes in the
amount and proportions of DDT and its
metabolites, DDE and DDD, in the marine
environment off southern California, 1949-
1972. Fish.Bull. 72:275-93.
6. McDermott, D.J., T.C. Heesen, and D.R.
Young. 1974. DDT in bottom sediments around
five southern California outfall systems.
Rept. TM217, So. Calif, coastal Water
Res. Proj., El Segundo, California.
7. Young, D.R., C-S. Young, and G.E. Hlavka,
1973. Sources of trace metals from highly
urbanized southern California to the
adjacent marine ecosystem. In Cycling and
control of metals. Nat. Envir. Res. Ctr.,
Cincinnatti.
8. Young, D.R., T.C. Heesen, D.J. McDermott
and P.E. Smokier. 1974. Marine inputs of
polychlorinated biphenyls and copper from
vessel antifouling paints. Rept. TM212,
So, Calif. Coastal Water Res. Proj., El
Segundo, California.
9. Horn, W., R.W. Risebrough, A. Soutar and
D.R. Young. 1974. Deposition of DDE and
polychlorinated biphenyls in dated sedi-
ments of the Santa Barbara Basin. Science
184:1197-99.
10. Young, D.R. and T.C. Heesen. 1974. Inputs
and distributions of chlorinated hydro-
carbons in three southern California
harbors. Rept. TM214, So. Calif. Coastal
Water Res. Proj., El Segundo, California.
-------
WATER CONTAMINATION DETECTION
I. Lysyj
Rockwell International/Rocketdyne Division
Canoga Park, California
Abstract
A laboratory and field study was conducted to de-
velop a means for detecting gross discharges and spills
of waste materials into open waters and to monitor
gross changes in the quality of the receiving water. A
pyrographic method was used for direct analysis of
water samples. A key element in the conceptual design
of such an alarm system is the ability to differentiate
between the natural chemical background of water and
the pollutant composition introduced into the water as
a result of waste discharge. To evaluate the possibil-
ity of detection of industrial waste discharge into a
waterway by pyrographic means, a combined EPA-Rocketdyne
team conducted a field study in the Occonee River in
Georgia. It was found, in all cases, that pyrographic
patterns of discharged waste were different, and could
be differentiated from pyrographic patterns of natural
unpolluted waters, providing a means for determining
general or "gross" water quality of a receiving stream.
Introduction
During the past few years, significant progress was
made in adopting modern methods of analytical chemistry
to environmental quality control. Examination of recent
publications in Analytical Chemistry and Environmental
Science and Technology reveals thata substantial frac-
tion of published papers deal with environmental moni-
toring problems, signifying the fact that intense
research efforts are currently directed by both the
government and industry to the advancement of the art
in this field. As a result of this activity, the shift
from wet chemical procedures to instrumental techniques
of analysis is continuing.
The use of modern methods of analysis, such as
spark source spectrometry for analysis of inorganic
constituents and gas chromotography-mass spectrometry
methods for detection of traces of organic matter, has
revealed the highly complex and multifaceted nature of
the chemical composition in environmental matrixes. As
the methods of analysis become more sensitive, progres-
sively larger numbers of chemical species are found in
water and air. Many of those are known to be toxic,
or otherwise detrimental to the environment. Conse-
quently, pollution control personnel are faced with
difficult decisions about monitoring chemicals present
in the environment (e.g., which chemicals to consider,
how to monitor them, and at what levels to monitor).
While the desirability of detecting and measuring each
and every pollutant present in the environment cannot
be disputed, economical and practical considerations
pose severe limitations on reaching such a goal. At
the present time, large capital requirements and the
need for highly trained professional personnel limit
the use of advanced instrumental analytical techniques
primarily to the government laboratories. A typical
industrial, municipal, or commercial laboratory con-
ducting water analysis relies heavily on traditional
wet chemical techniques as specified in ASTM and EPA
publications.
Simplified and inexpensive intrumental approaches
to the characterization of water quality are needed to
assist pollution control personnel in dealing with the
increasingly complex problem of water quality control.
Specifically, a well-defined and easily measurable
"indicator of general water quality" would be of great
utility in the field. Such an indicator should be cap-
able of disclosing the degree of deterioration of water
quality and to permit responsible personnel to imple-
ment corrective actions, including more detailed exam-
ination of the situation by both traditional and new
instrumental techniques, leading to actions required
to correct the situation.
The general or "gross" indicator of water quality
could be chosen to reflect the principal component(s)
of industrial and municipal waste. If such a principal
component is found in significant concentration, we
might suspect the presence of other polluting materials.
In bulk tonnage, organic waste probably constitutes the
largest part of all pollutants discharged into water-
ways. It would appear then, that the measurement of
the organic content of water (such as total organic
carbon) could constitute such an indicator. The stud-
ies conducted in 1968-1972 in the southeastern part of
the United States1 disclosed, however, that natural
unpolluted waters, with thriving ecological communities
and normal levels of dissolved oxygen, can contain sig-
nificant concentrations (e.g., 15-30 ppm) of naturally
derived organic matter which cannot be considered as a
pollutant. The measurement of "gross" water quality
must then be capable of differentiating between the
natural organics and organics derived from waste sources.
The method of pyrographic analysis, as applied to
analysis of water samples, has such a capability. The
detailed description of this technique has been reported
in a series of publications^'3' . In performing pyro-
graphic analysis, a water sample containing organic
matter is subjected to elevated temperatures, and the
organics are pyrolyzed. The pyrolysis gases (princip-
ally methane, ethylene, ethane, propylene, and propane)
are separated by gas chromatography and detected as a
series of peaks by a flame ionization detector. The
pattern formed by the peak ratios is characteristic of
the material analyzed; hence, the method provides both
quantitative and limited qualitative information on the
composition of matter.
Pyrographic analysis of natural, unpolluted waters
and waste containing discharges has disclosed that pat-
terns produced by waste-containing waters are markedly
different from those produced by natural water, both in
qualitative and quantitative character. This finding
provides a means for differentiating between natural
and waste derived organics found in waterways.
In the early stages of development of pyrographic
methodology, information produced by a pyrograph was
handled manually by observing and relating peak inten-
sities and peak ratios to the phenomenon studied. Later
this technique was fully automated and computerized and
became capable both of differentiating between natural
and waste organics, and providing information on quanti-
ties and the source of waste material present in the
sample1. It is believed, however, that to be most use-
ful to the field worker, instrument techniques must be
conceptually and instrumentally simple, and be available
at low cost. For this reason, some of the earlier work
on pyrography was re-examined to determine its utility
in general water quality measurements for everyday use
in field water quality control laboratories.
1
4-2
-------
Experimental
Table 2. Waste Treatment Effectiveness
An attempt was made in 1968-1969 to relate pyro-
graphic data to the composition and character of natural
(unpolluted) and polluted waters. The objective of this
study was to determine the feasibility of using pyro-
graphic data as general indicator of water quality. A
10-mile reach of the Middle Occonee River in Northern
Georgia was used in this study. The investigation con-
sisted of two case studies, conducted at the same time.
The first case study involved pyrographic characteriza-
tion of organic waste discharged from poultry process-
ing plant into one anaerobic and three aerobic ponds.
The second case study involved pyrographic characteri-
sation of untreated poultry waste discharged into the
river. In both cases, the deteriorating quality of
the water was found to be relatable to observed pyro-
graphic patterns.
Water samples were collected at various points at
each site, and preserved in an ice-chest in dry ice
prior to analysis^ Fifty microliter aliquots of water
samples were injected into the pyrograph and analyzed.
The results, in the form of a series of peaks, were
displayed on a strip chart recorder. All observed
peaks were integrated and absolute values were reported
in areas. 'To obtain specific pyrographic patterns, all
peaks in each pyrogram were related to the second
(ethylene) peak by assigning a value of 100 to it.
Anaerobic and Aerobic Waste Decomposition
The first case studied involved a poultry process-
ing plant that was discharging up to 1,000,000 gallons
of wastewater a day into a series of open waste treat-
ment ponds. The wastewater was reaching first anaero-
bic pond and then was flowing through a series of three
aerobic ponds before discharging into the river. Six
samples were collected. Sample 1 was upstream unpol-
luted water. Sample 2 was untreated waste effluent
from the processing plant. Sample 3 Was collected from
the anaerobic pond. Sample 4 was collected from the
second (middle) aerobic pond. Sample 5 was effluent
from the third aeroboc pond at the point of discharge
into the river. Sample 6 was collected from a point
approximately 0.1 mile downstream from the point of
discharge.
The samples were analyzed pyrographically, and the
resulting data are shown in Tables I and 2.
Table 1. Pyrographic Study of Waste
Treatment Process
Information that reflects the chemical nature of
the organic matter can be deduced from the pyrogram's
peak ratios. As shown in Table 1, the significant
changes in peak ratios are observed between peaks 1
(methane) and 2 (ethylene). For the natural water,
this ratio is about 60/100, and probably reflects a
composition of approximately 70-percent carbohydrates
and 30-percent protein, with small quantities of lipid
materials. For the untreated poultry waste, which is
Sample
Total
Peak Area
Assigned
Percent
Poultry Processing Plant
Untreated Waste
9.0
too
Anaarobtc Ho)ding Pond
3-5
33
Second Aerobic Holding Pond
1.5
17
Waste Effluent Into the River
1.1
)2
River 0.1 Mile From Discharge
0.18
2
Upstream Water
0.10
1
predominantly proteinacious in composition, the ratio
is 42/100. Under anaerobic conditions, this ratio
changes to 141/100, indicating a dramatic change of
chemical composition. As the level of organic matter
diminishes in the aerobic ponds and level of oxygen
rises, the Tatio of first to second peak diminishes
also. It is 100/100 in the second aerobic holding pond,
80/100 in the third, and reaches 5S/100 after mixing
with receiving water in the river. This progression
indicates a gradual shift from septic conditions to aero-
bic conditions. By assigning a value of 100 percent to
the untreated waste, one can obtain a quantitative meas-
ure of the effectiveness of the waste treatment pro-
cesses. The areas representing the organic matter in
each of the samples and their values relative to un-
treated waste are shown in Table 2. From these data,
it appears that the maximum reduction in organic matter
takes place in the anaerobic pond, which probably func-
tions also as a settling pond for particulate matter.
The organic matter from anaerobic and hence, pol-
luted, waters yield considerably higher proportion of
methane, than organic matter from well equilibrated
natural aerobic sources. Consequently, amounts of me-
thane as reflected in ratio of methane to ethylene re-
flect a degree of water septicity. It should be possi-
ble then to design a graduated scale of water quality
based on water septicity, taking as one extreme fully
anaerobic septic conditions, and on the other extreme
fully aerobic conditions as they prevail in the natural
unpolluted waters. From the ratio of methane to ethyl-
ene peaks as observed on pyrographic patterns, the de-
gree of water quality can be obtained,
Assimulation of Waste by a Waterstream
The overall changes in the composition of poultry
¦waste were studied in an open, moving waterway. In this
case, a poultry-rendering plant was discharging untreated
wastewater into the river, dramatically reducing water
quality' of receiving stream. The physical manifestation
of the water quality deterioration was characterized by
the presence of particulate organic matter in the stream
and by septic odor, the intensity of which varied from
day to day. Four water samples were collected for the
study. One sample was from an upstream (to waste dis-
charge) site. The second sample was that of untreated
poultry waste. The third sample was collected 2 miles
downstream from the point of discharge, and the fourth
sample was collected 7 miles downstream from the
discharge.
The pyrographic analysis was performed on each
sample and results are reported in Tables 3 and 4.
Examination of Tables 3 and 4 reveals similarity in
pyrographic patterns between corresponding samples of
the first case study (decomposition of waste in waste
treatment ponds) and corresponding samples of this
study. For example, the pattern for untreated waste
from poultry processing (first case) and poultry render-
ing (second case) operations are quite similar with
Sampie I;
S«mp>« 2.*
3s
Sampla V*
SaropW 5:
Sample t:
Peak
No.
Uprtrttm
Water
Poultry
Untreated
Wait#
Anaerobic
Pond
Second
Aerobic
Pond
Affluent
From Third
Aerobic
Pond
0.1 Mlt«
Downstream
1
60
42
r*t
100
89
55
2
100
100
100
100
100
too
3
31
n
*0
61
hk .
H
6
(2
9
37
5
3
28
15
6
10
Ut*
2
4-2
-------
Table 3. Pyrographic Study of Fate of
Untreated Waste in Moving Stream
P««k
He.
Upstream
W»t%r
Wast*
Blschirg#
2 HUti
0ownjtr««m
From CHscharg*
7 MfUt
Downstream
From Discharge
55
50
9>*
53
100
100
100
100
33
37
U2
26
2
10
15
11
14
9
Table 4. Waste Qecipation in Open Water Stream
Sample
Total Peak
Area
Ass i gned
Percent
.
Discharge of untreated
waste, poultry rendering
plant
12.7
100
2.
.2 miles downstream
from d i scharge
9-97
7.6
3.
7 miles downstream
from d ischarge
0.27
2- )
k.
Upstream of discharge
poi nt
0.17
1.3
methane to ethylene ratio of 60/100 in one case and
55/100 in the other. A sample collected at a point 2
miles downstream from the waste discharge (poultry ren-
dering plant) produced a pyrogram quite similar to the
pyrographic pattern for the second aerobic pond of the
poultry processing plant operation. Seven miles down-
stream from the waste discharge, the ratio of methane
to ethylene is 53/100 and is not unlike the ratio of
55/100 observed for a sample collected 0.1 mile down-
stream from the treated waste discharged from poultry
processing plant, This similarity in pyrographic pat-
terns suggests that a 7-mile stretch of Middle Occonee
River below the untreated waste discharge from poultry
rendering plant functions, in fact, as a waste treat-
ment facility leading to eventual assimilation of dis-
charged waste.
The condition of the river 2 miles downstream from
the point of waste discharge approaches septic condi-
tions; this is reflected both in the pyrographic data
and visual observation made at the site. On a gradu-
ated scale of water quality (one extreme fully anaero-
bic conditions, other extreme fully aeTobic equili-
brated waters) the 2-mile downstream point from the
untreated waste discharge appears to be similar to the
water quality of the second aerobic waste treatment
pond.
From the data reported here, further conclusions
can be made. On the basis of total pyrographic response
values (Table 4), we can calculate that the organic load
of river increases S.7 times as a result of discharge of
untreated waste from the poultry rendering plant, and
that waste materials constitute approximately 82 percent
of the organic load. The analysis of pyrographic data
for the subsequent 5-mile stretch of the river indi-
cates that waste matter is subjected to active processes
of assimilation and dilution. Overall reduction of or-
ganic load is approximately 70 percent, and at the point
7 miles downstream from the discharge, fully aerobic
conditions are re-established.
Conclusions
To fill the need for simple, direct and inexpensive
means for determining general or "gross" water quantity,
a conceptual approach based on measuring contribution of
organic waste matter to the receiving waters is proposed.
Such an approach is based on differentiating organic
waste contributions from the natural organic content of •
the waterway by means of pyrographic analysis.
The other simplified form of pyrographic analysis
can be used to develop a scale of relative septicity as
an indicator of "gross" water quality. Such a scale can
be based on ratios of methane to ethylene produced dur-
ing analysis.
Additional experimental and field work must be car-
ried out, however, before either scheme can be reduced
to practice and become a simple tool of measurement in a
typical water analysis laboratory.
References
1. Ihor Lysyj and P. R. Newton, EPA-R2-73-227, U. S.
Environmental Protection Agency, Washington, D.C.
(1973).
2. Ihor Lysyj and P. R. Newton, "Multicomponent Pattern
Recognition and Differentiation Method: Analysis for
Oil in Natural Waters," Anal. Chem. 44, 2385-2387
(1972).
3. Ihor Lysyj, P. R. Newton, and W. J. Taylor, "Instru-
mentation-Computer System for Analysis of Multicom-
ponent Organic Mixtures," Anal. Chem. 43, 1271-1281
(1971).
4. Ihor Lysyj, "A Pyrographic Instrument for Analysis
of Water Pollutants," Amer." Lab. 7, 23-25 (1971).
Acknowledgement
This work was supported by EPA Contract 14-12-802.
3
4-2
-------
TRACE ELEMENTS IN THE COMBUSTIBLE FRACTION OF URBAN REFUSE
Harold E. Marr, Stephen L. Law, and David L. Neylan
U.S. Department of Interior
Bureau of Mines
College Park Metallurgy Research Center
College Park, Maryland 2071»0
Abstract
The combustible fraction of urban refuse is being
considered for use as a supplement to coal for energy
production, but little data is available that can be
used to evaluate its pollution potential, in order
to fi11 this need for environmental impact infor-
mation, the Bureau of Mines has initiated research to
determine the concentrations of the trace, minor, and
major elements in the combustibles separated from
municipal refuse. Sampling and analysis of the
combust+bles are discussed, and comparisons are wade
with the elemental concentrations In coal and fly ash.
Introduct ion
Valuable metals, glass, and other materials,
including a combustible fraction composed mainly of
papers and plastics, are being recovered from urban
refuse in a pilot plant operated by the Bureau of
Mines at its College Park, Maryland, Metallurgy
Research Center.1 The most effective use of the
paper and plastic fraction is one problem presently
confronting those Involved In resource recovery. In
line with the Nation's concern about its energy
resources, power plant officials across the country
are seriously considering the paper and plastic
fraction of municipal solid waste as a possible fuel
source to supplement coal. Several cities, such as
St. Louis, Mo., Lynn, Mass., Chicago, 111., and
Rochester, N.Y., are already tooling up to use their
waste material as an energy resource.
The principal factors involved in the consideration
of refuse as fuel are primarily its availability as a
nondepleting resource, the economics compared to those
of fossi! fuels, and the environmental impact. A
comparison of the proximate analysis of coal suitable
for fuel with urban refuse as fuel is shown in
table 1. Many coals contain a much higher sulfur
content than shown in table 1, and depending on the
area of use, must be blended with coals having a low
sulfur content before being burned.
Table 1. - Comparison of proximate analyses of coal
and the combustible fraction of
urban refuse as fuel
Moisture, %
Ash, I
Volatile, %
Fixed carbon, %
Btu/lb, as reed
Btu/lb, dry
Sulfur, %
Chlorine, %
Refuse (3.7)
15-30
5-25
30-60
1-15
2,000-5,000
6,000-9,000
. 1-.3
.3-1-5
Coal W
3.5-7.0
9-12
35-AO
50-55
10,000-12,600
D,000-13,300
1-2.5
<0.1
Data related to objectionable elements released
during burning of the combustible fraction of munic-
ipal refuse mixed with coal are presently limited.6
Therefore, it is expected that concern will be
expressed by environmental groups about the possible
pollution resulting from the use of refuse as a fuel
supplement for generating electricity and heat.
Related available information comes from the
sampling of municipal incinerators where emissions
are not necessarily representative of emissions from
refuse-derived fuels.
The objective of the present analytical research
program at the College Park Metallurgy Research
Center is to obtain Information regarding the poten-
tially toxic elements present in the combustible
fraction of urban refuse. Future investigations will
follow the disposition of objectionable elements
during the combustion of refuse-derived fuels blended
with coal. As part of the present program the major
sources of trace and minor elements in refuse are
being identified and analyzed. High-volume contrib-
utors, such as newspapers, magazines, and plastic
containers, are being analyzed separately, and the
concentrations of the elements of concern are being
compared to those for the combustible fraction as a
whole. These data will be available soon. Available
at the present time, and presented later in this
report, are data on the composition of the bulk paper
and the bulk plastics as separate fractions, and
these data are compared with the total mixed
combustible fraction and with coal. In addition, the
fly ash emitted from coal-fired power plants is
compared with the composition of particles emitted
from incinerators processing municipal solid wastes.
Because Incinerators burn the entire solid waste as
received and not just the cleaner combustible
fraction, their emissions may be considered as a
worst-case model of the emissions possible from using
100 percent urban refuse as a power source.
The Analysis of Raw Refuse
Most power plants that will utilize urban refuse as a
fuel supplement will perform some separation before
incineration.5 The Bureau of Mines pilot plant,
using conventional shredders, magnetic separators,
trommels, air classifiers, and cyclones (see figure 1),
separates the combustibles of urban refuse Into
compacted bales suitable, with possible reshredding,
for feedstock for power plants. The samples used
in our analytical study were collected from several
pilot plant runs over a period of several months.
The sampling procedures for the combustible fraction
have been described in detail elsewhere.3
Essentially, grab samples are taken randomly until 10
4-3
-------
to 20 percent of the total product is obtained from
each run. The gross samples are progressively coned
and quartered to a final sample size of about 3
kilograms. After air drying, the sample is shredded
and mixed. For chemical analysis, the shredded
material is riffled and split to a final sample size
of approximately 4 grams. Several different analyt-
ical procedures were evaluated including ashing
followed by fusion or wet acid digestion, or wet
digestion without ashing using nitric, hydrofluoric,
and perchloric acids. Atomic absorption spectro-
photometry, X-ray spectrography, and optical emission
spectrography are used to determine the concentrations
of the elements of interest. The reproducibility of
the wet digestion-atomic absorption procedure
including the sampling is illustrated by the data in
table 2. Sample A is a light combustible fraction
collected in cyclone 2 (see figure l). Sample 8 from
the third cyclone contains the heavier paper and
plastics fractions of that particular refuse sample.
Because of the presence of a significant portion of
noncombustibles in the form of wire and foil, the
variations in concentrations of aluminum, copper, and
iron reflect the limitation due to the small sample
size used in the analyses.
Table I. - Triplicate analyses from two portions of
the combustible fraction of urban refuse
Concentration, parts per million
A-l
A-2
A-3
B-l
B-2
B-3
Aluminum
9,950
10,300
12,200
9,630
8,770
7,960
Bar ium
47
48
57
-
120
79
B i smuth
-
-
-
21
22
19
Cadmium
9
7
7
94
92
97
Ca1c ium
3,310
3,210
3,830 37,400 39,000 37,200
Chromium
44
4 2
35
52
56
60
Coba 11
<4
<4
-------
refuse than in coal—an important consideration in
utilizing municipal refuse as a fuel supplement.
Among the trace elements, copper, lead, manganese,
and cadmium are found at higher concentrations in
municipal refuse than in coal.
Other differences are apparent in the tables, but no
attempt can be made at this point in our research to
adequately explain them. The tables as presented
are working estimates which are expected to be
considerably refined as more data become available.
A comparison Is made In table 5 of crustal abundance
of some elements to fly ash from municipal inciner-
ators and fly ash from coal-fired power plants. This
information is important In determining the
enrichment of undesirable elements in dust particles
of urban air above that which might be expected from
natural sources. Many of the data in tables 4 and 5
are marginal, and conclusions must be cautiously
drawn. The objectives of the referenced studies
differed considerably as well as the experimental
testing procedures. Further studies are certainly
warranted, and some are presently in progress, such
as the fttdwest Research Institute tests at the St.
Louis Union Electric Company and the University of
Maryland project on emissions from major air
pollution sources (21-23).
Many questions will be faced during the course of
future research. Will many elements in refuse
combustibles such as aluminum, copper, and iron
remain in the bottom ash and not contribute to
atmospheric pollution? Will close monitoring of
elements forming volatile chlorides such as cadmium,
lead, and zinc verify that as a potential polluting
fuel, refuse is harmless? Is the burning of refuse
combustibles actually less of a threat to the
environment than the burning of coal? At the present
time, answers are just beginning to pass beyond the
speculative stage.
Summary
This paper has compiled the results from several
studies on the elemental composition of the combus-
tible fraction of urban refuse and compared these
results with similar studies on coal. These data may
be used to begin an evaluation of the environmental
impact of utilizing urban refuse as a fuel supplement
in the generation of heat for energy production.
Considerable work remains to be done to characterize
municipal solid waste and to establish the dispo-
sition of the objectionable elements during the
combustion process. An Increasing confidence in this
valuable resource will encourage more communities to
conserve other fuels by supplementing fossil fuels
used for power production with fuels derived from
refuse.
Ac knowIedgmen t s
The authors are grateful for the analytical support
provided by the Analytical Services Group and John
Novak, James McConnell, William Artlcola, and Robert
Gabier of the College Park Metallurgy Research
Center, to the Union Electric Company for discussions
of the reports of testing for the solid waste
processing facilities of the Meramec Power Plant, to
Forrest Walker of the Pittsburgh Energy Research
Center, and to Paul Sullivan and the Secondary
Resource Recovery Group of the College Park
Metallurgy Research Center for many helpful discus-
sions as well as supplying and preparing the samples.
References
(1) Sullivan, P. M., and H. V. Makar, "Bureau of
Mines Process for Recovering Resources from Raw
Refuse," Proc. Fourth Mineral Waste Utilization
Symposium, Chicago, Illinois (197^), 128-1^7.
(2) Hooper, R. E., "A Nationwide Survey of Resource
Recovery Activities," EPA/530/SW-142 (1975),
74 pp.
(3) Schultz, H., P, M. Sullivan, and F. E. Walker,
"Characterization of the Combustible Portions
of Urban Refuse for Potential Use as Fuel,"
BuMines Rl 8044 (1975), 26 pp.
(4) Unpublished estimates from J. B. Janus, Office
of Coal Sampling and Inspection, Bureau of Mines,
College Park, Maryland.
(5) Klumb, D. I., "Solid Waste Prototype for Recovery
of Utility Fuel and Other Resources," presented
to 1974 Air Pollution Control Association,
Denver, Colorado, May 1971*•
(6) Kaiser, E. R., and A. A. Carotti, "Municipal
Incineration of Refuse With 2 Percent and 4
Percent Additions of Four Plastics: Polyethylene,
Polyurethane, Polystyrene and Polyvinyl
Chloride," Proc. 1974 National Incinerator
Conference, New York, N.Y. (1972), 230-244.
(7) Un ion Electric Company internal report.
(8) Sharkey, A. G., T. Kessler, and R. A. frledel,
"Trace Elements in Coal Dust by Spark Source Mass
Spectrometry," Trace Elements in Fuel, Advances
in Chemistry Series 14) (1975), 48-56.
(9) Abernethy, R. F., M. J. Peterson, and F. H.
Gibson, "Spectrochemical Analysis of Coal Ash
for Trace Elements, "BuMines ft! 728) 0969),
30 pp.
(10) Abernethy, R. F., and F. H. Gibson, "Rare
Elements in Coal," BuMines 1C 8163 (1963), 69 pp.
(11) Gibson, F, H., and W. A. Selvig, "Rare and
Uncommon Chemical Elements in Coal," BuMines
Tech. Paper 669 (1944), 23 pp.
(12) Selvig, W. A., and F. H. Gibson, "Analysis of
Ash From Coals of the United States," BuMines
Tech. Paper 679 (19*15), 20 pp.
(13) Berman, M., and S. Ergun, "Analysis of Mineral
Matter in Coals by X-Ray Fluorescence," BuMines
Rl 7124 (1968), 20 pp.
(14) Wesson, T. C., and F. E. Armstrong, "Elemental
Composition of Coal Mine Dust," BuMines Rl 7992
(1974), 25 pp.
(15) Zubovic, P., T. Standnichenko, and N. B. Sheffey,
"Minor Elements In American Coals, Geochemistry
of Minor Elements in Coals of the Northern Great
Plains Coal Province," Geo!. Survey Bull, II?7~A
(I960.
(16) Schlesinger, M. 0., and H. Schultz, "An
Evaluation of Methods for Detecting Mercury in
Some U.S. Coals." BuMines Rl 7609 (1972).
3
4-3
-------
(17) Lukens, H. R., H. L. Schlesinger, D. M. Settle,
and V. P. Quinn, "Forensic Neutron Activation
Analysis of Paper Rept to USAEC TID - ^*500
(1970), 50 pp.
(18) Brunelle, R. L., W. D. Washington, C. M. Hoffman,
and M. J. Pro, "Use of Neutron Activation
Analysis for the Characterization of Paper,"
J. of the AOAC, Vol 5^ (1971), 920-921).
(19) Results of analyses, College Park Metallurgy
Research Center, College Park, Maryland.
(20) Krauskopf, K. B., "Introduction to Geochemistry,"
New York (1967), 721 pp.
(21) Rubel, F. N., "Incineration of Solid Wastes,"
Voyes Data Corporation (197^), 2^6 pp.
(22) Gordon, G. E., et al, "Study of the Emissions
From Major Air Pollution Sources and Their
Atmospheric Interactions," Progress Report,
University of Maryland (197^), 322 pp.
(23) Greenberg, R. R., W. H. Zoller, and G. E. Gordon,
"Municipal Incinerators: Source of Toxic Elements
on Urban Aerosols," to be published (1975)•
(24) Kaakinen, J. W., and R. M. Jordan, "Determination
of a Trace Element Mass Balance for a Coal-Fired
Power Plant," Proc. NSF Trace Contaminants
Conference, Oak Ridge (1973). 165-181).
(25) Davison, R. L., D. F. S. Natusch, J. R. Wallace,
and C. A. Evans, Jr., "Trace Elements in Fly
Ash," Environ. Sc i. Technol . 8^ (19710, 1 1 07 -1 1 13 -
(26) Quillaumin, J. C., "Determination of Trace Metals
in Power Plant Effluents," At. Absorpt. Newsl.
il (1974), 135-1 ^0.
(27) Cuffe, S. T., and R. W. Gerotte, "Emissions From
Coal Fired Power Plants," Public Health Service
999-AP-35 (1970), 26 pp.
Unbgmed refuse
PRIMARY
SHREDDER
KEY ;
I—Atr to baghouse
Water to recycle
system
(classifier.
Very Ught paper
and ploihc
, MAGNETIC
SEPARATOR
CYCLONE
Magnetic metal
to detinning
CYCLONE
«, No 2 .
Massive metois
COMPACTOR
Giast, nonferrous metois
and heavy organics
with paper and plastics
MAGNETIC
SEPARATOR
Aluminum and
heovy organics
with paper and
plastics
SEC0N0ARY
SHREDDER
Heavy
nonferrous
Gloss, nonferrous metois,
heavy food and other organics
iSECONDARY AIR
I CLASSIFIER
r1-!™
CYCLONE
WATER
ELUTRIATOR
Glass and
aluminum
MINERAL
Organic
Paper and plastic
ELECTROSTATIC
SEPARATOR
Heavy
nonferrous metois
Heavy
organic;
COMPACTOR
Paper and plastics
ROLLS CRUSHER
SCREEN
Aluminum
FROTH FLOTATION
PRIMARY
AIR CLASSIFIER
Pure
gloss
it
-------
TABLE 4. - Comparison of coal, paper, and plastics used as feedstock for power plants
Coal Paper Urban refuse
Handplcked plastics Paper and plastics
References
8-
16
17-
19
19
7,
19
Major elements
(conc. in pet)
Typical
value
Range
Typical
value
Range
Typical
value
Range
Typical
value
Range
Aluminum
0.14
.1-2.0
1 .0
0.3-3-0
0.2
0.1-1 .0
1.1
0.3-1 -6
Calc i um
.03
.007-.50
• 5
.04-2.7
>.2
• 5
.23-1 -0
Chlor ine
.01
0-.1
.1
.05-2.5
0-50
.4
.3-1.5
1 ron
.1
.01-1.0
• 03
.01-.06
• 5
.02-1 .0
.2
•05-.7
Magnesium
.02
.01-.4
• 05
.002-.2
.1
.02-.5
.1
.05--8
Phosphorous...
.006
<.03
<.02
.1
.001--7
Potass ium
.2
.01-.60
• 03
.01-.10
.2
.02-.5
• 07
.03--2
S i1 icon
.2
.08-4.1
>2.
4.
1 .-10
Sod i um
.2
.01-.35
¦ 03
.01-.25
.6
.05-1
¦ 5
.15-.9
Sulfur
1 .2
1.-2.5
<•3
.02-.2
.2
.1-.3
Ti tanium
.006
.003*.18
1.0
.01-1.7
.1
.02-.4
.2
.07-.5
Z i nc
.003
.001-.10
.08
.005-.02
.01
.002-.2
.10
.04-.8
Minor ^elements
(conc. in ppm)
Typical
value
Range
Typical
value
Range
Typical
va 1 ue
Range
Typical
value
Range
Rare earths...
<50.
10.
<600.
Antimony
20.
1-1,800
3-
.02-250.
2.
2-200
45.
20-80
Arsen ic
45.
1-70
.001-9
<5
<3
Bar ium
80.
20-1,600
1 .
1-10,200
100.
20-1,000
50.
35-100
Beryi1ium
25.
.4-90
<4
<2
<2
Bismuth
.2
.02-3
<15
<10
22.
3-45
Boron
100.
1-270
1 .
.1-20
2.
1-20
15.
5-70
Bromine
2.
1-25
.01-160
Cadmi um
.5
.2-5
¦ 03
.01-19
2.
1-5
15.
3-70
Ces ium
.3
.1-9
6.
<35
<20
<20
Chromium
I.
.3-400
20.
.4-330
100.
20-200
30.*
10-175
Coba it
25.
.3-135
1-4
<8
5.
2-17
Copper
20.
1-180
12.
.04-100
100.
20-500
195.
30-450
Ga11i um
7.
.3-36
<140
<60
German i um
45.
.03-1,000
<15
<10
<6
Gold
.1
.02-.5
• 03
.01-90
<2
1nd ium
.3
<•9
-------
TABLE 5- - Comparison of crusta) abundance to fly ash from municipal
incinerators and coal-fired power plants
Crusta 1
abundance
Municipal
Incinerators
Fly ash
Coal power plants
References
20
21
-23
24'
-27
Major elements
(conc. in pet)
Typical
value
Range
Typical
value
Range
A1 urn inum
8.2
11 .
.1-12
10.
1-21
Ca ic ium
4.1
4.
3-10
3-5
.8-6.3
Chlorine
.013
.8
.5-1.0
.02
.01-.05
1 ron
5.6
5.
4.5-5-5
12.
8-18
Magnesium
2.3
1.
1-10
• 35
.02-1.4
Phosphorous...
.10
1.1-1.5
• 3
.26-.36
Potassium
2.1
4.3
1.2-4.9
S i 1i con
28.2
20.
14-34
Sod i um
2.4
1.4
1-10
.2
.1-1
Sulfur
.026
8.
.01-49
Titanium
.57
3.
.5-5
• 5
.001-.9
Z i nc
.007
1 .
.8-10
.1
.01-1.3
Minor elements
(conc. in ppm)
Typical
value
Range
Typical
value
Range
Rare earths...
100.
100.
10-10,000
300.
<500
Ant imony
.2
230.
200-2,300
20.
1-53
Arsen ic
1.8
40.
30-190
200.
120-1,700
Ba r i um
425-
2,500.
100-3,500
1,100.
500-10,000
Beryl 1ium
2.8
10-100
24.
7-60
B i smuth
.17
2.
2-5
Boron
10.
100-1,000
>800
Bromine
2.5
2,300.
200-2,300
10.
4-20
Cadmium
.2
65.
33-900
12.
1-35
Ces ium
3-
1.
.8-1.2
4.
2-6
Chromium
100.
1,600.
400-1,800
170.
70-3,300
CobaIt
25-
15.
12-18
70.
27-130
Copper
55-
1,000.
200-1,500
250.
120-490
Fluorine
625.
1,000.
420-2,000
Ga11i um
15-
40.
36-130
German ium
1.5
50.
33-66
Gold
<.05
2.
¦ 5-3
Hafnium
3-
5.
4-6
2.
1-10
1nd i um
.1
.6
0-7
Iodine
• 5
30.
0-60
100.
•7-150
Lanthanum
25-
43.
35-50
90.
32-100
Lead
12.5
10.
5-15
700.
30-1,600
Lithium
20.
350.
290-470
Manganese
950.
3,400.
1,000-10,000
200.
150-1,300
Molybdenum....
1.5
15.
10-20
Nickel
75.
200.
200-10,000
160.
100-I,600
Niobium
20.
20.
10-40
Platinum
A
O
Rubid ium
90.
200.
50-500
Scandium
22.
12.
10-15
35.
20-44
Selenium
.05
10.
3-20
10.
7-28
SiIver
.07
4.
1-300
3.
.5-8
Strontium
375.
800.
300-1,500
Tantalum
<.05
5.
3-6
1.5
1.3-3-6
Tellur ium
<.05
.1
.10-.11
Tba 11 ium
.45
4.
2-4
Tin
2.
10.
5.
2-19
Tungsten
1.5
19.
16-22
5.
3.7-8.8
Vanad ium
135.
130.
10-200
320.
150-1,400
Yttrium
33-
100.
50-400
Z i rcon ium
165.
60.
30-315
6
4-3
-------
SPARK SOURCE MASS SPECTROMETRY:
THE OPTIMUM REFEREE METHOD FOR MULTIELEMENT ANALYSIS
Charles E. Taylor
Environmental Protection Agency
Environmental Research Laboratory
Athens, GA 30601
Summary
Spark source mass spectrometry (SSMS) has
been applied to a variety of samples for
trace-multielement analysis at the Athens
Environmental Research Laboratory. Analysis
of water samples by SSMS can routinely yield
data on up to 82 elements simultaneously at
the part per billion level. Comparison of
SSMS analyses with those of instrumental
neutron activation analysis, plasma emission,
atomic absorption, and x-ray fluoresence has
revealed unexpected errors in the latter
methods associated with the complex and
variable matrixes of environmental samples.
************
"What trace elements are present in my
sample?" is a common request confronting an
analyst when he is presented with a sample of
environmental materials. One would think
that by 1975 methods of trace analysis would
be more widely applied to environmental
samples. We have sent men to the moon and
brought them back with samples. We know more
today about the chemical nature of certain
parts of the moon's surface than we know
about soils, sediments, and water from some
of the earth's most polluted surfaces. The
request for a total analysis of a sample is a
reasonable one since pollution becomes a
problem when the environment is adversely
affected by elements or their compounds in
various concentrations or combinations.
Many instrumental techniques are being
applied to trace analyses on environmental
materials; however, only a few can be usee? to
analyze for trace amounts of a broad range of
elements. Instrumental neutron activation
analysis (IHAA), plasma emission (PE), x-ray
fluorescence (XRF), atomic absorption (AA>,
and spark source mass spectrometry (SSMS) are
currently being used for multielement
analysis. Uach technique has strong and weak
points for any given sample matrix, which
depend upon the instrumentation and the
natural phenomenon on which the technique is
based.
Three criteria should be considered in
evaluating referee techniques for
multielement analysis:
• The technique must be capable of
measuring a broad range of elements.
• It should be applicable to both solid
and liquid samples.
• It should have low minimum detection
limits and suffer minimal matrix and
interelement interferences.
Of the techniques mentioned, ifJAA, PE,
XRF, and SSMS cover broader ranges of
elements. PE has the capability for 20 to 30
elements and xrf and INAA systems usually can
analyze for about 40 elements. SSMS
capabilities extend to 82 elements.
Multielement analysis with AA, however, is
more difficult than with INAA, PE, XRF and
SSMS. In addition, AA excitation lamps are
specific for a maximum of 4 elements, with
the majority of lamps capable of only single
element analysis.
INAA, XRF, and SSMS all require minimal
samples preparation for solids. PE
techniques are relatively undeveloped at the
present time for direct solids analyses.
Liquids can be analyzed directly with INAA,
XRF, and PE. SSMS requires evaporation of
liquids onto a support electrode material for
analysis.
INAA, PE, XRF, and SSMS are well
qualified with respect to detection limits
and freedom from interferences. In analyses
of cuts from the same sediments (Figure 1)
SSMS reported ^60 elements quantitatively,
and INAA reported 14 elements quantitatively
together with 27 others qualitatively. (When
current modifications are made the INAA
system will be capable of analyzing sediment
samples for about 45 elements.) On the same
samples, XRF analysis yielded quantitative
data for 20 elements. No data were available
from PE.
In 4 years of experience at the
Environmental Research Laboratory-Athens,
SSMS has been found to be a reliable routine
referee technique for multielement analysis
when used with INAA, PE, XRF, and AA on three
points:
• SSMS yields data on a wide range of
elements on a routine basis.
• SSMS has desirable detection limits in
complex matrixes.
• The operating parameters for
instrumentation are constant for all
samples, i.e., no optimization
procedures are necessary to enhance
the sensitivities for given elements.
The element capabilities for INAA, PE, and
SSMS available at the Athens laboratory are
presented in Figure 2.
The radiofrequency spark source is the
component that makes SSMS so broadly
applicable. In 1934 Dempster demonstrated
the use of an RF power supply to vaporize and
ionize material from a two-electrode source.
Based on his work, he predicted a technique
1
4-4
-------
with broad applicability and very good
sensitivity. During the 1950's, the Mattauch
and Herzog mass spectrometry design was
married to the spark source. The use of the
double focusing mass spectrometer to analyze
ions created by the spark source fulfilled
Dempster's prediction.
The basic components of the SSMS are the
source, the analyzer, and the detector, all
of which are maintained under high vacuum. A
radiofrequency power supply of -v-l megacycle
provides the source power to create ions. A
voltage drop of 10"s of kV's is imposed
across two electrodes, spaced a,10~3 cm
apart. A vacuum breakdown occurs between the
electrodes, initiating the electrical vacuum
discharge. The resulting positive ions are
accelerated from the source into an electric
sector, which acts as an energy analyzer.
From the electric sector, the ions are
deflected into the magnetic sector and are
separated according to the m/e (mass to
positive charge) ratio. Either photographic
or electrical detection methods are used in
SSMS systems, depending on the nature of the
analysis and the type of sample.
Computerized electrical detection is used
most frequently at our laboratory.
When SSMS is used as a comparison
technique, results will be similar to those
shown in Figure 1. Results from studies such
as those on the sediments have led the SSMS
at the Athens laboratory to be used primarily
as a trace survey analysis technique. It is
also used to help evaluate and referee other
methods in which matrix or interelement
effects are problems.
Group
la
Ha
nib
IVb
Vb
VIb
VHb
VIHb
lb
lib
Ilia
IVa
Va
Via
Vila
Inert
Gas
PERIOD
1
H
H
He
2
Li
Be
B
C
N
0
F
•
Ne
3
•
Al
•
Si
P
•
S
•
A
4
(HI
0
©.
•
Vj
•
Cr*
•
Mn
•
Fe*
•
•
Co
•
Ni
«
•
Cu
•
•
Zn
•
Ga
•
Ge
•
•
As
•
Kr
5
ft
Sr
•
Y
•
Zr
•
Nb
•
Tc
Ru
«
Rh
•
Pd.
(§£
In
•
Sn
•
Sb
(§1
CD
Xe
6
4
Ba
•
•
La
•
<3
Ta
•
Re
Os
Ir
Pt
Hg'
Tl
•
Pb.
Bi
•
Po
At
Rn
7
Fr
Ba
Ac
8 "
©
Pr
. f
Pm
©
(E^|Gd
Tb.
Ho
•
Er
•
Tm
»
Q
1-
5f
<5}
Pa
<3
Np
Pu
Am
Cm
Bk
Cf
Es
Fm
Md
No
Lw
NEUTRON ACTIVATION {OUALi
NEUTRON A CT! VA TiON
SPARK SOURCE
Figure 1
Group
la
{la
nib
IVb
Vb
VIb
VHb
VIHb
lb
fib
Ilia
IVa
Va
Via
Vila
Inert
Gas
PERIOD
1
H
H
He
2
Li
_*
•
Be
z 1
•
B
•
C
N
0
F
Ne
3
Na
Mg(
• •
Al
Si
•
P
s
•
CI
A
4
K*
i
"fca
i
Sc
... i
•
Ti
j
V
... J
» •
Cr
.... i
• a
Mn
-. .1
*Fe
•
• •
Co
•
• 4
Ni
• •
Cu
• •
Zn
Ga
Ge
• •
As
*Se*
Br
Kr
5
Rb
J!
Sr
i
•
Y
a
Zr
_ 1
Nb
i
Mo
•
Tc
Ru
•
Rh
Pd
A",
• •
Cd
In
•
Sn
Sb
3
Te
4
1
Xe
6
Cs
j
• «
Ba
•
La
Hf
- ,f
«
Ta
•
W
•
Re
•
Os
•
ir
_ •
Pt
Au
Hg
Tl
•
Pb
Bi
Po
At
Rn
7
Fr
Ra
Ac
K "
Ce
-j
Pr
.. »
Nd
—i
Pm
,
Sm
. <
Eu
«
Gd
... .*
Tb
*
Dy-
Ho
...«
Er
—i
Tm
.. s
Yb
... _t
Lu
.. . *
i-
5f
Th
^
Pa
U
.. .i
Np
Pu
Am
Cm
Bk
Cf
Es
Fm
Md
No
Lw
PLASMA EMISSION
NEUTRON ACTIVATION
SPARK SOURCE
Figure 2
2
4-4
-------
FREQUENCY DISTRIBUTIONS FOR COLIFORM BACTERIA IN WATER
Wesley 0. Pipes, Drexel University, Philadelphia, Pennsylvania
Pamela Ward, Water Pollution Control Federation, Washington, D. C.
S. H. Ahn, Consoer Townsend & Associates, Chicago, Illinois
Summary
Five sets of data from MF collform counts on
raw water from a water treatment plant were analyzed
to determine the best fit statistical frequency dis-
tribution. It was found that for all five cases the
data fit a negative binomial distribution and did not
fit a Poisson distribution. Since 1908 It has been
assumed that collform data would fit a Poisson dis-
tribution. The implications of assuming a Poisson
distribution for data which has a negative binomial
distribution are discussed.
Introduction
Collform determinations are the tests most wide-
ly used for estimating the potential of water for
transmitting waterborne disease. The two methods
used for obtaining coliforni data are the multiple
tube fermentation technique and the membrane filtra-
tion (MF) technique.
The original assumption that collform bacteria
are randomly dispersed in water was made by Phelps
in 1908 and this assumption Is the basis for the use
of the Poisson distribution for analysis of collform
2
data. McCrady using the Poisson distribution de-
vised a method of estimating the mean collform densi-
ty most likely to give any particular result from a
multiple tube fermentation test on a series of deci-
mal dilutions. This mean density most likely to give
the results obtained is called the most probably num-
ber (MPN) and the multiple tube fermentation tech-
nique is commonly called the MPN test. The statisti-
cal methodology for estimating MPN values for multiple
tube fermentation results has been extensively elab-
orated by a number of investigators over the past
several decades.
3
Velz , using a graphical technique, found that
MPN values for a series of water samples would fit
a log-normal distribution and suggested the use of
this distribution for analysis of coliform data.
4
Thomas stated that he had found that bacterial
counts from replicate pour plate determinations would
fit a gamma distribution and suggested calculating
the mean MPN values from the parameters of a gamma
distribution calculated from the multiple tube fermen-
tation test results. After the MF technique had come
into general use, he showed that if bacterial numbers
in a water sample fit a gamma distribution the MF
counts should fit a negative binomial distribution.^
MF coliform counts are often analyzed using the log-
normal distribution on an emperical basis but Thomas'
suggestion of using the negative binomial distribution
has not been adopted.
The purpose of this paper is to re-examine the
original assumption that the Poisson distribution is
suitable for description of the dispersion of coli-
form bacteria in water. The investigations described
were based on the hypothesis that the negative binomi-
al distribution would give a better fit for coliform
data than the Poisson distribution.
Methods
Coliform Data
The data needed for testing the hypothesis are
large numbers of coliform counts on some definable
body of water. That portion of Lake Michigan which
serves as a water supply for Evanston, Illinois, was
selected. Coliform counts on the raw water but not
on the finished water were used. With time and flow
as variables there are several different parent popu-
lations of coliform densities which can be sampled.
These include the coliform densities of a single
sample of water, the coliform densities in the water
pumped through the treatment plant in a few hours,
the coliform densities in the raw water supply over
some extended period of time. The parameter usually
estimated in a water treatment plant is the mean coli-
form density over a 24-hour period but all three of
the parent populations cited above are of interest in
water quality evaluations.
To examine the distribution of coliform densi-
ties in the water supply over an extended period of
time, the data from the daily coliform counts (usually
two grab samples per day) made by the laboratory tech-
nician at the water treatment plant for the period of
January 1972 through July 1973 were used.® The number
of collform counts for this first data set was 1149.
For examination of the distribution of coliform den-
sities in a single sample, multiple replicate deter-
minations were made on individual samples of slightly
more than 10 liters each.^ The number of replicate
determinations (n) was 62 for the second data set, 63
for the third, and 55 for the fifth. The fourth data
set consisted of coliform determinations on samples
(n ¦ 60) collected at five minute intervals from the
raw water supply as it was pumped into the treatment
plant. All coliform counts were made on 100 ml por-
tions using the membrane filter technique.
Data Analysis
The first data set was used for a test to deter-
mine if the best fit was achieved using the normal,
log-jjormal, Poisson, binomial, or negative binomial
distributions. The other four data sets were tested
only against the Poisson and negative binomial dis-
tributions. The parameters for each distribution
were estimated fTom the raw data. The data were then
arranged on a histogram according to selected class
intervals. The expected frequencies for each of the
class intervals were calculated from the theoretical
frequency distributions whose parameters had been cal-
culated from the raw data. The expected frequencies
were tested against the observed frequencies using
the chi square goodness-of-fit test.
The selection of class intervals has an influ-
ence on the outcome of the chi square test because
the number of class intervals selected determines the
degree of freedom for this test. The criterion for
selection of class Intervals was to make the intervals
as small as possible and still have a reasonable fre-
quency of counts in. each interval.
1
4-5
-------
O Negative Binomial
tog-Normal
O Poisson
f—| Binomial
x
u
§
| 0.2-
«
>
o
4)
oc
30
Coliform Count (per 100 ml)
Figure 1. Treatment Plant Data (n-1149) Compared With Frequency Distributions.
Results
A histogram representing the first set of data is
presented in figure 1. Also plotted on the histogram
are curves representing four theoretical frequency
distributions with parameters estimated from the data
set. From the figure it is clear that the negative
binomial distribution is the only one which comes
close to describing the actual distribution of the
data. The normal distribution waB also tested but
it fit the data so poorly that it did not seem worth-
while to include it in the figure. The chi square
goodness-of-fit test resulted in rejection of the
Poisson, binomial, normal, and log-normal distributions
at the 5% confidence level. The negative binomial dis-
tribution was accepted at the 10% level but rejected at
the 20% level.
0.5-
0,4-
P-3—
O Negative Binomial
O Poisson
Coliform Count (per 100 ml)
Figure2. Data From Replicate Portions n«62 of a Single Sample
2
4-5
-------
O Negotive Binomial
O Poiiion
Coliform Count (per 100 ml
Figure 3. Data From Replicate Portions (n«63)of a Single Sample
The analysis of the first data set showed that the
negative binomial appeared to be the best Model for
daily coliform counts on a raw water supply over an ex-
tended period of tine. However, it did not give any
information about measuring coliform densities in sin-
gle samples or in a water supply over a short period of
time.
The second and third data sets are presented as
histograms in figures 2 and 3 respectively and the
fourth data set is presented in figure 4. From figures
2 and 4 it is quite clear that the negative binomial
distribution provides a much better fit to the data
than the Poisson distribution for replicate determina-
tions on a single sample and for a series of determina-
tions on a raw water supply over a period of five hours.
The difference between the two distributions is not so
apparent from figure 3.
The parameters estimated for the second, third,
fourth, and fifth data sets and the results of the chi
square tests are presented in Table I. The negative
binomial distribution gave a lower ciii square value and
therefore a better fit for all four data sets. The
difference in the goodness-of-fit for the two distribu-
tions is less for small values of mean density (data
set 3) and for large values of the parameter k (data
set 5). These last two observations can be predicted
mathematically.
Discussion
It is reasonably clear that these data support the
hypothesis that the negative binomial distribution pro-
vides a better description of MF coliform counts than
the Poisson distribution. Both distributions are mem-
bers of the binomial family. The characteristic
O Negative Binomial
O Poisson
—H
0.2-
0.1-
Coliform Count per 100 ml
Figure4. Data From Samples n*60 of Raw Water Over A Five Hour Period
3
4-5
-------
TABLE I
STATISTICS
FOR MEMBRANE FILTER COUNTS
Data Set
2
3
5
n
62
63
60
55
X
6.32
0.67
15.48
4.98
s2
35.39
1.04
140.09
9.21
k
1.04
0.83
2.19
6.4
X2 for neg. binomial
10.85
5.43
36.20
6.8
degrees of freedom
15
2
23
9
rejection level
70%
10%
5%
70%
acceptance level
50%
5%
2%
50%
X2 for Poisson
176.34
12.67
294.49
17.56
degrees of freedom
9
2
22
8
rejection level
0.1%
1%
0.1%
2.5%
acceptance level
-
0.5%
-
1%
difference is that for the Poisson distribution the
variance is equal to the mean while for the negative
binomial distribution the variance is greater than the
mean. The results presented are equivalent to a demon-
stration that the variance is significantly greater
than the mean for these data.
If the Poisson distribution is assumed, only one
parameter, X, is estimated. This parameter is estima-
ted from the mean of the observed counts and is assumed
to be also an estimate of the variance. Assuming a
Poisson distribution for data which has a negative bi-
nomial distribution results in an underestimate of the
variance of the data.
The negative binomial is a two parameter distribu-
tion. The two parameters usually estimated are the
mean, p, and k. For this distribution the variance is
given by
Clearly 1/k is a measure of excess variance and small
values of k result in a variance much, much greater
than the mean. At high values of k there is little
difference between the Poisson and the negative bino-
mial distribution. The values of k obtained for the
last four data sets ranged from 0.83 to 6.3. An esti-
mate of the common k for all four data sets was 1.90.
Many more sets of data would be required to demonstrate
that it is reasonable to calculate a common k for con-
form data from a single source. However, if it is
assumed that k is constant for MF coliform determina-
tions and that it is in the range of 1 to 2, it is then
possible to demonstrate another consequence of assuming
a Poisson distribution for data having a negative bino-
mial distribution.
For the Poisson distribution the relative frequen-
cy of zero counts is
P(0) - e~g
For the negative binomial distribution the relative
frequency of zero counts is
P(0) = (1 + £)"k
For k » 1, P(0) ¦ -jp- ; for k « 2, P(0) » ~ 2 , etc. ;
2+y
For any value of vi the negative binomial will give a
higher frequency of zero counts than the Poisson dis-
tribution. Calculation of a MPN value is based upon
using the proportion of zero counts from a multiple
tube fermentation test as an estimate of the frequency
of zero counts and then using this estimate to estimate
the mean density. If a Poisson distribution is assumed,
for N cultures, S of which showed no growth
S/N = P(0) = e"x
and x = £n N/S.
Since assuming a Poisson distribution for data having a
negative binomial distribution gives a low estimate of
the frequency of zero counts, using the proportion of
zero counts to estimate the mean density will give a
low estimate. The quivalent expressions for estimating
mean density from the proportion of zero counts, assum-
ing a negative binomial distribution are
x = N/S - 1 for k ¦ 1
and x = 2vf7s - 2 for k - 2.
For instance, if S/N «= 0.2 (1 of 5 cultures showing no
growth) the estimates of x would be 1.61 for the Poisson
distribution, 4 for the negative binomial distribution
with k = 1, and 2.47 for the negative binomial with
k - 2.
The logarithmic transformation is often used to
"normalize" data which have a negative binomial distri-
bution. The log-normal distribution is often used for
analysis and interpretation of MF coliform counts.
This demonstration that MF coliform counts do fit a
negative binomial distribution provides a theoretical
basis for this procedure. Even so, the negative bino-
mial distribution appears to provide a better fit for
MF coliform counts than the log-normal distribution
and should be used for statistical procedures where
normality of the data is not required.
The negative binomial distribution has gained wide
use in recent years for description of population den-
sity data for a number of different species of plants
and animals. When population density data fit a nega-
tive binomial distribution, this is taken to indicate
that the population has a clumped or aggregated dis-
persion rather than a random dispersion. Undoubtedly,
coliform bacteria often exist in clumps of two or three
or a few bacterial cells physically adhering together.
However, this demonstration that MF coliform counts fit
a negative binomial distribution does not give any in-
formation about clumping in the sense of cells adhering
together. Each colony on a MF plate could have been
started by one or more bacterial cells just as each
positive tube in a multiple tube fermentation teat
4
4-5
-------
could he the result of one or more cells In the por-
tion of water planted In that tube. What has been
demonstrated is simply that for MF coliform counts
the variance is signficantly greater than the mean.
Inferences about the physical arrangement of coliform
bacteria in a body of water are unwarrented at this
time.
Conclusions
1. The data presented support the hypothesis that
MF coliform counts do fit the negative binomial
distribution but do not fit the Poisson distri-
bution.
2. If this hypothesis is true, the use of the
Poisson distribution to represent coliform
counts leads to an underestimate of the vari-
ability of these counts.
3. If this hypothesis is true, the calculation
of MPN values from multiple tube fermentation
tests leads to underestimation of mean coliform
densities.
Acknowledgments
The generosity of Mr. Harry Von Huben, Superin-
tendent of the Evanston Water Department, in making
available coliform data and facilities for obtaining
raw water samples. At the time this work was done
W. 0. Pipes was Professor of Civil Engineering, and
Pamela Ward and S. H. Ahn were graduate students in
the Department of Civil Engineering, Northwestern
University.
References
1. Phelps, E. B. "A Method for Calculating the
Number of B_. coli from the Results of Dilution
Tests," Am. J. of Public Hygene, 4: 141-145
(1908).
2. McCrady, M.H. "The Numerical Interpretation of
Fermentation Tube Results," £. Infectious Dis-
eases , 17_: 183-212 (1915).
3. Vela, C. J. "Graphical Approach to Statistics,
Part 4: Evaluation of Bacterial Density,"
Water and Sewage Works, 98: 66-78 (1951).
4. Thomas, H. A. "On Averaging the Results of
Coliform Tests," J. Boston Soc. Civ. Engrs., 39:
253-270 (1952).
5. Thomas, H.A., R. L. Woodward, and P. W. Kabler.
"Use of Molecular Filter Membranes for Water
Potability Control," J_. Am. Water Works Assoc.,
4£: 1391-1402 (1956).
6. Ward, P. S. "An Evaluation of the Statistical
Basis for Coliform MPN Calculations," M.S. Thesis
Department of Civil Engineering, Northwestern
University, Evanston, Illinois (1974).
7. Ahn, S. H. "Distribution of Total Coliform
Bacterial in Drinking Water Supply," M.S. Thesis
Department of Civil Engineering, Northwestern
University, Evanston, Illinois (1975).
5
4-5
-------
"TRACE METAL ANALYSIS BY ATOMIC EMISSION USING A D.C. ARGON PLASMA"
By
Walter G. Cox
Material and Chemistry Branch
Naval Underwater System Center
Newport, R.I. 02840
Introduction
Documentation of trace metals in even
smaller concentrations lead to the develop-
ment and refinement of such analytical tech-
niques as atomic absorption with cold vapor
and graphite furnace methodologies. The
challenge of an improved rapid, accurate low-
level direct measurement which would reach
or surpass the detection limits of atomic
absorption (with new modifications) has been
aptly met by D.C. Argon plasma atomic emiss-
ion. Matrix effects which have traditionally
plagued atomic absorption are minimized or
eliminated by D.C. Argon plasma atomic emis-
sion, and the overall acquisition and main-
tenance costs are comparable with an atomic
absorption system.
Instrumentation
Spectrometer. The Echelle grating
spectrometer has been described by several
workers.1' 2' 3> h > 5 The spectrometer
has a 0.75m focal length with a moveable
echelle grating for horizontal dispersion
and a quartz prism for vertical order separa-
tion. The combination of prism and grating
creates a two dimensional raster-like pattern
with each line equivalent to a segment of a
high resolution spectrum with the orders
vertically stacked. The entrance and exit
slits are mounted on slides for easy adjust-
ment with 20 combinations for the entrance
and 16 combinations for the exit. These
combinations allow for the best theoretical
selection of entrance and exit slits for op-
timum performance of the monochromator. A
front panel scroll displays the orders of re-
flection and allows for simplified wave length
selection. The resolution depends on wave
length and slit width and is 0.01SA at 2000A
with a 25 micron slit. The spectrometer it-
self requires bench space about three by 2.5
feet.
Argon Plasma Jet
The characteristics of the D.C. argon
plasma are well documented 6> 7> 8> 9 and
has been developed as part of a high resolu-
tion spectroscopic analytical system. The
user is provided with a 6000 to 7000°K argon
plasma that is convenient to use, very effic-
ient ind virtually free of chemical inter-
ferences found with conventional flames. The
abundance of energy available from the plasma
enables the direct analysis of non-metals
such as silicon, boron, phosphorus and carbon
(Figure 1) as well as refractory type metals.
Using the ionized spectral lines available
with the plasma often results in a greater
linear range (Figure 1).
Spectra Jet Module
This includes the plasma assembly, the
sample introduction system, and the power
supply. The sample introduction system con-
tains the spray chamber, nebulizer, alignment
controls, protective housing for the plasma,
and a collecting lens. The power supply has
all the power regulating circuitry as well
as current control meter and argon flow
meters.
Readout
In order to obtain analytical data ana-
log and digital readout modules are required.
This combination of components completes the
system for quantitative analysis. The spec-
trometer can also be equipped with a camera
for qualitative spectrographic analysis.
Multi-element Analysis
This feature is available for routine
repetitive analysis where the sample com-
position does not change and the same ele-
ments are to be determined each time.
Applications
To date I have analyzed 34 elements in
our laboratory, specifically Li, Na, K, Cs,
Be, Mg, Ca, Sr, Ba, Y, V, Cr, Mn, Fe, Co, Ni,
Cu, Ag, Au, Zn, Cd, Hg, B, Al, Ga, C, Si, Sn,
Pb, P, As, Sb, Bi, and Se. From communica-
tions with other researchers in the field,
approximately 70 elements have been analyzed
using D.C. Argon plasma atomic emission.
In addition to this instrument's versa-
tility by virtue of the many elements easily
analyzed, the linear dynamic range far exceeds
that of other common instrumental techniques.
This allows greater flexibility in quantita-
tive determinations of widely ranging element
concentrations, reduces analysis time. This
minimizes errors due to dilution or pre-
concentration further enhancing the technique's
accuracy. The analyst is not required to
maintain a complete working stock of element
lamps thus reducing operating costs. The only
limit to the full instrument capability is
the availability of standard solutions.
Several commonly analyzed elements are
presented (Figure 2) to illustrate D.C. argon
plasma atomic emission's versatility. The
analytical techniques for the elements in
Figure 2 follow.
Mercury
The commonly used atomic absorption cold
vapor technique 1c» ll» 12 was adapted to
the atomic emission system. It was necessary
to modify the sample introduction system by
closing the sample tube and introducing the
vapor directly into the large nebulizer ven-
turi which leads directly to the plasma. An
advantage of this system is that water remov-
al is not required which eliminates the need
for a drying tube and its potential hindrance
of mercury vapor flow. In our studies we have
measured mercury concentrations in sea water
4-6
-------
as low as 0.15ppb which should not be con-
sidered the ultimate sensitivity of the in-
strument. Calculated limit of detection is
approximately 57yg/l. Because of the high
resolution of the echelle grating spectro-
meter, (0.015A at 2000A with a 25 micron slit)
only the mercury signal is measured virtual-
ly eliminating matrix effects. For example,
chlorine interference does not affect mercury
measurements when an echelle grating spectro-
meter is used.
Lead
Direct measurement of lead in sea water,
by standard addition, is illustrated in Fig:
ure 2. No extraction or preconcentration was
necessary and concentrations as low as 315ppb
were measured. The calculated detection limit
in the sea water matrix was 35vig/l. Lead is
generally one of the least sensitive elements
to be measured by spectroscopic techniques.
D.C. Argon plasma provides a much improved
analysis technique for this element. For
example, I have measured a lead concentration
as low as 35jig/l in distilled water to docu-
ment lead contamination.
Cadmium and Zinc
Direct analysis of cadmium and zinc in
sea water were performed using the method
of standard additions. Figure 2 shows a cali-
bration curve for these metals. The calcu-
lated detection limit for cadmium was 12ug/l
and zinc was 12ug/l. Copper and chromium were
also determined and the detection limits were
6yg/l and 8pg/l respectively. The wide dyna-
mic range for zinc is a real advantage over
conventional atomic absorption whose linear
response does not exceed 2mg/l, so that less
dilutions are required for samples with high
zinc concentrations.
Discussion
This technique provides instrumental
analysis capability for some elements where
no other method currently exists. Analysis
for total silicon is such an example. This
system also allows analysis in matrices which
do not lend themselves to more conventional
methodologies. An example would be the analy-
sis of calcium in various phosphate concentra-
tions. Conventional atomic absorption rapid-
ly loses sensitivity in a phosphate solution11
and lanthanum must be added to overcome
this interference. On the other hand a
10,000 fold excess of phosphate to calcium
can be tolerated with D.C. argon plasma atom-
ic emission before any interference is notic-
ed.
Another feature of atomic emission is
the potential to use several emission lines.
Should a sample be too concentrated, the
analyst can select a less sensitive emission
line without the necessity of sample dilution
saving both time and reducing potential dilu-
tion errors. Should a spectral interference
from closely emitting elements be noted at
one wave length the operator simply selects
an interference free wave length.
The acquisition cost of the instrument
is comparable with that of conventional atom-
ic absorption, however, no hollow cathode
2
lamps are needed and only a single fuel
(non-explosive) is used. When these factors
are considered the operation and maintenance
is somewhat less than atomic absorption.
Analysis time is also less since lengthly
digestion procedures are not needed for the
majority of samples. Chemical pretreatment
is generally only needed in cases where
cold vapor techniques are employed such as
generation of arsine or mercury vapor. In
most other cases the plasma does the work for
the analyst when digestion is called for.
The spectrometer can be equipped with
a photographic attachment which allows a
rapid qualitative identification of sample
components. To convert to a qualitative
mode, a polaroid press camera is inserted
into the photographic attachment and extend-
ed the bellows to move the mirror for photo-
graphy into the light path. If photographs
are taken with type 55 P/N ASA 50 film, a
permanent record in the form of a 4x5" posi-
tive and a negative result. The negative can
be enlarged without appreciable loss of in-
formation .
Conclusions
The commercial development of the high
resolution echelle spectrometer in conjunc-
tion with the D.C. Argon plasma has brought
about a revival of optical emission spec-
troscopy which for many years has been over-
shadowed by the popularity of atomic absorp-
tion. It is now recognized that emission is
more sensitive for a larger number of elements
and use of the D.C. argon plasma provides
benefits such as a greater linear dynamic
range.
This technique has applications to a
wide variety of environmental samples of
unknown or varying composition.
We have analyzed for trace elements in
storm drain outfalls, sewage, sediments, soil,
animal tissue, oil, plastics, sea water, dis-
tilled water, brakish water, waste from
plating baths, trace metal impurities in alloys
and total phosphorus in storm drains. The
methods could easily be extended into the
field of air pollution monitoring for heavy
metals. In many cases prior chemical treat-
ment is not needed and the total analysis
time is considerably shortened. The initial,
operating and maintenance costs are equal
to or less than other instrumental techniques
providing equivalent results.
The echelle spectrometer with the D.C.
argon plasma combines high resolution, high
luminosity, and wide spectral coverage in a
bench top design and created new horizons for
the analyst in the field of trace metal analy-
sis .
-------
References
'M.S. Cresser, P.N. Keliher, and C.C.
Wohlers, anal, chem., £5, 111 (1973).
2George J. Matz, American Laboratory,
March 1973.
3P.N. Keliher and C.C. Wohlers, anal,
chem., 4^, 682 (1974) .
"•M.S. Cresser, P.N. Keliher and C.C.
Wohlers, Laboratory Practice, 2_6, 335 (1975).
SP.N. Keliher and C.C. Wohlers, Appliod
Spectroscopy 2_9, 198 (1975) .
'William G. Elliott, American Laboratory,
August 19 71.
7William G. Elliott, Presented at 1974
Pittsburgh Conference.
6P. Merchant, Jr., and C. Veillon,
anal, chem. acta., 70_, 1.7-24 (1974).
9W.E. Rippetoe, E.R. Johnson, and T.J.
Vickers, anal, chem., 47^, 436 (1975).
1""Methods for Chemical Anlaysis of
Water and Waste", EPA-625-/6-74-003.
llJ.E. Hawley and J.D. Ingle, Jr., anal,
chem., 42, 719 (1975) .
12W.F. Fitzgerald, W.B. Lyons, C.D. Hunt,
anal, chem., £6, 1882 (1974).
¦II
3 4-6
J
-------
10000,
E Zinc 2025.51 A.U.
io<5o
100
10000
= Cadmium 2144,38 A.U.
1000
100
ml l 1. I I I I III I—I I I
0.01
0.10
I I Hill lOl I i I I i I ill i I I I I
10000
1000
DO
10 10 0.01 QiO
CONCENTRATION (mg/l)
I0000^_
1.0
I I I
10
; Lead 2833.07 A.U.
- Mercury 2536.52 A.U.
CO
h Z
LxJ
1000
100
inl i i i i 11 ill I in
0.01
0.10
1.0
I M I Hill lOl i I I I Mill
10 0.01
0.10
,i i mini i li
IjO
10
Figure 2 - Calibration curves for zinc, cadmium,
lead and mercury in sea water.
4
4-6
-------
AN EVALUATION OF AUTOMATIC WASTEWATER COMPOSITORS
Daniel J. Harris And William J. Keffer
U.S. Environmental Protection Agency, Region VII
Surveillance And Analysis Division
25 Funston Road
Kansas City, Kansas 66115
Summary
The apparent characteristics of raw municipal
wastewater were found to be highly dependent upon
choice of sampling equipment and technique. The phi-
losophy and field compliance monitorinq procedures
whigh the Surveillance and Analysis Division has
adopted as a result of equipment evaluation studies
and some 150,000 hours of operational experience are
discussed.
Introduction
The Environmental Protection Agency, Region VII,
Water Section is responsible for planning and conduct-
ing the sample collection activities of the Surveil-
lance and Analysis Division, which provides the water
quality data needs of the agency for the four-state
region of Iowa, Kansas, Missouri, and Nebraska.
During the past three years the section has experi-
enced a dramatic increase in the demand for compliance
monitoring data as a result of the National Pollutant
Discharge Elimination System (NPDES) permit program.
Because of this demand, the section has curtailed
manual methods of sample collection and relied upon
commercially available automatic wastewater sampling
equipment.
The section has accumulated approximately 150,000
hours of field experience with fifty compositors of
sixteen makes and models from nine manufacturers. As
wastewater chemistry information accumulated from the
utilization of these compositors, discrepancies in
data were noted which appeared to result from varia-
tions in equipment performance.
This paper summarizes the results of four
studies1 which were conducted by the Water Section to
evaluate these compositors. Also discussed are the
philosophy and field sampling methodology which have
been adopted by the section as a result of experience
and these evaluations.
Compositor Comparison Studies
Table I presents an inventory of the commercially
available wastewater samplers which the section has on
hand or has gained some experience with through the
courtesy of the manufacturer. The names and addresses
of these manufacturers can be found in the appendix.
Table I also shows the inside diameter of the sampler
intake line and the linear velocity of the wastewater
in this line during a sampling cycle.
TABLE I
INVENTORY OF AUTOMATIC WASTEWATER SAMPLERS
Sampler
Intake
Tube
ID
mm^
Liquid
Intake
Velocity
cm/sec(k)
Sigmamotor WA-2
3.17
7.9
Sigmamotor WD-2
3.17
7.9
Brailsford EV-1
4.76
0.45
Brailsford DU-1
4.76
0.45
Brailsford EP-1
4.76
0.45
Hants Mark 3B
6.35
75(d)
ISCO 1391-X
6.35
21
ISCO 1392
6.35
61
ISCO 1580
6.35
61
Sirco MKVS7
9.5?
98
Manning S-4000
9.52
189^
Pro-Tech CG-125P
3.17
207
QCEC CVE
6.35
61-152
N-Con Scout
, 6.35
7.6
N-Con Surveyor
12.70
36
N-Con Sentinel^
NA
Variable
(a) Multiply by 0.0"?P4 to obtain inches
(b) Multiply by 0.0328 to obtain fps
(c) Loaned courtesy of manufacturer
(d) Mean
(e) Maximum
Table II summarizes the nonfilterable solids
(NFS) data from four comparison studies in which
time and flow-proportioned samples, collected concur-
rently with some of the samplers listed in Table I,
were compared to NFS data resultinq from samples col-
lected simultaneously by manual methods. Althouqh
these studies also examined 5-day biochemical oxyqen
demand (BOD5) and chemical oxyqen demand (COD), this
discussion is limited to the NFS data since experience
4-7
-------
has indicated this parameter to be one of the most
difficult to sample. The data in Table II is a sum-
mary of the NFS ratios resulting from the daily mean
NFS concentrations from studies which lasted from one
to four days.
TABLE II
RATIOS OF NONFILTERABLE SOLIDS
Sampling Method/Sampling Method
Ratio
Richards-Gebaur STP
(24-hour composites - 3-day means)
Raw Waste
ISCO Flow Comp/Manual Flow Comp
0.94
QCEC Time Comp/Manual Flow Comp
1.69
Manual Time*Comp/Manual Flow Comp
0.92
Primary Effluent
Hants Flow Comp/Manual Flow Comp
2.04
Slgmamotor Time Comp/Manual Flow Comp
0.84
Manual Time Comp/Manual Flow Comp
1.06
Secondary Effluent
Hants Flow Comp/Manual Flow Comp
1.22
Brailsford Time Comp/Manual Flow Comp
0.96
Manual Time Comp/Manual Flow Comp
1.02
Lincoln. Nebraska STP
(24-hour composites - 3-day means)
Raw Waste
QCEC Time Comp/ISCO Time Comp
l.«2
Secondary Effluent
Brailsford Time Comp/ISCO Time Comp
0.92
Ashland, Nebraska
(24-hour composites)
Raw Waste (4-day mean)
Hants Flow Comp/ISCO Flow Comp
2.72
Secondary Effluent
Hants Flow Comp/ISCO Flow Comp
2.20
Kansas C1tv. Kansas. STP
Raw Waste (20-hour mean)
QCEC Time Comp/ISCO Time Comp
1.99
Sirco Time Comp/ISCO Time Comp
1.24
Richards-Gebaur Air Force Base Study
The first study was conducted at the trickling
filter treatment plant on the Richards-Gebaur Air
Force Base. The three sampling stations at this
treatment plant were: (a) the raw waste, (b) the
effluent from the primary clarifier, and (c) the final
effluent.
A QCEC Model CVE sampler was installed to collect
time-composite samples (15-minute cycle time) at the
influent and concurrently, an ISCO Model 1391-X was
used to collect discrete samples at 2-hour intervals
for manual flow proportioning and compositinq. A
Slgmamotor Model WD-2 (15-minute cycle time) and a
Hants Mark 3B (2-hour cycle time) were used at the
primary effluent to collect, respectively, time and
flow-composite samples. The secondary effluent was
sampled with a Brailsford DU-1 (4-minute cycle time)
which took time-composite samples and with a Hants
Mark 3B (2-hour cycle time) for manual flow compos-
iting. At each of the three stations grab samples
were manually collected at 4-hour intervals for dis-
crete analysis and for flow compositing in order to
provide additional data for comparison.
An examination of the AFB data 1n Table II would
indicate that there were significant variations in the
NFS concentrations of the samples collected by differ-
ent equipment and manual methods. Three things are
apparent from the AFB data: (a) the variation of the
ratios from unity generally decreased as the waste-
water passed through the treatment plant, (b) the
samples collected with the Hants and QCEC samplers,
whether time or flow composited, were hiaher than the
manually flow-composited samples, and (c) the time-
composited QCEC sample on the raw waste was higher
than the manual, flow-composited sample.
Attention 1s called to the last observation.
The grab samples collected at 4-hour intervals from
the raw waste (data not presented) indicated a defi-
nite decrease in strength during the early morning
hours. Since the QCEC time-composite sample included
portions of the low-flow, low-strength waste which
were equal in volume to the aliquots collected during
daylight hours, the resulting data should have, in
theory, been biased low in relation to the manual,
flow-composite data. However, it can be seen that
the resulting QCEC/manual flow ratio was 1.69.
Standard Methods2 reports the precision of the
NFS parameter resulting from interlaboratory analyti-
cal quality control samples. Based upon this infor-
mation, 16 of the 27 (59 percent) NFS analyses (data
not presented) of the samples collected from the
Richards-Gebaur study were outside the results (± one
standard deviation) obtained with the manual flow-
proportioning technique.
Table III shows the NFS removal efficiencies of
the Richards-Gebaur sewage treatment plant using the
various sixteen combinations of sampling techniques.
It can be seen, depending upon the sampling combina-
tion, that the apparent treatment efficiency varied
from a low of 36 percent to a high of 72 percent.
Comparing similar sampling methods, the removal effi-
ciency based on the manual flow composites was 52 per-
cent, while the time-composite data from the OCEC-
Brailsford combination indicated a removal efficiency
of 72 percent.
2
4-7
-------
TABLE III
Kansas City, Kansas
APPARENT REMOVAL EFFICIENCIES OF
RICHARDS-GEBAUR FACILITY WITH VARIOUS COMBINATIONS
OF 24-HOUR SAMPLING METHODS
Lincoln, Nebraska
A second comparison test was conducted at the
municipal trickling filter plant in Lincoln, Nebraska.
The raw wastewater was time composited, concurrently,
with an ISCO 780 (1-hour cycle time) and a QCEC-CVE
(14-minute cycle time). At the final effluent an ISCO
780 (1-hour cycle time) and a Brailsford DU-1 (6-
minute cycle time) were used to collect time-composite
samples. An examination of Table II would indicate
that the resulting mean ratios of three days of sam-
pling with the QCEC/ISCO and Brailsford/ISCO combina-
tion were 1.82 and 0.92, respectively.
Ashland, Nebraska, Sewage Treatment Plant
A third comparison study was conducted at the
Ashland, Nebraska, municipal treatment plant. At the
influent and effluent an ISCO 1391 and Hants Mark 3B
were paired and set to simultaneously take discrete
samples at 2-hour intervals for manual flow propor-
tioning and compositing. The mean Hants/ISCO ratios
resulting from four days of sampling on the influent
and effluent were 2.72 and 2.20, respectively.
The last comparison study was conducted at the
Kaw Point Sewage Treatement Plant in Kansas City,
Kansas. The raw waste was sampled for a period of
about 20 hours using a QCFC-CVE, Sirco MKVS7, and an
ISCO 1391. Each unit took equal volume aliquots at
cycle times which were 15, 40, and 60 minutes, respec-
tively. The resulting QCEC/ISCO and Sirco/ISCO ratios
were 1.99 and 1.24 respectively.
Pi scussion
In every case, the comparison studies on the raw
municipal waste indicated variations in water chemis-
try data which were greater than could be explained by
laboratory analytical error.
An examination of Table I would indicate that
those compositors which produced the highest concen-
trations of solids had higher liquid intake velocities.
The QCEC, Hants, Sirco, and ISCO 1391-X samplers had
sampling velocities of 61 to 152, 75, 98, and 21 cm/
sec, respectively. It would appear that the slower
acting peristaltic and piston pump type samplers were
either not capturing settleable materials or that
after introduction to the intake line settling veloci-
ties were higher than the liquid intake velocities.
It is also possible that the QCEC sampler may have
biased solids levels because of the larger number of
small-volume aliquots (approximately 25 ml cycle)
which it takes over a 24-hour period. The sampler may
have reduced the pressure on the incominq liquid to
the point that dissolved gases were passing out of
solution and entraininq solids as the gas bubbles rose
to the surface.
To the credit of the professional chemist, labo-
ratory analytical error has been reduced to the point
of insignificance when one considers the errors which
can be introduced by improper sample collection tech-
niques. Unfortunately, sampling methodology has not
received the same professional attention.
The implications of the data variability indi-
cated by the comparison studies are apparent. It is
obvious that those individuals responsible for surveys
and sample collection activities can use any of the
generally-accepted sampling techniques and equipment
and still intentionally or unintentionally manipulate
apparent wastewater chemistry characteristics and
facility removal efficiencies.
It is little wonder that there are so many dis-
agreements among various responsible Federal, state,
city, and individual groups regarding water chemistry
characteristics and facility performance. When varia-
tions in sampling methodology and laboratory sys-
tematic and random errors are further compounded by
errors in flow measurements, differences can become
astronomical. Without an adequate monitoring program
and tight controls on samplinq techniques, equipment,
and laboratory procedures, data interpretation can be
reduced to little more than an exercise in futility.
Water Section Philosophy and Sampling Methodology
Although the results of the sampler comparison
studies are not conclusive and additional work is
needed, it is the opinion of the Water Section that
high-vacuum, sampling equipment produces more repre-
sentative samples. On waste sources with appreciable
concentrations of large and/or heavy settleable mate-
rial such as a raw municipal wastewater, the section
makes every effort to install a high vacuum unit when
Sampling Method Combination
% Removal
Influent
Effluent
NFS
Manual Flow Comp
52
Manual Flow Comp
Hants Flow Comp
Mean of Manual Grabs
41
51
Brailsford - Time Comp
54
Manual Flow Comp
49
ISCO Flow Comp
Hants Flow Comp
Mean of Manual Grabs
37
4B
Brailsford - Time Comp
50
Manual Flow Comp
48
Mean of Manual Grabs
Hants Flow Comp
Mean of Manual Grabs
36
47
Brailsford - Time Comp
50
Manual Flow Comp
71
QCEC Time Comp
Hants Flow Comp
Mean of Manual Grabs
65
71
Brailsford - Time Comp
72
Mean and Coefficient
of Variation
Mean, mg/1
53
Coefficient
Variation, %
21
3
4-7
-------
compatible with site conditions and data requirements.
Of the present samplers on hand, the section prefers
the QCEC-CVE, Manning S-4000, and the Sirco MKVS7
compositors for sampling these wastewaters. Because
of equipment shortages in the field, the section also
occasionally uses the ISCO 1392 and 1580 on raw wastes.
The overall ability of the section to obtain a
complete 24-hour composite sample is about 80 percent.
The major cause of sampler failure, in addition to
human factors, has been plugging of sampler intake
lines on raw wastes. Experience has indicated that
plugging of intake lines occurs more frequently on
those units with small diameter intake lines, low
intake velocities, and no purge cycle. Consequently,
the section uses the Sigmamotor, Brailsford, N-Con,
and ISCO units (Table I) primarily on treated waste-
waters with low levels of suspended material. Although
high-vacuum samplers can be effectively used on these
wastes, the comparison studies would indicate that
well-treated effluents with no visible solids can be
representatively sampled with the lower intake velocity
compositors.
Wittr present sampling technology, the section is
of the opinion that flow compositing of raw municipal
wastewaters and other wastes with appreciable settle-
able sol Ids is neither necessary nor justified. The
variations in sampler performance and manual sampling
techniques completely mask actual changes in wastewater
chemistry characteristics. The magnitude of the data
discrepancies indicated by the comparison studies does
not warrant the extra time and expense involved in
installing sophisticated sampling equipment and flow-
measurement devices.
The comparison studies on treated wastes would
indicate that wel 1-treated, sparkling effluents with no
visible solids are amenable to flow-proportional sam-
pling and that a suitable compositor can be selected
without regard to variations in performance.
Flow-proportional sampling currently accounts for
less than 5 percent of the sample collection activi-
ties of the section and is limited to providing data
for enforcement activities and for sampling of indus-
trial dischargers which have extreme changes in
process water volume and characteristics as a result
of batch processing.
Because the majority of the permits which are
issued are based upon average maximum weekly and
monthly effluent limitations, the Water Section is
routinely conducting five and six-day compliance
monitoring investigations of municipal and industrial
dischargers. These monitoring periods, which include
weekends, are felt to produce the minimum amount of
data needed for determining the compliance or noncom-
pliance of a discharger.
The data variability indicated by the comparison
studies warrants a professional in the field. At new
locations which have not been previously sampled it is
a policy of the Water Section to have a professional
present to select the sampling point, to inspect the
flow-measurement equipment of the facility or deter-
mine a suitable measurement method, and to supervise
installation of the sampling equipment. It 1s felt
that this practice reduces the risk of compositor mal-
function and missed samples, improves the representa-
tiveness of the data, and results fn a more detailed
and Informative report.
The practice of using inexperienced, unsupervised
personnel to collect samples for analysis by highly-
paid professional chemists is a misappropriation of
technical and economic resources which can only result
in unrepresentative data.
Conclusions
1. Currently available sampling equipment cannot
be relied upon to produce representative
samples.
2. High vacuum samplers produce more representa-
tive samples and should be used on raw
municipal wastewaters and other wastes with
significant levels of large heavy suspended
material.
3. Any sampler compatible with site conditions
and data requirements can be used to sample
well-treated effluents with no visible
sol ids.
4. Flow-proportional samplinq of raw municipal
wastewaters with currently available samplinq
equipment is neither necessary nor justified.
5. Current sampling equipment and methodoloqies
need to be refined to improve data repro-
ducibility and accuracy.
6. Apparent wastewater chemistry characteristics
and facility removal efficiencies can easily
be manipulated by choice of sampling equip-
ment and methodology.
7. Experienced professional personnel are
required in the field to obtain representa-
tive samples of wastewater.
Acknowledgement
The EPA, Region VII, Water Section wishes to
acknowledge the cooperation of the Cities of Kansas
City, Kansas; Lincoln, Nebraska; and Ashland,
Nebraska, and Richards-Gebaur Air Force Base in allow-
ing the section to conduct sampler comparison studies
at their wastewater treatment facilities.
Disclaimer
Mention of brand name of equipment does not con-
stitute endorsement or recommendation of product by
the Environmental Protection Agency. The information
and findings presented in this paper are not to be
construed as representing official equloment design or
modification specifications.
References
1. Harris, D. J. and Keffer, W. J., "Wastewater
Sampling Methodologies and Flow Measurement
Techniques," U.S. Environmental Protection
Agency, Region VII, Kansas City, Missouri (1974).
2. "Standard Methods for the Examination of Water and
Wastewater," 13th Ed., Amer. Pub. Health Assn.,
New York, New York (1971).
Appendix
Names and Addresses of Sampler Manufacturers
Sigmamotor Model WA-2 and WO-2
Siqmamotor, Inc.
14 Elizabeth Street
Mlddleport, New York 14105
4-7
-------
Brail sford Model EV-1, DU-1, and EP-1
Brailsford and Company
Milton Road
Rye, New York 10580
Hants Mark 3B
Testing Machines
400 Bayview Avenue
Amltyville, New York 11701
I SCO Model 1391 and 1392
Instrumentation Specialties Company
P. 0. Box 5347
Lincoln, Nebraska 68505
Sirco MKVS7
Sirco Controls Company
401 Second Avenue West
Seattle, Washington 98119
Pro-Tech C6-1Z5P
Pro-Tech, Inc.
Roberts Lane
Malvern, Pennsylvania 19355
QCEC Model CVE
Quality Control Equipment Company
2505 McKinley Avenue
Des Moines, Iowa 50315
N-Con Scout, Surveyor, and Sentinel
N-Con Systems Company, Inc.
Clean Waters Building
New Rochelle, New York 10801
Manning S-4000
Manning Environmental Corporation
120 8uBo1s Street
P. 0. Box 1356
Santa Cruz, California 95061
4-7
-------
MICROSCOPICAL IDENTIFICATION OF ATMOSPHERIC PARTICLES
Walter C. McCrone
McCrone Associates, Inc.
Chicago, IL 60601
The principal use of the light microscope is to
represent research, and indeed science, in company
brochures, annual reports and sales literature. The
light microscope, especially the polarizing microscope,
is seldom used and almost never at full potential. It
is, however, in trained hands, an extremely powerful
tool and one which can quickly solve many important
problems. It is, for example, uniquely qualified for
the identification of individual atmospheric dust
particles. Its uniqueness for that application is
based on the ability of a trained microscopist to
identify individual particles in terms of the actual
compound present rather than the constituent elements.
Not only are the particles identified by composition
but also often by source.
There are many substances with almost identical
chemical formulas but with very different origins.
Silica is an outstanding example, the microscope will
not only tell us that we have silica (Si02) but it will
differentiate between each of the diverse forms of
silica. These include diatoms, radiolarla, sponge
spicules, silica cells from plants, vitreous silica,
silica gel, quartz, sea sand, quartz desert sand,
cristobalite, tridymite and other, usually less pure,
forms of silica such as chalcedony, cryptocrystalline
quartz, opal, flint and amethyst. There are many
similar examples, e.g.. cellulose, which occurs in
thousands of forms, usually fibrous, but also as starch
grains; each can be identified as such and as a part of
a particular plant. Flyaeh can not only be identified
in terms of the fuel originally burned but often the
burning conditions can be very closely specified. The
process by which the particle was formed is often
readily apparent, e.g., is a given particle of zinc
oxide, titanium dioxide or any other pigment, a ground
product or precipitated? The shape of the particle
will nearly always answer such questions.
Why then isn't the polarizing microscope more
often used for the characterization of atmospheric
dust. There are two reasons! First, Relatively few
analysts are aware of the potential of this modestly
priced tool and second, if aware, few analysts are
willing to spend the time necessary to learn to apply
it. McCrone Associates' personnel have done everything
possible to correct this situation. We present papers
like this one, write books like The Particle Atlas,
teach courses and develop new analytical procedures for
the characterization and identification of small
particles. Although the application of these methods
in the real world is our bread-and-butter we have no
secrets. We are very willing to teach anyone the
methods we use. We would like to see more general
application of the light microscope. There are plenty
of particles to keep us all busy.
What then do we do when we have a sample of
atmospheric particles? We first mount the sample lev a
liquid having known optical properties. We usually use
one of the Aroclors since they have refractive Indices
in a useful range and since they are viscous enough to
disperse the particles so they can be individually
studied. We then examine the sample with a polarising
microscope, but with low magnifications, usually about
100X. We very seldom exceed a total magnification of
400X unless the particles are very small (<1 pm). In
any case, we look at the particles and try to identify
them by recognition of shapes, sizes and colors that
we've seen before. If we can remember what it was the
last time we saw it then we know what it is now. The
procedure is no different than the recognition of
individual people that we see macroecopically. If
you've seen them often enough and if they have once
been identified then there is usually no difficulty in
remembering who they are. We do, of course, have prob-
lems when it's a particle that we only see every other
year or so and, of course, some of us are better at
"remembering faces than names". Obviously, the more
time one spends looking at known and unknown samples
the better he will be at this kind of identification.
It is a fact that with practice a microscopist can
expect to be able to identify at sight literally
thousands of individual substances.
When one's memory-bank fails it is then time to
go to more sophisticated procedures like micro x-ray
diffraction, electron microscopes and microprobes. You
can imagine the escalation in cost, however, when this
step becomes necessary. It is very advantageous
economically to be able to identify particles by light
microscopy. A good light microscopist assumes that the
sophisticated tools don't even exist and he makes every
effort to identify the particles in his sample without
leaving his light microscope.
His microscope, incidentally, with all the
necessary accessories such as refractive index media,
filters, light sources, camera, dispersion staining
objective and refractive index liquids may cost under
$2,000. Good microscopy requires a good microscopist,
not an expensive microscope. This is not to say that a
more expensive microscope would not do the job, but it
certainly doesn't do the job proportionately better
than the less expensive microscope. I would not
hesitate to say that a good microscopist can do just as
good a job of identifying particles with a $1,000
polarizing microscope as he can with the $25,000-50,000
variety. The lower priced model would also be far
easier to use; in fact, only a highly trained
microscopist should be turned loose on one of the
expensive instruments. My own view is that every
microscopist .should do his apprenticeship on the
$500-1,000 microscope and graduate to the more
expensive model as he becomes more fully appreciative
of its flexibility and versatility. The more expensive
microscope becomes a fringe benefit, as it were, after
a year or two of "faithful service" and significant
accomplishment.
What then do we look for when we want to identify
a small particle? First, are the three S's: size,
shape and surface. The size of ground materials may
not be very significant, in general, but size of
fibers, spores, pollens etc. are reliable identifying
chracteristics. Corn and rice starch, e.g., have
similar shapes but rice starch averages only half as
large. Sand grains are large rounded granules, whereas
pigments are always very much smaller and flyash can
cover the entire microscopical range of size. The
shape of particles is one of the prime Identifying
characteristics often by Itself, enabling certain
identification, e.g., paper fibers, diatoms, balsam
poplar seed hairs, oil soot and fluorlte. Surface
markings are also very useful in identifying, e.g..
foundry sand, bast fibers, coniferous wood pulp and
magnetite in pulverized coal-fired boiler flyash.
Next perhaps, is color both by reflected and
transmitted light. The identification of human hair is
1
5-1
-------
based usually ori shape, size and surface but further
characterization by race is based on color. The colors
of paint pigments such as rutile, cinnabar, azurite,
malachite etc. are very essential identifying
characteristics for those materials. By itself, color
may not be sufficient but without it identification
would be much more difficult.
Transparency is another useful characteristic.
Many small particles are completely opaque, e.g.,
magnetite, metal particles, oil soot and graphite; some
are translucent, e.g., glassy flyash particles; most
are transparent, e.g., most minerals, fibers, abra-
sives, pigments etc. Since they are transparent one
can quickly estimate their refractive index relative to
the liquid in which they are mounted (usually Aroclor
with an index around 1.65-1.66).
The Becke line quickly tells whether the particle
has an index greater or less than the liquid and the
degree of contrast indicates how much higher or lower.
Many substances, e.g., fluorite, Teflon®, diatoms and
cellulose acetate have much lower refractive indices
than Aroclor. Many, e.g., white lead, zinc oxide,
rutile and carborundum have very much higher indices.
By far, most of the world's small particles have
indices less than Aroclor but some of these will have
indices close to Aroclor. Calcite (limestone) has one
refractive index almost exactly equal to Aroclor,
calcium phosphate, apatite, has all of its indices very
slightly less than Aroclor. Amosite asbestos and most
of the components in Portland cement have indices
slightly higher than Aroclor.
Next, one crosses the polars and observes first
whether the particle is isotropic, e.g., glass, salt
and diatoms, or whether it's anisotropic, e.g., 95% of
the particles in the world. The degree of anisotropy
(birefringence) should also be noted since some, like
quartz, cellulose acetate and apatite have low bire-
fringence whereas Dacron®, nylon, limestone and rutile
have,very high birefringence.
All of the characteristics so far noted require
very little time, perhaps less than a minute and, at
this stage, most particles should have been recognized.
If not identified at this point it may be that the
microscopist needs some suggestions as to the possible
substances and this is best done by first classifying
the particle based on the characteristics already
measured; to do this he notes in the following order
six parameters: transparency, color, birefringence,
index relative to Aroclor, flatness and elongation. A
six-digit number made up of ones or zeros, depending
upon the findings (positive or negative) in each of the
six catagories characterizes that particular particle
type. He then refers to Volume IV of
The Particle Atlas which lists a large number of common
substances having those same characteristics. Glancing
down this list he may be reminded of some substance he
should have thought of and thereby identify the
particle. The chances of identifying the particle at
this stage is experience-dependent. If he Is still
unable to identify the particle there are still
additional microscopical procedures he can use. Often
he will feel reasonably certain that the particle is
one of two or three possibilities and he must then
think of some additional test which will differentiate
between those possibilities. The test he chooses will
depend on the choice he has to make. He may use a
small magnet to differentiate between graphite and
magnetite, he does this by warming the preparation so
that the Aroclor is quite fluid and then (while he
watches the particle through the microscope) he moves a
small magnet in an arc around the microscope objective
close to the preparation. If the black particle is
magnetite it will, of course, behave like a compass
needle whereas graphite will respond in no way.
Another simple procedure which differentiates between
many possible substances is the so-called "squoosh"
test. The preparation is again warmed to soften the
Aroclor and then (while watching the particle through
the microscope) a needle is used to depress the
coversllp into contact with the particle. Pressure
applied at this point will cause the particle to deform
if it is soft or fracture if it is hard. Its behavior
under these circumstances is a good identifying
characteristic. An elastomer, e.g., will not only
flatten as pressure is applied but will recover its
original shape when pressure is released. Other soft
materials, e.g., polyvinyl polymers, will squeeze into
flat layers and remain thus on release of pressure.
Glassy spheres will crack, showing concholdal fracture,
and many minerals will cleave along specific di-
rections .
Refractive indices are among the most useful
identifying characteristics and usually justify
spending a few minutes to measure. If the indices are
only slightly lower than Aroclor the preparation can be
heated in a hot stage to the temperature at which the
particle and the Aroclor have the same index. One can
then calculate from the known temperature coefficient
of index for Aroclor what the refractive index of the
particle was at room temperature. If necessary, he can
remove the particles from the preparation by warming to
soften the Aroclor, then sliding off the coverslip and
picking out the desired particle with a needle. The
particle is then transferred to a clean slide and
washed clean before mounting in a second liquid.
Usually, however, there is sufficient sample so that a
second aliquot can be placed in the second liquid. It
is usually possible to guess the desired index for the
second liquid so that a near match is obtained and
either dispersion staining or the hot stage will give
the desired refractive index.
Another useful technique for particle
characterization is crystal rolling; this is done with
either Aroclor 1260 as the mounting fluid or by warming
Aroclor 5442 to soften it sufficiently so that sliding
the coverslip with a needle will cause the particle to
roll. In this way, different views of a crystal can be
obtained and particular views such as those looking
down on optic axis can be rolled into position. This
can be done either watching the crystal itself if it
has a recognizable crystallographic shape or it can be
done, between crossed polars, by rolling the crystal
into a position of low birefringence or it can even be
done while viewing the interference figure itself and
rolling until an optic axis lies in the field of view.
The interference figure obtained in this way is a very
useful identifying characteristic and will often
suggest or confirm identity of many minerals and other
crystalline substances. In conjunction with shape and
refractive indices it seldom fails to identify such
substances.
There are other supplementary tests that can be
used, e.g.. solubility, which is measured on individual
subnanogram particles by a procedure originally sug-
gested by Laskowski. He places the particle on the
underside of a coverslip over a small cell on the
microscope stage and then Introduces a drop of any sol-
vent and watches the particle to see if it picks up
vapor of that solvent to form a liquid droplet in which
it then dissolves. The test is rapid and the particle
is easily regenerated by removing the coverslip and
exposing it to the air for a few seconds. Another test
is density. The particles are mounted in, say, meth-
ylene iodide between slide and coverslip. Particles
having densities higher than 3.3 will sink to the
2
5-1
-------
bottom of the prep and those with lower densities will
rise to rest against the coverslip.
Finally, there are microchemical tests which can
be applied to single particles by a procedure published
by McCrone. There are quick precipitation reactions
for any of the cations or anions and even for many of
the common organic compounds.
A trained microscopist using these tests will be
able to identify a large proportion of the particles
that he encounters; only a very few will require sup-
plementary instrumental analysis. His success depends
on his skill as a microscopist, the amount of time he
is able to spend studying known substances and his
ability to remember the identity of particles once seen
and characterized.
low.
A few photomicrographs of common particles fol-
3>
0
I
.• * c 'Kc-
Hi «M / '
Figure 3. Carborundum
r-,"J
( , * & . <* • •
v
-------
MONITORING OF ATMOSPHERIC AEROSOL MASS AND SULFUR CONCENTRATION
E. S. MACIAS
Department of Chemistry
Washington University
St. Louis, Missouri 63130
and
R. B. HUSAR and J. D. HUSAR
Department of Mechanical Engineering
Washington University
St. Louis, Missouri 63130
Summary
A technique for simultaneous measurement of at-
mospheric aerosol mass concentration in two size
ranges and sulfur content has been developed. The
mass measurement was performed with a _two state on
line mass monitor with jjerosol size separator (TWO-
MASS) employing the beta attenuation technique. This
instrument independently analyzes the mass concentra-
tion of two particle size fractions. Coarse particles
are inertially separated by impaction and the fine
particle fraction is collected on a high efficiency
glass-fiber filter. The instrument has been gravi-
metrically calibrated with several laboratory aerosols
of differing chemical and physical properties as well
as ambient aerosol from the St. Louis area.
Filter samples collected by the TWOMASS in the
fine particulate mode (d < 3 ym) were analysed for
total sulfur content by the flash-vaporization flame
photometric technique. This method was calibrated
with five water-soluble sulfur compounds. The mass
measurements were taken with a tiuie resolution of 10
min and with a sensitivity of 5 yg/m3. The sulfur
samples were integrated over two hour intervals with
sensitivity of 0.19 yg/m3.
Introduction
The control of air pollutants in past years has
been focused on gaseous species and coarse particu-
lates. However, continuing deterioration of visibil-
ity in many urban areas, strong evidence of the ad-
verse effects of fine particles on humans, and possi-
ble effect on climate have led to a concern for con-
trolling fine particulates (particle diam £ 3 ym).1
A fundamental requirement of an effective control is
the ability to monitor the pertinent physical and
chemical parameters of the atmospheric aerosols. The
mass concentration in the coarse and fine particulate
fractions and the sulfur content of the aerosol are
considered parameters of major importance.
In this paper we present an on-line method of
simultaneous measurement of mass concentration of
atmospheric aerosols in two size ranges and the subse-
quent determination of their sulfur content.
Experimental Procedures
Particulate Mass Analysis
The mass of atmospheric particulates in two size
ranges, fine particles with submicron diameter and
coarse particles with supermicron diameter, was deter-
mined with a two stage £n-line mass monitor with
aerosol size sampler (TWOMASS)This instrument is
shown schematically in Fig. 1. Because a complete
TWOMASS
IMPACTION
STAGE
St DETECTOR^
S; DETECTOR
FILTRATION
STAGE
TO PUMP
AMP/
DISC.
PRE AMP
COUNTER
TIMER OOUNTER
AMP/
DISC.
PREAMP
TAPE
PROGRAMMABU
CALCULATOR
PLOTTER
Fig. 1. Schematic diagram of TWOMASS including the
data acquisition and analysis system.
description of TWOMASS is given elsewhere3 ,l* only the
essential details are given here.
The flow system separates particles into two size
1
5-2
-------
fractions. Coarse particles (diam > 3 ym) are Im-
pacted on a glass fiber filter with cellulose back-
ing. Fine particles are collected on the same type
of high efficiency-low mass density filter. The flow
rate through the TWOMASS is set at 12 Jt/min using a
GAST rotary vane pump. The impaction and filtration
stages of the TWOMASS are each fitted with independ-
ent source-detector systems which are run simultan-
eously.
The beta detector is a solid state ruggedlzed
silicon surface-barrier detector. It has been chosen
because of its high count rate capability, low noise
characteristics, low cost and simplicity of operation.
As shown in Fig. 1 the output of each detector is sent
to a low-noise fast-risetime preamplifier and then to
a low-noise gaussian-shaping amplifier with a lower
level discriminator set to remove noise. The outputs
of the two discriminators are sent to two counters
interfaced to a programmable calculator. The calcu-
lator determines the mass using the following ex-
pression
M = ~~ MI /I) (1)
fy o
m
where M is the mass concentration in units of Wg/m^»
A is the cross sectional area of the collector j f is
the flow rate, is the mass attenuation coefficient,
I is the count rate of the previous interval and I
o
is the count rate of the current interval. The calcu-
lator plots, prints, and stores this data on magnetic
tape.
The beta intensity transmitted through the filter
paper is measured as the particulates are being de-
posited; mass concentrations are obtained at the end
of each 10-min counting period. The system is auto-
mated with two motors moving each filter tape at two
hour intervals. Note that with this arrangement the
tape is not moved while particulates are being de-
posited or measured and thus, the tape movement does
not effect the mass measurement.
Typical counting rates on each detector are
10,000 cps with the unloaded filter tape in place.
The resulting sensitivity of TWOMASS for 600 sec
sampling times is 5 Ug/nr*. The instrument has been
calibrated with laboratory and atmospheric aerosols
against a microbalance which yielded a linear re-
sponse with mass. The mass attenuation coefficient
for St. Louis Aerosol determined from these measure-
ments is - 0.26 ± 0.02 cm2/mg.
Particulate Sulfur Analysis
Analysis of water soluble particulate sulfur in
the fine particle samples from the TWOMASS instrument
was determined using a flash vaporization-flame
photometric detection method described in detail else-
where.5 Only the essentials of this method are given
below. The sulfur analyzing system is shown sche-
matically in Fig. 2 below. It consists of a flash
vaporization vessel, flame photometric detector, elec-
tronic integrator, and a strip chart recorder. The
sample vaporization is performed by capacitor dis-
CLEAN AIR
TUNGSTEN
BOAT
BRASS
METAL
POST
2.5 V ADJ.
GLASS TO
METAL SEAL
PEAK AREA
INTEGRATOR
INFOTRONICS
CRS 100
SULFUR
DETECTOR
MEL0Y FLAME
PHOTOMETRIC
INSTRUMENT
STRIP CHART
RECORDER
TO PUMP
PEAK AREA
Fig. 2. Schematic diagram of the sulfur detection system consisting of the flash vaporizer, Meloy SA-160 flame
photometric detector, strip chart recorder and the peak area integrator.
2
5-2
-------
charge across a tungsten boat, resulting in resist-
ance heating to 1100°C. Vaporized gaseous decomposi-
tion products of sulfur compounds are carried to the
flame photometric detector by a stream of clean, char-
coal filtered air at a flow rate of 2 cm3/sec.
The detector used is the Meloy SA-160 flame
photometric total sulfur sensor. Its operation is
based on the chemiluminescence of activated molecula.r
sulfur species, produced in a hydrogen hyperventilat-
ed diffusion flame.6 The Meloy sulfur analyzer out-
put is linearized and the sulfur content is registered
directly on the strip chart recorder and the peak area
integrated by an electronic integrator.
Samples of ambient particulate matter were col-
lected on TWOMASS filter tapes. The filter deposits
were extracted in 0.25 to 1 ml double distilled de-
ionized water at 25°C.
Sample extracts and standard solutions were
transferred with a 5 pi microsyringe to the tungsten
boat and heated at 60°C for 30 seconds until dry. The
residue was then vaporized by capacitor discharge.
The sulfur compounds reported to be the dominant
components of the submicrometer range urban aerosol
are sulfuric acid, ammonium sulfate and bisulfate,
zinc and zinc ammonium sulfate.7-10 Calibration of
this technique using standard solutions of these sul-
fates in the range of 0.16-34 lig/ml of sulfur yielded
a correlation coefficient of 0.996 for the above five
sulfates.
The lower detection limit of the method is given
by the purity of the distilled water used and not by
the sensor. A 5 pi sample of distilled deionized
water gave a signal equivalent to 0.3 ± 0.05 ng sul-
fur, corresponding to a solution concentration of
0.06 vg/ml. This value is a factor of five lower
than the measured sulfur content of filter blank
extract.11
The precision of the flash vaporizatiati-FPD
measurement excluding the extraction procedure was
found to be 2.5%. The precision of the entire proce-
dure, Including the extraction step, would probably
be poorer, by a factor of two.
Field Evaluation
Field measurements were carried out in April 1975
in St. Louis at Btation number 112 of the Regional Air
Monitoring Systems (RAMS) in order to test these
methods for determining aerosol mass and sulfur con-
centration. The TWOMASS determined the mass concen-
tration in 10 min sampling intervals on-line ar ' v.oved
the filter tapes every two hours. The two hour filter
deposits were subsequently analysed for sulfur. In
these experiments the light scattering coefficient
(t>scat) and relative humidity were also measured simul-
taneously. These parameters measured during a five
day period are shown in Fig. 3.
The data exhibit a rather striking temporal varia-
tion of fine particulate mass and sulfur content over
the sampling period. Data from 18 April indicate the
arrival of a polluted air masB, drastically changing
both the mass concentration and light scattering coef-
ficient within a one hour interval. A rain storm at
1600 on that day caused the sharp drop in fine par-
ticulate mass concentration. It should be noted that
such observations on this time scale were not pre-
viously possible with conventional gravimetric measure-
ments or beta attenuation instruments. After the
arrival of a clean air mass on 18 April the fine par-
ticulate mass concentration dropped to an extremely
low value (3-5 yg/m3) as shown in Fig. 3. Even at
these low concentrations, meaningful results can be
obtained with the TWOMASS operating with 10 min count-
ing intervals.
Analysis of the data in Fig. 3 shows a strong
correlation between fine particulate mass and light
scattering coefficient with a correlation coefficient
of 0.96 for two hour averages.
The data indicate that sulfur in the fine partic-
ulates expressed aB sulfate (yg/m3) constitutes
15-50% of the fine particulate mass concentration.
It is interesting to note, however, that this fraction
exhibits substantial temporal variations.
Conclusion
The use of the TWOMASS for monitoring the mass
concentration of atmospheric aerosol in the coarse and
fine mode and the flash-vaporization flame photometric
detection method for particulate sulfur analysis has
been shown to be a promising method for measuring
these three aerosol parameters. Future development
of the method should include the automation and on-
line detection of the sulfur content.
References
1. Abatement of Particulate Emissions from Stationary
Sources, National Academy of Engineering, National
Research Council, Report COPAC-5 (1972).
2. The TWOMASS is presently being manufactured by
Meteorology Research Inc. (MRI), Altadena, Calif.
3. E.S. Macias and R.B. Husar, In: Proceedings of
the Second International Conference on Nuclear
Methods in Environmental Research, Vogt, J.R.
(ed.) Columbia, Mo. (1974).
4. E.S. Macias and R.B. Husar, In: Proceedings of
the Symposium on Fine Particles, B.Y.H. Liu (ed.),
Minneapolis, Minn. (1975).
5. J.D. Husar, R.B. Husar, and P.K. Stubits, Anal.
Chem. (in press) (1975).
6. D.P. Lucero and J.W. Paljug, ASTM Special Techni-
cal Pub. 555, Philadelphia, Pa. (1974).
7. C.E. Junge, J. Geophys. Res., ^5, 227 (1960).
8. M.O. Amdur and M. Corn, Amer. Ind. Hyg. Assoc.
J. 24, 326 (1963) .
9. T.G. Dzubay and R.K. Stevens, presented at the 2nd
Joint Conference on Sensing of Environmental Pol-
lutants, Washington, D.C. (1973).
10. R.J. Charlson, A.H. Vanderpol, D.S. Covert, A.P.
Waggoner, and N.C. Ahlquist, Science 184, 156
(1974).
11. J.D. Husar, R.B. Husar, W.E. Wilson, J.L. Durham,
W. Shepherd, J. Anderson, and S. Gregg, EPA
Report (in press) (1975).
3
5-2
-------
ST. LOUIS AEROSOL
APRIL 1975
FINE PARTICLE MASS (//g/m3)
40
20
-A
scat
2.0
SULFATE MASS ( .^g/m3)
RELATIVE HUMIDITY (%)
100
60
20
4/17
19 4/20
TIME (days)
Fig. 3. Field evaluation data of the St. Louis aerosol including determination of fine particle mass, light
scattering coefficient, sulfate mass and relative humidity.
Acknowledgements
The authors gratefully acknowledge the assistance
of Robert Fletcher, Roland Head, Connie Rutledge and
Pamela Stubits in this work. Part of this work was
supported by the U.S. Environmental Protection Agency,
Research Triangle Park, North Carolina, Division of
Chemistry and Physics, Field Methods Branch, under
grant number RB03115-01-0. Special thanks are due the
Environmental Protection Agency and the St. Louis
Regional Air Pollution Study for their cooperation in
a portion of this work.
4
5-2
-------
MEASUREMENTS WITH A PROTOTYPE MASS DISTRIBUTION
MONITOR FOR PARTICULATE AIR POLLUTION
W. Stober and F.J. MBnig
Fraunhofer-Gesellschaft
zur FBrderung der angewandten Forschung
Institut fiir Aerobiologie
5948 Schmallenberg-Grafschaft, Germany
Summary
A monitoring device for the gravimetric determina-
tion of aerosol mass distributions in the atmosphere
is described. The instrument combines the size-separat-
ing capability of a spiral duct aerosol centrifuge with
the mass sensor function of a thin piezo-electric
quartz crystal. This arrangement permits telemetric
read-outs during sampling and does not have the draw-
backs of quartz sensors in connection with impactors
or precipitators at normal gravity. Performance data
and test results simulating particulate air pollution
measurement are given.
Introduction
In recent years, a spiral duct aerosol centrifuge
has been developed (StSber and Flachsbart'0), which
facilitates the dynamic size spectrometry of airborne
particulate matter in the important size range between
0.08 and 6 urn of aerodynamic diameter. The new device
distributes the aerosol particles along a collection
foil in such a way that particles of equal aerodynamic
diameter are deposited at the same foil location. Thus,
a continuous spectrum of decreasing sizes is obtained
along the foil. For some selected standard operating
conditions of the original spiral duct centrifuge,
Figure I shows the calibration curves relating the foil
locations, in distances Jt from the leading edge of the
foil, to the aerodynamic diameters of the deposited
particles. Details of the design and the performance
of centrifuges of this type have been described else-
where and were reviewed recently (Stober®) so that it
may suffice here to describe briefly the essential
features.
When the instrument is in operation, a laminar flow
of clean air enters the spiral duct of the spinning
rotor (Figure 2) through an off-center inlet before the
aerosol is entrained at the axis of rotation. The ratio
of the aerosol flow to the total flow determines de-
cisively the size resolution of the centrifuge. There-
fore, the aerosol sampling rate is usually a small
Figure I: Calibration Curves of a Spiral Duct Centri-
fuge
Figure 2: Rotor of a Spiral Duct Centrifuge
fraction of the total flow rate. The collection foil
covers the outer wall of the spiral duct where the par-
ticles are deposited under the influence of both, the
centrifugal forces and the winnowing effect of the
clean air. By properly accounting for the operating
conditions, the sampling rate and the sampling time
(Stober and Flachsbart'', Kops et al.6) it is possible
to relate the size-selective deposit concentrations of
particles to the size-selective airborne concentrations
of the sampled aerosol. Thus, besides the maintenance
of the operating conditions an aerosol size distribu-
tion analysis by means of a calibrated spiral duct cen-
trifuge requires only the measurement of the deposit
concentrations along the collection foil. This is done
either by (electron) microscopic inspection and count-
ing of the deposited particles per unit area or, if
feasible, by weighing the deposits of selected areas
along the collection foil. Both procedures require the
collection foil to be removed from the centrifuge.
Hence, the analysis is a batch-wise operation.
If the spiral centrifuge is to be developed into a
convenient instrument accomodating modern preferences
for real-time instrumentation and automation, the
batch-type analytic techniques mentioned above have to
be abandoned. For this reason, an effort was made to
replace the tedious, batch-wise size distribution ana-
lyses by a procedure permitting an automated analytical
measurement. Since modern air pollution criteria and
standards are demanding information in terms of air-
borne mass concentrations, the replacement of the
gravimetric micro balance determinations of the foil
concentrations of deposited mass by some automated
method appeared to be the most desirable approach. A
promising route for this is the utilization of piezo-
electric quartz oscillators which have been applied to
the measurement of homogeneous surface deposit concen-
trations for a long time (Sauerbrey®) and, more
5-3
D ¦ 19 lilers/min
0 • 10 liters/mm
5 lilers/min
-------
recently, were incorporated in sampling instruments
measuring total aerosol mass concentrations (Chuan ,
Olin and Sem7, Carpenter and Brencbley1).
The following is a report on a feasibility study and
some prototype measurements probing the employment of
piezo-electric quartz oscillators as sensors for meas-
uring and monitoring size-selective aerosol mass depo-
sit concentrations along the outer wall of a spiral
duct centrifuge.
Sensitivity Requirements for Mass Distribution Monitor-
ing with a Spiral Duct Centrifuge
Based on the performance data of the original spiral
duct centrifuge (StSber and Flachsbart'°), a sampling
rate of 3 liters min"' at a size resolution of some
15 % can be achieved under suitable operating condi-
tions (rotor speed 3000 rpm, total flow rate 19 liters
min~'). They correspond to the upper calibration curve
in Figure 1.
Assuming an atmospheric particle size distribution
as proposed by Junge*
dn
dD„
1
ae uae
the mass frequency density would be
dm
1
Dae
dDae
and, along the collection foil, we find
dm 1 dDae
di
"ae
dl
(!)
(2)
(3)
From Figure I, it can be concluded that approximately
D a
ae I
Hence,
and
dm
di
<*Dae
it
7
and by truncating the distribution between 0.1 and
5 ym, the total mass becomes
*1
"J
max
d In I
- In
*iBax
lmin
(4)
(5)
(6)
Figure I indicates approximately
£ < 160 cm and I . > 9 cm
max "* nan —
Present regulations on surveillance of particulate
air pollution call for 24-hour mean values of total
airborne mass concentrations. A minimum requirement for
a spiral centrifuge with quartz sensors would then
necessitate a capability of the sensor to detect a de-
posit at a surface concentration of cf • in a 24-hour
sample of relatively clean air, which may11 be assumed to
have an airborne mass concentration around 20 micro-
grams m~3. For 24 hours of sampling, this would give
values of
-2
m„ ¦ 86.4 yg and » 62 ng cm
i w t mm
Then, for 4 99 J confidence level of detection, a stan-
dard deviation of
cf min
2.58 * >t1
± 17 ng cm
(7)
corresponding to the apparent average deposition in
24-hour control runs without actual sampling would be
permissible.
Theoretical Sensitivity of Quartz Crystals for Measur-
ing Surface Mass Deposits
The basic properties of piezo-electric crystals are
described in the literature'» 8. por the determination
of surface layers deposited on piezo-electric quartz
sensors, disk-shape crystals are used. The quartz is
cut along its crystal planes in such a way that the two
surfaces of the disk will oscillate mechanically in a
thickness-shear mode with the piezo-electric resonance
frequency when the external electric field applied bet-
ween the two sandwiched electrodes on the disk surfaces
oscillates with a similar frequency. The crystal then
forces the external circuit to oscillate exactly at the,
resonance frequency. Quartz disks oscillating in the
shear mode as described are called AT quartz crystals.
For AT quartz crystals, the resonance frequency is
simply
f - f (8)
where N » 167 kHz x cm is a material constant for
quartz and s is the thickness of the crystal disk.
Thus, the frequency will shift with the thickness of
the crystal and it is easy to show that
df
f
ds
s
(9)
so that
However, since, the fundamental oscillation in the
shear mode does not produce any nodal points on the
crystal surfaces, the material closely adjacent to the
crystal surface influences the frequency only by its
inertia, i.e. mass, and not by any elastic properties.
Therefore, as confirmed by Sauerbrey®, we can also ex-
pect for a homogeneous surface layer of density p
In
*max
*mi«
2.9
Thus, with the actual width b = 3 cm of the foil de-
posit, the surface concentration will be
1
b'l
In*
0.12 -i
jt
and for the smallest sizes around 0.1 urn we obtain
c_ . » 7.2 * 10
f mm
-4
"T
as a minimum surface concentration, which must not be
below the detection limit for a piezo-electric quartz
sensor.
TT * ™ ^ *1" CO)
f PqS
where pq - 2.65 grams cm" 3 is the density of the
quartz crystal. Further generalizing, we obtain with
the sensitive surface area F as determined by the area
of the crystal electrodes
df
T
dm
Pq
sF
(11)
where
dm ¦ p F ds
is the mass deposited homogeneously in a layer of
thickness ds over the area F. This should hold for
small amounts of all masses which are deposited on the
surface in such a way that they fully participate in
the piezo-electric oscillation- The surface concen-
5-3;
-------
tration
dC_
dm
F
(12)
(13)
can then be measured by the frequency shift df as
2
p s df
dC = -
F N
Thus, the sensitivity of the quartz crystal as a de-
tector for the concentration of homogeneous surface
deposits is proportional to the sensitivity in measur-
ing the frequency shift. The detector sensitivity can
further be increased by reducing the thickness of the
crystal, i.e. by increasing the resonance frequency.
AT quartz crystals of 0.0167 cm thickness with a
resonance frequency of 10 MHz are easily commercially
available. Their thickness renders them rather fragile
and, therefore, the resonance frequency of 10 MHz may
constitute a practical upper limit.
For these crystals, we obtain from equation (3)
dCL
"F , , -2-1
-T7- ¦ 4.4 ng cm sec
at
(14)
and since the frequency of 10 MHz oscillators can
easily be determined within less than 1 Hz, it appears
that the sensitivity condition (7) can be met. It
would require
df - 4 Hz
Figure 4: Rotor Lid with Outer Wall Insert and Quartz
Sensors
Since the piezo-electric resonance
AT quartz and the frequency shift due
layer can be triggered in the parallel
with an aperiodic oscillator stage, a
circuit^ was used as shown in Figure 5
consists of an oscillator stage Ti, a
T2 and an output stage T3.
miniature transistors.
1:
The active
frequency of the
to any surface
resonance mode®
Pierce-Colpitts-
This circuit
separator stage
components were
Prototype Design
An AT quartz disk of 0.9 cm diameter with a nominal
piezo-electric frequency of lOMHzwas selected as the
basic sensor unit. Commercially available blanks are
coated with gold and form electrodes of a pattern as
indicated in Figure 3. The diameter of the circular
center is 0.5 cm. The graph further reveals schemati-
cally the mounting of the sensors. Thin brass foils
soldering tog
quartz crystal
soldering tag
Figure 3: Quartz Sensor Assembly
served as soldering strips and held the crystals flush
in suitably cut and etched depressions of a copper-
sandwiched card board paper. For the prototype model,
a total of four crystals were mounted to a strip of
this card board paper which was attached to the rotor
lid of the centrifuge in such a way that it lined the
outer wall of the spiral duct when the lid was in-
serted. Figure 4 is a photograph of the prototype lid.
It shows two of the four quartz crystals with their
sensitive areas on the center line of the wall and
facing the duct.
10 MHz
quartz
Figure 5: Pierce-Colpitt Oscillator Circuit
Four of these units linking individually to the four
crystals were installed on the rotor lid together with
a decoder and an electronic counter. Figure 6 shows the
graph of the electronics. Counter and decoder will ac-
tivate the oscillators one at a time. Sequential
switching is triggered by an external flash light puls-
ing a photo transistor mounted on the rotor. The re-
sonance frequency is transmitted by a common circular
antenna on the rotor and picked up by an external re-
ceiver. Figure 7 is a photograph of the integrated
electronic circuitry on top of the rotor lid.
r
~i
X
X"
L 0
.'!!!! _J
Figure 6: Block Diagram of Telemetry Electronics
5-3
-------
Figure 8: Ideal Temperature Characteristics of AT
Quartz Crystals
[•C]
Figure 7!" Integrated Components on Rotor Top
A high precision electronic counter was used for the ex-
ternal measurement of the resonance frequencies of the
crystals. The counter permitted a direct determination
of Jhe crystal frequency of 10? Hz. It had a precision
of - 0.1 Hz for a time base of 10 seconds. The abso-
lute error per day was * 0.3 Hz if the ambient tempera-
ture was kept within 2 °C of the nominal value.
Crystal Frequency Variations by Ambient Influences
The precision of the electronic counter is not the
limiting factor for the accuracy of measuring the fre-
quency shift related to the build-up of a surface
layer. Instead, there are a number of ambient influ-
ences which may cause variations of the resonance fre-
quency of a crystal without any change of the surface
deposit. Thus, for a reliable measurement of the de-
posit concentrations the ambient influences have to be
checked and kept small.
The most obvious ambient cause of frequency shifts
is the variation of the crystal temperature. Unfortu-
nately, these changes are not unique for all quartz
crystals but depend upon the precision to which the
surfaces of the ground quartz disks match the crystal-
lographic plane of the quartz lattice. Changes of the
cutting angle by a few minutes of arc are very influ-
ential, although, on the other hand, precisely cut AT
quartz disks have a favorable tempegature characte-
ristic at a cutting angle of 9 ¦ 35 10' as shown
according to Sauerbrey in Figure 8. In practice, how-
ever, commercial batches of AT quartz disks showed a
variety of actual temperature characteristics. Figure 9
gives the results for three randomly picked crystals
used for test runs. Apparently, this variety of thermal
frequency shifts would impair any effort of suppressing
the temperature influence as otherwise possible by mea-
suring the beat frequency to a reference quartz
crystal in thermal equilibrium with the sensor quartz.
For this reason, the prototype design was laid out to
measure the absolute frequency of the individual quartz
disks. The latter were selected with regard to the
mggnitude of their thermal frequency shift at 25 ' *
5 C. A quality like crystal Qj3 in Figure 9 was deemed
acceptable for the prototype.
As reported in the literature (Daley and Lundgren^)
the frequency of AT quartz plates is influenced by the
relative humidity of the air in contact with the
A f
f
x 10
-7
/ Q19
1
1
v 1
\ 1
1 /
1/
1
q12
x/l
10 2]
//\ f
y j \30
LO T [°C]
/1
/ 1
\
/ 1
! 1
1
1
1
Q13
20
15
10
5
0
-5
-10
-15
-20
Figure 9: Actual Temperature Characteristics of AT
Quartz Crystals
Figure 10: Frequency Shift and Relative Humidity
5-3
-------
crystal surfaces. For the quartz disks used in the pro-
totype design a decrease of close to 1 Hz was typical
for an increase of ! % rH. Figure 10 gives some data
and their reproducibility for a typical 10 MHz crystal.
It indicates that, in practice, the mass deposits
should be measured under constant if not standardized
relative humidity conditions. This is all the more pre-
ferable since the sampled deposit itself may also
change weight with varying relative humidity.
Test Runs with Laboratory Aerosols
The prototype design was tested in sampling runs
with laboratory aerosols simulating realistic situa-
tions. In these runs, the spiral duct centrifuge was
operated at 3000 rpm and a total flow rate of 10 liters
min-'. Under these conditions, three of the sensors
(Q2> Q3. Q4) would receive aerosol particles of aero-
dynamic diameters of 1.1 ym, 0.8 ym and 0.5 ym, respec-
tively, while the fourth sensor (Qj) was mounted as a
reference crystal at the beginning of the spiral duct
where no aerosol will deposit. The aerosol sampling
rate was 1 liter min"' corresponding to a size resolu-
tion of better than 10 %.
Nebulized aqueous solutions of fluorescein were used
as a laboratory aerosol. This test aerosol has several
advantages for basic evaluations as described else-
where'^. It was selected for this study because it
could be subjected to a simple, fluorometric procedure
to determine the mass deposit concentration indepen-
dently of the piezo-electric measurement. For this
purpose, sampling foils were mounted on the wall of the
spiral duct upstream and downstream adjacent to the
quartz sensors. The foils were removed after each run
and their fluorometric surface concentration values
were averaged and compared with the piezo-electric
measurements.
[H«]
1500" 15,47
1*26
2 500 • 11,05
500-
[l
••
• e
0.°
. i:.
0 to 20 30 (0 SO to 70 60 90 100 ItO 120 t[m«$
Figure 11: Deposit Increase During Sampling of Dense
Fluorescein Aerosol (30 mg nf^)
Figure 11 shows the frequency shifts during a 2-hour
sampling test with a fluorescein aerosol of very high
concentration ( = 30 mg/m^). The ordinate scale gives
the frequency shift and the corresponding theoretical
value of the surface concentration of the deposited
mass. The fluorometric mass determinations were consi-
stently below these values varying between 98 % at
1.1 vim and 75 7. at 0.5 ym. This may indicate that the
particle mass may not have been composed of pure
fluorescein alone. However, the results certainly show
that heavy particulate air pollution (500 yg/m^) may be
sampled continuously for more than 120 hours without
overloading the quartz crystals.
A series of test runs were made at airborne concen-
tration levels comparable to moderate air pollution
(50 yg/m^) situations. Figure 12 shows the results for
sequential sampling periods of approximately 6 hours.
Apparently, significant frequency shifts are obtained
in a fraction of that time although the sampling rate
was merely 1 liter min-'. Very little drift occurred
during the stand-by times, however all crystals reveal-
ed a reproducible frequency "jump" between stand-by
periods without rotation and actual operation.
o
20
to
60
eo
100
120
140
160
180
200
220
240
260
200
300
,'iiop'; ' 'Vtop"U ,"»top" ft
' ¦ ; 'jt run / I \ run
A *0 J
i t op", j
t I
-j J v fun
1»h . 6*1 1»h th T75h «5h I86h S«l>
-5°A-9'
\'2
'3
A MHz)
Figure 12: Sequential Sampling of Dilute Fluorescein
Aerosol (50 yg m~^)
These results encourage the design and construction
of a suitably dimensioned instrument based on the con-
cept of the spiral duct centrifuge and incorporating
a number of quartz sensors for size-selective mass
monitoring of airborne concentrations covering the
range from light particulate air pollution to heavy
source emissions.
References
1. Carpenter, T.E., and D.L. Brenchley, Amer. Ind. Hyg.
Assoc. J. 33 : 503 (1972)
2. Chuan, R.L., J. Aerosol Sci. J_ : 111 (1970)
3. Daley, F.S., and D.A. Lundgren, Amer. Ind. Hyg.
Assoc. J. 36 : 518 (1975)
4. Junge, C., Tellua 5 : 1 (1953)
5. Koch, H., Transistorsender, Franzes-Verlag, Munich
1969
5
5-3
-------
6. Kops, J., L. Hermans and J.F. van de Vate, J, Aero-
sol Sci., 6 ;— <1975) in press
7. Olin, J.G., and G.J. Sem, Atmosph. Environm. 5_ :
653 (1971)
8. Sauerbrey, G.Z., Zeitsch. Physik 155 : 206 (1959)
9. Stober, W., Design, Performance and Applications
of Spiral Duct Centrifuges, Proceedings of the
Symposium of Fine Particles, Minneapolis, Minn.,
May 28-30, 1975
10. Stober, W., and H. Flachsbart, Environm. Sci.
Technol. 3 : 1280 (1969)
11. StSber, W., and H. Flachsbart, J. Aerosol Sci.,
2 : 103 (1971)
12. Stober, W., and H. Flachsbart, Atmosph. Environm.,
7 : 737 (1973).
6
5-3
-------
THE APPLICATION OF PATTERN RECOGNITION TECHNIQUES
TO THE CHARACTERIZATION OF ATMOSPHERIC AEROSOLS
S. P. Perone, M. Pichler, P. Gaarenstroom
Department of Chemistry
Purdue University
W. Lafayette, Indiana 47907
and
J. L. Moyers
Department of Chemistry
University of Arizona
Tucson, Arizona 85721
Summary
Measurements were made of the amounts of 24 com-
ponents in 24-hour samples of atmospheric particulates
collected over a 12-month period in the greater Tucson
area. Techniques of pattern recognition were used to
examine the data base, which also included meteorolog-
ical information collected daily over the same period
of time. Clustering methods grouped similar components
together. Principal Component Analysis showed that
most of the variance was contained in only a few
dimensions.
Introduction
Atmospheric particulates were collected from 11
locations in and around Tucson, Arizona. Each sampling
period lasted 24 hours. One of the locations was
located 50 miles from Tucson and served to establish a
background level for component amounts. Samples were
collected over a period of about one year. Analysis
of particulate chemical composition was done at the
Uni-versity of Arizona. The 24 components measured
+ - 2-
were: total mass, NH^ , N03 , SO^ , Si, Ti, Cs, Li,
Rb, Al, Zn, K, Fe, Ca, Mg, Na, Pb, Cu, Mn, Sr, Ni, Co,
Cr, and Cd.
It was the purpose of the work reported here to
apply computerized pattern recognition techniques to
the inteVpretation of particulate composition data.
One specific objective was to identify the number of
different polluting sources and the related pollutants.
Another objective was to investigate multi-dimensional
meteorological effects on particulate composition. A
variety of pattern recognition methods have been
applied, as described below. In this discussion a
"pattern" will consist of all measurements (chemical
composition of particulate samples and meteorological
data) collected for a particular day at a particular
location. There were 449 patterns. The measurements
are referred to as "features". Most studies have been
conducted with partial patterns consisting of only
compositional features.
Data Treatment
Since no training set of known and classified
patterns exists, supervised learning techniques1 of
pattern recognition cannot be used. Methods of unsu-
p
pervised learning or cluster analysis were utilized to
see whether patterns or features could be grouped
together; how many clusters should be formed; and what
did members of a cluster have in common.
Clustering can be done on either patterns or
features. The methods used are nearly identical. A
measure of similarity is needed between every pair of
patterns or features. A distance measure, usually
Euclidean distance, is the similarity between two
patterns which have been represented as two vectors in
n-space, where n is the number of features per pattern.
The correlation coefficient is a similarity measure
between two different features of a number of patterns.
Since a human's ability to recognize clusters in
2 or 3 dimensions surpasses that of any clustering
algorithm, different kinds of mappings are used to
reduce the dimensionality of the data to 2- or 3-dimen-
sional space. Mappings are used to display the
arrangement of patterns, but with slight modification
can display features as well.
Clustering of Features
Two clustering methods were used to determine
those features which behaved similarly. These were a
hierarchical clustering technique and a nonlinear
mapping. Correlation coefficients were the basis of
the similarity measurement. To make the similarity
analogous to distance, three similarities were tried:
(1) l-rij, (2) 1-1 rij |, (3) l-(r..j)2, where is the
correlation coefficient of features i and j. Two
identical features would have a similarity of zero.
In this work there was little difference in the results
produced when using different similarity measures. The
main difference between measure (1) and the others is
the treatment of negative correlations, but in this
data base negative correlations were infrequent and of
small magnitude.
A hierarchical clustering method is initialized by
having each cluster contain one of the elements to be
clustered. Clusters are connected one at a time
beginning with the two most similar until all the
clusters are joined into or>e. A dendrogram is a dis-
play of the order in which the clusters were joined
and the similarity between the clusters that were
joined. If the similarity between two clusters is the
same as that of their most similar pair of elements,
one from each cluster, the strategy is called "nearest-
neighbor" or "single-linkage". The furthest neighbor
method uses the similarity of the most dissimilar pair
in the clusters as the similarity of the two clusters.
Another method uses the average similarity between the
elements in the two clusters and is called the
2 3
centroid method. '
Figure 1 is a dendrogram for chemical features
from one location in the Tucson area. The dendrograms
reveal two fairly tight clusters of chemical features.
These are Si, Ti, Al, Sr, Li, Mn, in one cluster and
K, Fe, Ca, Mg, Na in another. These two clusters plus
Rb, Cr, and mass form a looser cluster. Around this
larger cluster are Cs, Co, and Ni. Seven features are
.f.
far removed from any of the others. They are NH^ ,
SO,,2", Cu, Sn, Cd, NO,", and Pb. With the exception of
+ 2-
the highly correlated pair, NH^ and S04 , these
features have generally low correlations with other
1
5-4
-------
features.
correlation plot tocflrroN z.
HIERARCHICAL FEATURE CLUSTERING - LOCATION 2
flTFl "rCT
MS SI TI AL SR LI MN K fE. CH NA MO RB CR CS CO NI NH SO CU NO PB ZN CD
Figure 1. uenrogram of features from a location in
Tucson. Nearest neighbor clustering was used. Special
symbols: MS = total mass, NO = NO^", SO = SO^2 , and
NH = NH4+.
Clustering algorithms work best when the clusters
are in the shape of hyperspheres. Other shapes make it
difficult to distinguish which are clusters. A human
interaction is the best way to recognize clusters of
other shapes, but the display must be restricted to 2
or 3 dimensions.
A nonlinear mapping was used to display the
relationship of features to one another. Nonlinear
mapping has been devised to display patterns in two
dimensions by preserving interpattern distance in 2-
space as much as possible to what the distance was in
n-space.4,5 In this work the technique was modified to
observe the features instead of the patterns. To dis-
play features, the similarity measurement as defined
above was used in place of distance. Results of the
nonlinear mapping agreed closely with the clusters as
revealed by the hierarchical clustering. Figure 2 is
a nonlinear mapping of the chemical features for the
same location as in Figure 1. Features which the
hierarchical clustering algorithm joined into clusters
are found close together in the mapping.
Features which cluster probably reflect a common
source. The similar behavior of these features also
indicates that it should not be necessary to monitor
all of the species in the same cluster if the goal is
to gauge the output of that source.
Correlations of Chemical Features with Weather
Meteorological measurements were available for
each sampling date. The 24 weather measurements were:
the maximum, minimum, and average of temperature,
relative humidity, dewpoint, visibility, and wind
speed, as well as resultant wind speed, wind direction,
direction of fastest wind, precipitation, pressure,
daytime sun, percent sunlight, daytime sky cover, and
24-hour sky cover. Correlations between any weather
measurement and chemical features were generally low,
indicating no simple linear relationships. However,
lack of variability 1n the weather conditions and
averaging effects of the 24 hour sampling period make
the weather data less useful for correlation studies.
CO Cb
Figure 2. Nonlinear mapping of features at a location
in Tucson. Same special symbols as in Figure 1.
Intrinsic Dimensionality
Principal component analysis was chosen as a tech-
nique to ascertain the intrinsic dimensionality of the
chemical composition data. In principal component
analysis each of the n observed features, zj> is
described linearly in terms of n new uncorrelated com-
ponents, Fr F2, ..., Fn.6
J jl 1 j2 2 jn n
The components are ordered according to the contribu-
tion each makes to the total sum of the variances of
the variables. For highly correlated data a few com-
ponents will account for a large percentage of the
total variance. Each aj.. 1s called the loading of
variable j on factor i, but is also the correlation
coefficient of variable j with factor i.
To calculate a^.. requires finding the eigenvalues,
Xj, and eigenvectors, $>., of the correlation matrix of
the data, R.
R*i = Vi
From ^ = (c*j., c^.., ..., ani)T, each a^ is calculated
by ad1 = A'i a j 1,
The contribution of a component, F^, to the total
variance is the ratio of its eigenvalue, A.., to the sum
of all the eigenvalues.
fraction of total variance due to F. = —1—
i n
Z Xn
PB1 p
When the correlation matrix of the particulate
compositions was submitted to principal component
analysis, it was found that most of the variance was
5-4
-------
contained in only a few dimensions. For a typical
sampling location about 60% of the variance is contain-
ed in two dimensions, 75% of the variance in four
dimensions, and 90% of the variance in eight dimensions.
Table 1 lists the fraction of the variance of ten
dimensions for a single sampling location.
Table 1.
Fraction of Variance in
Direction of
First Ten Eigenvectors
Location 2
Eiqenval
ue % of Variance
Cumulative %
1.
13.22
55.1
55.1
2.
3.52
14.7
69.8
3.
1.45
6.1
75.8
4.
1.2,3
5.1
80.9
5.
1.04
4.4
85.3
6.
0.72
3.0
88.3
7.
0.57
2.4
90.7
8.
0.52
2.2
92.8
9.
0.38
1.6
94.4
10.
0.31
1.3
95.7
Two-dimensional Displays of Patterns
Linear and nonlinear mappings of the patterns were
made to visually examine them in two-dimensional space.
One kind of linear mapping is to examine only the
coordinates of the patterns of two features. These
displays illustrate the scatter of the patterns about
a line which would represent perfect correlation of
the two features.
Another kind of linear mapping is an eigenvector
plot. The patterns are transformed by multiplying each
by the transpose of a matrix consisting of the two
eigenvectors with the largest eigenvalues. This
results in the linear mapping which preserves the most
variance. The eigenvector plot of two locations is
shown in Figure 3. One location (labeled E) is the
outlying location, while location 2 is within the
City of Tucson. It is seen that the locations are
almost separated in the direction of the first
eigenvector, but in the second eigenvector direction
the patterns are completely overlapped.
ClOENVECTOR PROJECTION OF LOCUTIONS 2 I 11
One rule of thumb used for estimating the number
of significant principal components is to use the
number of eigenvalues which exceed 1. This rule would
indicate the number of significant components to be 4,
5, or 6, depending upon the sampling location.
Table 2 lists the correlations of the chemical
features with the first two principal components.
Those features which correlate well with the first
principal component indicate that this component is a
measure of primary particulates of a crustal origin in
the atmosphere. The second principal component corre-
,4+ and S042"
condensation particulates of anthropogenic origin.
Correlations of the chemical features with other
principal components are not as high.
Table 2. Correlations of Chemical Features with
First Two Principal Components
lates best with NH, and SO, which are found in
Feature
Corr. with F^
Corr. with F,
Si
.96
-.13
A1
.96
-.13
Mn
.95
-.11
Mg
.94
.00
Li
.94
-.08
Sr
.93
-.08
Fe
.92
.01
Ti
.91
-.21
K
.91
.08
Cr
.89
.02
Ca
.87
.08
Rb
.86
-.01
Na
.86
.14
Mass
'Y:'
-.01
Cs
-.12
Co
1 j'.
-.01
Ni
. t3
.37
no3"
.22
.52
Zn
.18
.58
Pb
.09
.71
Cd
.09
.46
Cu
.02
.72
nh4+
-.04
.83
2-
S04
1
O
-P*
QO
2
2
2
E 2 2
E E E 2 #
EE 2 2 E 2 2 2 22
e£i|zE f2 2 Z ?£
E \ E 2 2 Z 5?
2
2 2
2
Figure 3. A linear two dimensional display of patterns
from 2 locations. The patterns labeled E are from the
outlying location.
Non linear mapping was also performed to see if
any of the patterns within a location would cluster
together, but no clustering was observed.
Conclusions
The clustering, mapping, and principal component
techniques used point to a crustal origin of a large
number of the atmospheric particulate components in the
Tucson area. A few components, such as Cs, Co, Ni,
probably have several sources and one of these is
crustal. The lack of correlation among Cu, N03", Pb
+ 2-
Zn, Cd, and the NH^ , S04 pair indicates no single,
common source is responsible for the components.
-------
To reconcile the obvious importance of meteorolog-
ical factors in determining pollution levels to the
lack of correlation between chemical and weather
features in this study, it appears as if a multi-
variate approach will have to be taken which considers
several weather features at a time to describe chemical
feature behavior.
Results from the pattern recognition treatment of
the data can be useful in designing future sampling
projects. The number of components that are measured
could be reduced since behavior of several groups of
the chemical features were similar. Use of pattern
recognition in the interpretation of data requires
careful consideration as to the design of the experi-
ment to make the data analysis effective.
Acknowledgements
This research was supported by National Science
Foundation Grant Number MPS74-12762, and by The Office
of Naval Research contract N00014-67-A-0226-0021.
Partial support by the Electric Power Research
Institute is also gratefully acknowledged.
References
1. B. R. Kowalski, C. F. Bender, J_. Am. Chem. Soc.,
94, 5632 (1974).
2. B. Everitt, Cluster Analysis, Ch. 2 (1974).
3. G. N. Lance, W. T. Williams, Comp. J., 9, 373
(1967).
4. J. W. Sammon, IEEE Trans. Computers, C-18, 401
(1969).
5. K. Fukunaga, Introduction to Statistical Pattern
Recognition, Ch. 9 (1972).
6. Harry H. Harman, Modern Factor Analysis, 2nd ed.
U. of Chicago Press (1967}"]
4
5-4
-------
THE USE OF A DUAL BEAM LASER TRANSMISSOMETER
AS A MEANS OF MONITORING AIR QUALITY
William C. Malm and Kenneth O'Dell
Department of Physics
Northern Arizona University
Flagstaff, AZ 86001
Abstract
This study will report on comparative air quality
measurements made at Lake Powell and Flagstaff. The
results of measurements made by a dual beam laser trans-
missometer, calibrated using a very precise «1$ error)
two source telephotometer, will be compared to measure-
ments of visibility (and thus the extinction coeffi-
cient) using the integrating nephelometer, as well as
contrast techniques utilizing cameras and photometers.
Additionally, theoretical calculations of aerosol mass
loading and size distribution bas'ed upon the trans-
missometer results will be compared to conventional
measurements of these air quality parameters.
Introduction
The assessment of air quality in isolated areas is
a very different problem than that of assessing urban
air quality. For example, the visual impact of the
Grand Canyon and the Lake Powell areas is highly de-
pendent upon visibility ranges which often exceed 180
km.l- The addition of very small amounts of atmospheric
aerosol would significantly degrade visibilities in
this range and degrade the aesthetic value of the area
accordingly. In spite of this, surprisingly little effort
has been directed toward obtaining the data of a type
that would be useful in establishing meaningful stan-
dards which, with the aid of informed management, would
result in the maximum use consistant with the preser-
vation of these unique resources for future generations.
In assessing the air quality of the above men-
tioned areas, the following limitations of certain tra-
ditional methods have been noted:
1) High-volume samplers, segregators, and
cascade impactors are all point measure-
ments, usually time-averaged over a period
of several hours.
2) Integrating nephelometers give quick read-
out, but are still point measurements. In
clean air, commercial instruments may give
readings that are in error by as much as
+100$ of the recorded value. Additionally,
these instruments ignore the absorption
component of the extinction coefficient.
3) The camera-telescope techniques used to
record photographic range are not well suited
for continuous monitoring of the atmosphere.
This measurement must be made during daylight
hours and has the additional restriction of
cloudless skies.
U) Sampling methods using filtration techniques
do not typically capture particles less than
.ltA<.in size and will typically ignore a
liquid aerosol.
These limitations are particularly serious when one is
attempting to understand the dynamics of air quality
in relatively clean atmospheres.
Our studies indicate that a transmissometer using
two lasers at two wavelengths may prove to be a super-
ior method of monitoring atmospheric aerosol loading,
size distribution and visual range in remote areas. A
dual beam laser transmissometer (DBLT) allows contin-
uous monitoring of all these parameters over extended
ranges 21+ hours a day -- a criteria not met by any
other of the above mentioned techniques.
Theoretical Approach
Numerous methods are being used for monitoring
visual range2j3,U and others have been proposed for
determining aerosol size distributions.5)6,7 Most re-
ly on the interaction of electromagnetic energy and
the molecules with which the atmosphere is composed of
and with the aerosol or particulate matter contained
within the atmosphere. A beam of E 4 K radiation pass-
ing through the atmosphere will be attenuated either
by absorption or scattering. In the case of clean at-
mospheres and in the wavelength range of visible
radiation, attenuation due to absorbtion is essentially
nonexistent. However, scattering occurs at all visi-
ble wavelengths and is thus primarily responsible for
reduced visibilities and increased extinction coeffi-
cients.
The attenuation of a beam of electromagnetic radi-
ation is usually described by:
-bt(h,A)*
I(h, X) = I0O) 6 (1)
where I and I0 are the observed and initial intensities
respectively, bt is the total extinction coefficient
i.e., bt, {hiX) = babsorbed + ^scattered) and * is the
distance through which the beam has traveled, h and
X are the elevation above sea level and wavelength
respectively. If babSorbe'd = 0, the remaining contri-
bution to the extinction coefficient is molecular (Ray-
leigh) and aerosol or particulate scattering.
The scattering characteristics of aerosols depend
on their physical size, shape and index of refraction.
The atmosphere normally contains aerosols of different
sizes as well as varied aerosol mass loadings.
To determine the effect of aerosol size distribu-
tions and aerosol mass loading on extinction coeffi-
cients the scattering theory of Mie® was used. Mathe-
matical formulations of the Mie theory will be kept to
a minimum since the reader can refer to the clear, con-
cise discussion of van de Hulst.9 Given the size par-
ameter X =i1Tr/>. (r being the radius of the aerosol and
the wavelength of the incident electromagnet wave);
the complex index of refraction of the aerosol, n*;
and the scattering angle measured from the forward di-
rection, 9; the following quantities can be calculated:
oo
Si(X,n->^0) = §>_ „4;,1(am77m + biti7m) (2)
1 m(|m + i; j
oo
S2(X,n*,Q) = + 1 (kmTTni + amTm), (3)
m=l m(m + 1)
1
5-5
-------
Qext = 2 (2m + i) Re (am + bm), (U)
U m=l
and
oo
Qsca Si (2m +1} (H2 + M2)- ^
Sj_ and S2 are the dirnensionless, complex amplitude
functions related to the components polarized in planes
perpendicular and parallel respectively, to the scatter-
ing plane; Qext and Qsca are the total extinction and
scattering cross sections divided by the geometrical
cross sections i.e., the efficiency factors, am and bm
are the complex scattering coefficients while 7fm and
X m are the angular dependent coefficients. The half
order Bessel functions, circular and hyperbolic func-
tions that are contained in am and bm were evaluated
using the usual recursion formulas. The quantities am
and bm were terminated when (|am|2 + |bmp)/n < 10"12.
Using the above formalism, the extinction coeffi-
cient is given byt r?
,b(A)ext =TT \ r2 Qext U» n*) dN
(6)
where dN is the number of particles per cubic meter per
increment in r. dN can be represented by a power law
proposed by Junge
(7)
In this equation, N, r, c and 6 are respectively the
number of particles/unit volume, radius, and constants
to be determined. dH is the differential number
dlnr
density per increment in Inr. Vhen In dN is plotted
dlnr
against lnr, a straight line is derived with a slope
equal to (h . ft is representative of the relative ratio
of small to large particles; the ratio increasing as
increases. Using equation 7 the extinction coeffi-
cient becomes: ro
k (A )ext =77"c \ Qext dlnr (8)
The determination of (i and c specifies the size
distribution and allows the calculation of the aerosol
mass loading via the following equation:
0-7
<
hi
lli O'fa
a
U-b
M = U/3 17 c
J
r3"<*
dlnr
(9)
'At
;// \l
0-2.5
/ /
\ A 0-1.5
tiki \ >A
/ p"3.5
/
V u V
/ /
/ / y
in .---i' 1 1 1 1
1 1 1 1 _i_.
RADIUS IN MIIMR,
FIGURE 1 Argument of Equation 8 plotted as a function
of particle radius for Beta = 1.5, 2.5 & 3.5
Special nomograms have been constructed to yield
Beta, c, and aerosol mass loading from the ratio of
b633nm/bl|i42nm> These are shown in Figures 2, 3 and k
respectively. (All calculations are based on r2=10M,
n* = 1.5, and/'3 2.6 gm/cirK, all representative num-
bers for Flagstaff and Lake Powell.)
If the extinction coefficient due to aerosols is
accurately described by wavelength dependent equation
8, the two unknowns, <3 and c, can be determined by
measuring the atmospheric attenuation at two wave-
lengths, subtracting contributions due to Rayleigh
scattering, and simultaneously solving the two result-
ing equations.
Some care must be taken in establishing integration
limits ri and rg. To determine the importance of these
limits, the arguments of equation 8 were plotted (see
Figure 1) as a function lnr for different 'a. Since
measured aerosol size distributions described by Junge's
law extend down to at least .01** , and since Figure 1
shows the argument of equation 3 to be negligible below
.OlM, rj was chosen to equal .Olx^.. The selection of
r2, the upper limit of integration, is not as straight
forward. For /ft larger than about 2.5 the argument tends
to zero rapidly. However, for^< 2.5 an upper inte-
gration limit must be chosen that is representative of
local aerosol size distributions.
-J I i i i I I 1—1 l_J I,, I, ,1 I l l l l I l l i l L
BETA
FIGURE 2 Ratio of the scattering coefficients measured
at 633 nm and U+2 nm as a function of Beta.
2
5-5
-------
C=FACTOR X B5CAT
LJ
1X10
FIGURE 3 Beta is plotted against a factor which when
multiplied by the extinction coefficient in
km yields the value of c. c is defined in
Equation 7-
FIGURE It A nomogram representing the aerosol mass load
as a function of bscat km-1- for different
values of Beta.
Heasurewents
The output beams of a HeNe (633 nm) and HeCd laser
(U22 nm) were directed through a 100 Mpinhole and a
positive lens (see Figure 5). The pinhole cleaned the
beam of interference patterns introduced by the laser
while the positive lens produced a beam with a diver-
gence of aoout one degree. The divergence of the en-
tire Gaussian profile created an uniformly illuminated
central "spot" of about 6 meters in diameter at 3.1 km.
Both lasers were mounted on a single base that could
be moved using screw adjustments. Additionally, the
HeNe had individual adjusting screws.
TRANSMITTER
LASER -i LENS PINHOLE
MOUNT
'//////t VERT. ADJ.
SCREW
RECEIVER
I MM
PINHOLE
PM
ALIGNMENT
PORT
FIGURE 5
The receiver was a 3 1/2 inch refractor telescope
with a 1 mm pinhole at the focal plane. Appropriate
narrow band interference filters, mounted between the
pinhole and photomultiplier, served to eliminate back-
ground light,. Using these filters, a daytime signal
to noise ratio of 10:1 was achieved.
Since laser outputs tend to vary, the signal
strength was monitored at the transmitter. The output
of the photomultiplier was monitored on a strip chart
recorder that was shunted with a 1 megohm resistor as
well as appropriate capacitors that allowed a variable
time constant.
The calibration of the DBLT is carried out by the
use of a two source telephotometer accurate to < 1%.
The detection equipment includes either a 6-inch re-
flecting telescope or a l^-inch refractor, interference
filters or a Beckman spectrometer, an MI 65260 photo-
multiplier, high-voltage supply, and associated pulse-
counting equipment. '
The basic observational procedure, illustrated in
Figure 6, consists of observing a single incandescent
light source placed alternatively at two different dis-
tances; the extinction coefficients are then determinec
for the intervening atmosphere after the data has been
corrected by the inverse-square law. The coefficients
are expressed as total extinction per kilometer. Dis-
tance separations from 3 to 21 km have been success-
fully used. HallI2 discusses this equipment and pro-
cedure in more detail.
TELEPHOTOMITER
FILTERS
OR
SPECTROMETER
PULSt COUNTER
(
DISCRIMINATOR
POWER SUPPW
jPHOTO" \
MULTIPLIES f-
LAMP
- -i ¦
LIGHT
SOURCE
3 nd
POSITION
3
FIGURE 6
3
5-5
-------
Aerosol mass concentrations were made using the
conventional high-volume samplers while aerosol size
distributions were made using the casella cascade im-
pactor Mk Ila.
Discussion of Results
On April 27, 1975, measurements were made at Lake
Powell with the two source telephotometer, cascade im-
pactor and high volume sampler. Due to equipment prob-
lems, the DBLT was not operated at that time, nor were
contrast measurements obtained due to local cloud cover.
A typical cascade impactor measurement is shown in Fig-
ure 7. The results are representative of numerous other
size distribution measurements^ made in the area. The
slope of this curve is approximately 1.3. There is a
sharp drop in the number of particles per meter cubed
per increment in inr at 10m. thus justifying the upper
integration limit of r2 = 10m. The decrease in the
number density at .kM. is due to instrument design.
3..
2 *
CK
b »
S
1—
3 0*4 0.7 0 S 3 < 5 B ® # ® J OS 70 ffl
RADIUS
FIGURE 7 A Junge plot for a typical number distribu-
tion. A normalized dN/dlnr is plotted against
particle radius (in microns).
Atmospheric extinction coefficient measurements
made on the nights of 14-27-75 and U-28-75, as obtained
from the two source telephotometer, are presented in
Figure 8. The lower portion of the graph represents
the total extinction coefficient while the upper half
shows the extinction due to aerosols.
A measured aerosol scattering coefficient ratio
(b^-jnm/bi j o™) approximately 1 yields from Figure
2 a value of A =» 1-3, a number which is in close agree-
ment with the value measured by using the cascade im-
pactor. Using the value of Figure 3 yields
the value cel.75 x 10~3, a number somewhat less than
the measured value of 2.3 x 10~3. This slight discrep-
ancy is expected since the cascade impactor measurement
was taken over land and near a lightly traveled roadway
(two cars in a 12 hour period), while the transmission
readings were made over Lake Powell. Figure U yields a
value for aerosol loading of about 20«ng/m3. This
value is slightly lower than the measured 32Mg/m3 and
is again probably due to the dust generated by the
slight activity in the localized area of the high volume
measurement.
«.02-
.00-
-.02-
.15-1
a -io
O
,05'
RAYieiGH and OZONE
4-27-75
4-28- 75
ACROSS LAKE POWELL
~r~
.30
—r~
.40
T
—I—
.60
—r
.70
.50
X iJim)
EXTINCTION COEFFICIENT / WAVELENTH
EL. 1.2 KM
FIGURE 8 Results of extinction measurements made at
LaKe Powell on I4.—27 and U-28, 1975 using the
two source telephotometer. These previously
unpublished data were obtained by Hall and
Riley.
Similar measurements using the two source tele-
photometer and DBLT were made at Flagstaff on the night
of August 12, 1975 and the morning of August 13, 1975.
Particle size measurements yielded a/5 i 2.2 while high
volume samples showed an aerosol mass loading of 3Q«g/
m3. Representative strip chart traces as obtained from
the DBLT are shorn in Figure 9-
k
FIGURE 9
The top, middle and bottom recorder traces are
signals processed at the receiver at 633 nm from 11:00
to il:L}5 p.m., kh2 nm from 12:00 to 12 :U5 a.m. and
t>33 nm from 8:30 to 9:15 a.m. respectively. The 2-4
minute five percent fluctuations of the top (633 nm)
5-5
-------
signal are due to spatially and temporally changing
aerosol concentrations (smoke from fire places etc.).
The middle trace (442 ran), taken immediately after the
first trace, shows similar fluctuations in time with
smaller amplitude variations. This should be expected
since a greater portion of the signal is due to Rayleigh
scattering. Tne bottom trace (633 nm) was recorded
during daylight hours and shows slight temporal vari-
ations -- a signal representative of the rather thor-
ough mixing that has taken place. The extinction coeffi-
cient due to aerosol, as measured by the calibrated
DBLT, varied from a maximum of .021 km~^ at 7:00 a.m.
to an averaged minimum of .016 km"-*- at 1:00 p.m.
On Aug. 12, 1975, 1:H5 pm, the measured value of
^33 and was -°16 and -°19 km"1 respectively. The
ratio of b^jj/b^^ was approximately equal to .84 which,
using Figure 2, yields a&z2.k- This number is in :
close agreement with /i -2.2 as obtained from a casella
cascade impactor measurement made at the same time.
The small discrepancy is probably due to the different
air masses that were sampled by the two instruments.
The impactor measurement was made at Lowell Observatory
(located on Mars Hill, which is slightly displaced from
downtown Flagstaff) while the DBLT sampled the air vol-
ume over Flagstaff. The air over Flagstaff is expected
to contain more pollutants and hence would imply a
larger & . Using the value of ft = 2.4, a value of c =
1.6 x 10"? and an aerosol mass loading of 35»«.g/m3 is
obtained from Figures 3 and 1; respectively. Both these
values are somewhat nigher than the cascade impactor
value of c = .9 x 10"' and the high volume measured
30>ug/m''. This discrepancy is again probably due to
the location of the size distribution and aerosol mass
loading measurements. These point measurements were
made at Lowell Observatory while the transmissometer
sampled air over Flagstaff.
Conclusion
References
!-K. D. 01 Dell and R. G. Layton, Plateau 46, 133, (197U).
2w. A. Baum and L. Dunkelman, J. Opt. Soc. Am. 45,
166, (1955).
^D. H. Hohn, Appl. Opt. 14, 404, (-1-975).
%.C. Ahlquist and R. J. Charlson, 3, 551, (1969).
^G. Yamamoto and M. Tanaka, 8, 447, (1969).
^B. M. Herman, S. R. Browning and J. A. Reagan, 28,
763, (1971).
fa. E. Shaw, J. A. Reagan, and B. M. Herman, 12_, 374,
(1973).
8G. Mie, Ann. Phys., 25, 372, (1908).
9h. C. Van de Hulst, Light Scattering by Small Parti-
cles, Wiley, New York, 407 PP, (1957)•
^C. E. Junge, Air Chemistry and Radioactivity, New
York, Academic Press, 382 pp, (1963)
^¦K. W-lleke and K. T. Wiitby, J. Air Pollution Control,
25, 529, (1975).
J. S. Hall and L. A. Riley, Proceedings of AIAA 10th
Thermophysics Conference, 75., 683, May 1975.
^W. C. Malm, E. G. t<her, R. Cudney, To be published
in Lake Powell Research Project Bulletin, NSF-RANN.
Based upon preliminary measurements, it appears
that a dual beam laser transmissometer is a superior
method of monitoring certain air quality parameters.
The advantages of tne above described instrument are
two-fold:
1) "Instantaneous" values for the extinction
coefficient, visibility, aerosol mass load-
ing and aerosol number distribution are ob-
tained from one instrument, and
2) these directly measured and derived air
quality parameters are spatially averaged
(as opposed to point measurements) over
long distances.
Determination of the visual ranged using contrast
techniques has the limitations of cloudless skies and
can only De made during daylight hours -- other trans-
missometer measurements can only be made at night. The
integrating nephelometer, though it gives instantaneous
values of the scattering coefficient, is a point meas-
urement, as are high volume and particle size measure-
ments. Measurements made at one point in space are not
representative of potential air quality changes that
may take place over large distances. The high volume
sampler has the additional restriction of being time
averaged over periods of hours and sometimes days. If
the dynamics of air quality change are to be studied,
a change in aerosol mass loading must be witnessed on
time scales shorter than those currently being used.
Future development of this technique will center
around automating measurements and the use'of various
other light sources (such as mercury lamps);^ that will
afford measurement of the extinction coefficient at
various wavelengths.
5
5-5
-------
DESIGN CONSIDERATIONS AND FIELD PERFORMANCE FOR AN IN SITU,
CONTINUOUS FINE PARTICULATE MONITOR BASED ON
RATIO-TYPE LASER LIGHT SCATTERING
by
Gerhard Kreikebaum
Frederick M. Shofner
Environmental Systems Corporation
1212 Pierce Parkway
Knoxville, Tennessee 37921
ABSTRACT
Design considerations are given for an in situ, ratio-
type, laser light scattering instrument for particle
sizing in the 0.2 < d < 3.0 micrometers size range.
Mie scattering computations show a monotonic response
in the desired sue range and a relative independence
of the particle's index of refraction on sizing. Other
advantages of the ratio technique, including a good size
resolution, are described. The system reported here
operates in situ, up to particle concentrations of 10®/
cm^ and up" to temperatures of 200°C. Example field data
include particle size and mass distributions measured
downstream from an electrostatic precipitator serving a
coal-fired power plant.
INTRODUCTION
Accurate measurement of particle sizes and densities in
process streams and in the ambient atmosphere consti-
tutes a general and fundamentally important problem.
Interest in this problem area is currently increasing
as it is becoming more broadly appreciated that physical
processes involving particulates depend sensitively upon
the particle size distribution. It is further being
recognized that it is desirable to measure particle
size distribution in situ because in many cases it is
very difficult, if not practically impossible, to sample
process stream particulates without modifying the
particle size distribution and/or the physical and
chemical characteristics of the particles. Also, the
advantages of on-Hne measurements are self-evident;
electro-optical instrumentation generates real time
signals for prompt operator evaluation and action and,
ultimately, control signals which may be "fed back" to
improve product output rate and efficiency and reduce
emissions, for example.
THEORETICAL CONSIDERATIONS OF RATIO-TYPE SCATTERING
Hodkinson^ proposed and Gravatt^ investigated and uti-
lized the ratio concept for light scattering measure-
ments of particle size. The technique basically
involves the simultaneous measurement of the intensity
of light scattered by a single particle at two angles
in the forward scattering lobe. The major advantages
which accrue from the ratio approach are: (1) A mono-
tonic response in the important size range around 1
micrometer using a practical source with a wavelength
of about 1 um and a practical angle pair. (2) Relative
Independence of refractive indices. The technique
makes use of the fact that, for any given particle
diameter, the angular distribution of scattered inten-
sity within the forward lobe changes little as a func-
tion of refractive index, whereas the scattering
intensity at any one angle varies more with refractive
index. Thus, particles can be sized more independently
of refractive index than by a single angle measurement.
(3) Independence of variations in particle illumination
which are attributed to laser power fluctuations and
position of the particle within the laser beam.
Ratio Method: Particle Diameter Range
Figure 1 shows the ratio of scattering coefficients for
a number of angle pairs. The particle diameter is cal-
culated for a laser wavelength of 0.904 un>. These
ratios are calculated from the exact Mie theory of light
scattering my homogeneous spheres and in this case for
particles whose imaginary component of the refractive
index is larger than about one; that is, for strongly
absorbing particles.
O
z
tt
Ui
F 0.6
h-
3
£0.4
O
| 0.2
PARTICLE DIAMETER, micrometer
Fig.l. Ratio of scattering coefficients versus particle
diameter for various combinations of scattering angles.
The diameter range covered by any angle pair is mainly
determined by the larger angle since the ratio goes
through a minimum at a diameter'for which the scatter-
ing intensity at the larger angle goes through its
first minimum. The largest particle diameter which
can be measured by the ratio method depends on how
close to the forward direction two simultaneous measure-
ments can be made without interference from the illumi-
nating laser beam. The lower limit is dictated by the
largest angle at which enough scattered light can be
collected for an adequate signal to noise ratio.
In order to size particles with diameters up to 3 ym
the 14° and 7° angle pair was selected. This diameter
range was chosen because the so-called fine particulate
with diameters below 3 um is respirable and thus of
major concern.
Index of Refraction Dependence
Detailed calculations have been carried out for the
selected angle pair (14° and 7°) in order to show how
the scattering intensity ratio depends on the refrac-
i.o
3.0
4.0
2.0
50
1
5-6
-------
tive Index. The results show that the ratio is inde-
pendent of refractive index for all highly absorbing
material with an imaginary component or absorption co-
efficient n2 > 1 of their refractive indices n=n]-in2-
Figure 2 illustrates the effect of the imaginary re-
fractive index component for a real component of
n-| = 1.6. The ratios oscillate around the monotonical-
ly decreasing ratio for highly absorbing materials.
The amplitude of oscillation is largest for non-absorb-
ing particles and the maximum sizing error introduced,
if the ratio for highly absorbing particles is used for
sizing, is about 20% in this case. Possible sizing
errors are reduced as the imaginary refractive index
component increases. For example, for a refractive
index of n = 1.6-1 0.05 the maximum sizing error is
only about 10%. This behavior is similar for other
real components of refractive indices.
f-
z
bJ
1.0
O
Lu
U.
Ul
o
o
o
2
0 8
0.6
-
cc
\~
J—
<
o
CO
0.4
-
U-
o
0.2
-
o
1—
§
Figure 3 shows a block diagram of the electro-optics.
The collection optics are of an annular design in order
to collect all light scattered into the two forward
angles.
X" (V904^r*
COLLECTION DETECTORS
OPTICS AMPLIFIERS ratio
CIRCUIT
SIZE
ANALYSIS READ-
CIRCUIT OUT
LASER
AND
OPTICS
SENSOR HEAD | CONTROL UNIT
Fig.3. PILLS IV electro-optical block diagram.
METER
CONTROL
ELECTRONICS
^-SAMPLING
VOLUME
PARTICLE DIAMETER, micrometer
Fig.2. Ratio of scattering coefficients versus particle
diameter for n = 1.6-in2-
DESIGN CONSIDERATIONS FOR AN IN SITU. FINE PARTICULATE
SIZING/CONCENTRATION MONITOR
Design Objectives
1. Sizing Range: 0.2 urn < d < 3.0 pm.
2. Particle Density: < 106/cm3, for d > 0.2 urn.
3. Free Stream Gas Velocity: 2 to 20 meters/second.
4. Stream Gas Temperature: < 200°C.
Also, the instrument sampling probe had to fit into a
4" sampling port, be field reliable, and portable by
normal men.
Electro-Optics
Besides the angle pair (14° and 7°) to cover the 0.2
to 3.0 ym size range, angular ranges of 1° were chosen
to give an adequate signal to noise ratio for the
smallest particle size. For in situ particle sizing
the scattering volume, that is, the volume external to
the instrument from which scattered light can be
detected, has to be small enough so that there is only
a small probability of the presence of more than one
particle in the volume at the same time. To meet this
requirement for particle densities of up to 106/cm3 a
scattering volume of about 2 x 10"7 cm3 was selected.
The volume is totally optically defined by the beam
diameter of about 20 um and in axial direction by the
"fields of view" of the light collecting optics.
Fig. 4. PILLS IV fine particulate size and concentration
monitor.
Mechanical
Figure 4 shows a photograph of the in situ particulate
monitor designated PILLS IV. The device is an exten-
sion to the PILLS (particulate Instrumentation by L_aser
Ljight Scattering) technology3'^ to fine particulates.
The instrument's sensor is'contained in the end of a
3-1/2" O.D. stainless steel tube. The particulate laden
gas stream flows through the slotted section of the tube
and the sampling volume is located halfway between the
cones protruding into the slotted section, essentially
in the unperturbed gas stream. The optical transmitter
and receiver are located on either side of the slotted
section and need to be protected from the high gas temp-
erature and from the particulates. The cooling is ac-
complished by a large volume flow of filtered air inside
the tube. A small portion of the air is used for purg-
ing the optical surfaces. To ensure that the optically
defined scattering volumes coincide for both angles,
even for severe temperature changes, special attention
has to be given to the mechanical design of the trans-
mitter and receiver.
2
5-6
-------
In order to make the system easily portable the stain-
less steel probe has quick disconnect provisions from
its support case which houses a filter, a blower and
the supporting electronics for the sensor head.
PROCESSING AND DftTft PRESENTATION
All data processing is accomplished in a separate unit
which is connected to the sensor package by an umbil-
ical cable. The data presentation on this unit is a
hard copy print-out of particle size distribution.
Some of the unique features of the all digital unit are
32 pre-programmed as well as arbitrarily field-settable
particle sizing ranges and a number of analog outputs
of particle count rates in any selected size range.
Since more than one sensor can be interfaced with the
signal processing unit the system can be used, for
example, for real time evaluation of the fractional
efficiency of control equipment.
Although the ratio technique is theoretically indepen-
dent of the light intensity incident on a particle,
ratios formed from small scattering yields can be in
error due to system noise. Therefore, the data proces-
sing unit includes a ratio qualifier circuit, which
only passes a ratio for further processing if the ratio
originates-from a large enough scattering yield. The
magnitude of the minimum scattering yield depends on
the ratio itself. By this means, the ratio approach
covers, for example, the 0.75 < d < 3.0 um range. Also
because of system noise, particle sizing in the lower
area of the total sizing range is based on scattering
at one angle only. This is necessary since the ratio
in this range is rather insensitive to size (see
Figure 4). Indeed this fact is used to qualify the
single angle sizing which covers the 0.2 < d < 0.75 um
ran9®- — RATIO
1.0
as
0.9
0.7
os 0.3 ai
i i r i i I"
-id-0.21 /im.FWHM-
0.9
1.0
2.0 3.0
PARTICLE DIAMETER , micrometer
F1g.5. PILLS IV response function.
LABORATORY AND FIELD DATA
Figure 5 illustrates the size resolution of the instru-
ment. The figure shows the system response to mono-
dispersed polystyrene latex spheres of 2 ym 1n diameter.
We believe that this kind of size resolution for an
in situ system can best be obtained with a ratio ap-
proach where the effects of varying particle Illumina-
tion as a function of particle position in the scat-
tering volume are greatly reduced. The size resolu-
tion Ad/d (FWHM of the response function r(d) ) 1s
typically no larger than + 203! 1n any part of the size
range covered. ~
The entire system has undergone extensive field testing
and the sensor has been exposed to all of the severe
conditions for which it was designed without any notic-
able degradation in performance.
10
*i§
TJ
s
10'
SIZE
DISTRIBUTION
»
C
§
i.
10'
H s
10J
MASS
distribution
J
0.3 OS I 2 3 8
PARTICLE DIAMETER, micrometer
Fig. 6. Example particle size and mass distribution.
TOTAL GRAIN LOADING
jrolnt / ACf
0.5 1.0 2.0 3.0
PARTICLE DIAMETER. micrometer
Fig. 7. Example mass distribution.
Some example field data are given 1n Figures 6 and 7.
Figure 6 shows particle stee distribution and mass
distribution downstream from an electro-static precip-
itator 1n a coal-fired power plant. The data in Fig-
ure 7 was obtained from a PILLS IV system built for
the EPA and generated in a wind tunnel test facility
at EPA, .Research Triangle Park, In July, 1975. The
tunnel contained a controlled particulate loading using
redispersed fly ash obtained from a precipitator on a
5-6
-------
coal-fired power plant. The figure shows particle mass
concentration for various total grain loadings in the
tunnel. As expected, most of the total mass is con-
tained in sizes larger than 3 urn. Of interest is the
small mass fraction in the 0.5 to 1.0 ym size range.
CONCLUSIONS
Design considerations as well as field performance of
an in situ, continuous, fine particulate monitor based
on ratio-type light scattering have been presented.
The principal advantages of the ratio techniques are a
monotonic response in a size range determined by the
scattering angle pair and the light wavelength used and
a relative independence of the particle's index of re-
fraction. Extensive Mie scattering computations show
these features, particularly for optically absorbing
materials. Another major advantage is the independence
of laser power fluctuations and of particle position in
the scattering volume, which is most important for an
in situ measurement where the particles cannot be intro-
duced into the laser beam at a preferred position and
which is necessary for a good response function result-
ing in a good size resolution.
Design considerations were given for an instrument to
cover the important size range of respirable particles
below 3 um in diameter. Using an infrared laser, the
0.2 to 3.0 pm range can be covered selecting the 14° and
7° forward scattering angles. The proper choice of
angles and wavelengths is important since the ratio
technique has a relatively limited particle sizing
range.
Other, design features are an air cooling and purging
concept to protect the sensor from a particulate laden
gas stream at elevated temperature. Field data show
that the instrument performs reliably under the extreme
temperature and particle loading conditions for which
it has been designed.
ACKNOWLEDGEMENTS
The authors would like to thank Mr. William Kuykendal
with the Environmental Protection Agency, Research
Triangle Park, North Carolina, for many helpful discus-
sions as well as for the permission to publish data
obtained by EPA with the PILLS IV instrument.
REFERENCES
1. Hodkinson, J. Raymond; "Particle Sizing by Means of
the Forward Scattering Lobe", Applied Optics, Vol.5,
No. 5, May 1966.
2. Gravatt, C. C. Jr., "Real Time Measurement of the
Size Distribution of Particulate Matter by a Light
Scattering Method", J. of the Air Poll. Control
Assoc., Vol. 23, No. 12, December 1973.
3. Shofner, F.M., Watanabe, V., and Carlson, T.B.,
"Design Considerations for Particulate Instrumenta-
tion by Laser Light Scattering (PILLS) Systems",
ISA, Vol. 12, No. 1, 1973.
4. Shofner, F.M., Schrecker, G.O., Carlson, T.B., and
Webb, R. 0., "Measurement and Interpretation of
Drift Particle Characteristics", Cooling Tower
Environment - 1974 Symposium, March 4-6, 1974.
4
5-6
-------
NEW ANALYTICAL PROSPECTS FOR AN OLD DETECTOR
Lawrence L. Altpeter, Jr.
North Star Division, Midwest Research Institute
Minneapolis, Minnesota
SUMMARY
Examination of the current pulses produced by
individual particles in a conventional flame ionization
detector (FID) has led to the development of a novel
multi-electrode FID. In this device, the individual
particle appears to provide its own event marker as it
is consumed in the flame. This marker pulse is
followed by another current pulse which may be due to
the actual charge carriers, delayed by their time of
flight from the flame to the electrode. The objective
of this work has been to determine whether or not
individual particles produce ionization current signa-
tures which can be related analytically to particle
size and/or composition.
In addition, a performance study has been con-
ducted with the conventional FID in which analytical
parameters such as size limits of detection, response,
sensitivity, and selectivity have been evaluated.
INTRODUCTION
The ideal atmospheric pollution monitor would per-
form on a real time and continuous basis while provid-
ing interference-free analytical data for each pollut-
ant. In the case of aerosol pollutants, the analytical
data might include particle characteristics such as
number concentration, size, and composition.
The use of the flame ionization detector (FID) as
a real time and continuous monitor for gases and vapors
at the trace level is well-known. Similar use of the
FID for monitoring aerosols has received relatively
little attention.^ This is because: 1) the device is
count-rate limited; 2) its response varies with compo-
sition, so that accurate size information would not
appear possible.
In a series of experiments conducted at this
laboratory, conventional and novel-design FID's were
challenged with various homogeneous, monodisperse
aerosols covering the size range 0.5 to 10 pm. High
speed waveform studies were performed by connecting a
wideband amplifier to the ground side collection
electrode and observing the output current pulse from
individual particles as they were consumed in the FID.
In the detector configurations we have tested, this
current pulse has a peak amplitude of 10~® to 10-6
amperes and a duration of several milliseconds. When
the electrode geometry was parallel plates,the current
pulse was observed to be modulated. With the help of
a multi-electrode FID, or so-called MFID, this modu-
lated current pulse appeared to be a composite signal
in which the charge carrier current and its time of
flight could be observed.
EXPERIMENTAL
Apparatus
Conventional FID Detectors. The conventional FID
employed throughout these studies is shown schemati-
cally in Figure 1. Significant features included
parallel plate electrodes, a pre-mixed stoichometric
air-hydrogen flame, a ground-side wideband amplifier
and removal of combustion products by evacuation. A
ceramic burner with a 1 mm orifice was centered be-
tween two stainless steel parallel plate electrodes
which were separated by 1.6 cm. Gaseous flow rates
EXHAUST
SIGNAL OUT
±BIAS
-5-VOLTAGE
t
SAMPLE
IN
Figure 1. Schematic Illustration of
Conventional FID Aerosol
Monitor (Parallel Plate)
included 100 cc/min hydrogen, 250 cc/min aerosol-laden
air, and 300 cc/min secondary air. A polarizing voltage
of 500 VDC was applied. Details of the actual working
configuration have been shown elsewhere.6
Multi-Electrode Flame Ionization Detector (MFID).
A functional sketch of the first generation MFID is
shown in Figure 2. The MFID was designed to examine
EXHAUST
rfii
SELECTABLE
INPUT
BUS
3 r BIAS
VOLTAGE
IMP
[secondary air
SAMPLE IN '
Figure 2. Schematic Illustration of MFID
the cloud of charge carriers being detected by the
conventional FID. The MFID was identical to the
conventional FID with the exception of the collection
electrode. Within the same overall area, the collec-
tion electrode of the MFID was transformed into a
mosaic of electrodes. Each electrode was insulated
from its neighbors and individually addressable. Those
electrodes not in use were grounded. The effect was
that of producing an overall electric field identical
to that employed for the conventional FID while per-
mitting one to examine the spatial distribution of ion
currents impinging on the surface of the collection
electrode of the conventional FID. Details of a
working model have been presented elsewhere.®
SECONDAR
1
3-7
-------
Experimental System
A schematic layout of the experimental system em-
ployed in these studies is shown in Figure 3. Briefly,
.EXHAUST
SCOPE VISICOROER
hfe
(SIGNAL!
'2 [TRIGGER)
SAMPLE
FOR
AUXILLARY
BENCHTOP
FID
STUDIES
SHUT OFF VALVE
INTAKE AIR
AIR— FILTER
'/V
-/v
Figure 5.
Schematic Layout of Experimental
System for Evaluation of FID
'V
the FID was centered in an atmospheric simulation
chamber which had a volume of 3.7 (128 cubic foot).
A Berglund-Liu Monodisperse Aerosol Generator (Model
3050, Thermo Systems, Inc.) produced homogeneous,
monodisperse (og - 1.067) aerosols which were blended
with clean air and delivered to the chamber to produce
aerosols of controlled size, number concentration, and
composition. Amplified signals from the FID were
monitored visually with an oscilloscope while a perma-
nent record was obtained with an oscillograph, connected
In parallel with the oscilloscope. Further details
have'been presented elsewhere.6
Sample Preparation
Aerosol samples were prepared in three categories:
inorganic, organic, and biological. The latter con-
sisted, of a set of commonly occurring microorganisms.
Inorganic and organic aerosols were prepared with de-
ionized and distilled water which was doubly filtered
through Whatman 40 paper and Millipore 0.45 ym paper.
Biological aerosols were removed from their nutrient
broth by centrifugation and then resuspended in tap
water which was dechlorinated, sterilized, and doubly
filtered as described above. These aerosols were
measured promptly to minimize the extent of lysis.
RESULTS
FID Output Signals
The following set of figures illustrate some
preliminary results for certain selected aerosols.
Each waveform represents the output from a single
particle. A calibrated grid appears in which the dis-
tance between each horizontal line as well as each
vertical slash mark is one centimeter. The triggering
level control of the oscillograph was set to record
only those signals which exceeded the noise level.
Consequently a portion of the leading edge of each
waveform is missing.
The results for several different metallic salts
are shown in Figures 4 through 6. In the case of Csl,
the current pulse is devoid of modulaton as shown in
Figure 4. The ordinate is presented as millivolts per
cm, the control setting of the oscillograph. It can
be seen that ample signal is produced by the 4 pm
particle. Duration of the event is 2 milliseconds,
approximately. Data for NaCl (Figure 5) reveals a
sharp leading peak with a single trailing shoulder.
Again, good signal-to-noise is apparent and the
Figure 4. FID Output Signals from 4 pm Csl
Aerosol. Oscillograph Sensitivity
200 mV/cm. Time Base 1 msec/cm.
Figure 5. FID Output Signals from 4 pm NaCl
Aerosol. Oscillograph Sensitivity
100 mV/cm. Time Base 1 msec/cm.
Figure 6.
FID Output Signals from 4 um BaClj
Aerosol. Oscillograph Sensitivity
20 mV/cm. Time Base 1 msec/cm.
5-7
-------
duration of the event is 2 milliseconds. Moving from
Group IA to Group IIA of the periodic tahle can
produce several changes as shown in Figure 6 for BaCl2.
The signal now appears as a leading spike followed by
two shoulders. The signal amplitude is greatly re-
duced, reflecting perhaps the higher ionization poten-
tial of barium. Duration of the event remains un-
changed.
It turns out that the waveform produced by BaCl2
was observed for many other species. Thus, an organic
aerosol Such as latex (Figure 7) revealed a similar
contour composed of a leading spike and two trailing
shoulders.
, V
iv
Figure 7.
FID Output Signals from 0.8 ym Latex
Aerosol. Oscillograph Sensitivity
20 mV/cm. Time Base 1 msec/cm.
Figures 8 through 10 display the output signals
for several microbiological aerosols. Sax-ratio,
mavcesaens, 1 pm in size approximately, is shown in
Figure 8, Data from somewhat larger organisms such as
that of a yeast (Saoaaromyoes aeveviaiae) and a mold
(/\apepgillu8 nigef) are shown in Figures 9 and 10,
respectively.
Figure 10.
FID Output Signals from A.n. Aerosol.
Oscillograph Sensitivity 500 mV/cm.
Time Base 1 msec/cm.
FID Response, Sensitivity, and Detection limits
The response of the FID to different kinds of
aerosols was examined over the size range from 0.5 to
10 urn. Sodium chloride and latex were selected
arbitrarily as examples of inorganic and organic
aerosols. Similarly, several microorganisms were
selected to span a size range for microbiological
aerosols. Twenty output signals were recorded for
each particle size of a given aerosol specie. Detector
response was measured by the area under the curve for
each current pulse. The results for NaCl and latex
are shown in Figure 11.
Figure 8. FID Output from S.m. Aerosol (0.5 to
1.0 vm by 1.0 Mm. Oscillograph
Sensitivity 20 mV/cm. Time Base
1 msec/cm.
:l
Figure 9.
FID Output Signals from S.o. Aerosol (4
to 5 ym approximately). Oscillograph
Sensitivity 200 mV/cnu Time Base 1 msec/
AERODYNAMIC SIZE I Jirtl)
Figure 11. Response of Conventional FID (Parallel
Plate) as a Function of Aerosol
Particle Size
3
5-7
-------
In the case of NaCl, the results of duplicate
runs on two separate days are shown for each size that
was tested. It can be seen that on a log-log format
that response Is linear. A linear regression analysis
of the data yielded the equation:
log10Y = 0.72 + 1.54 log1QX
a)
with a correlation coefficient of 0.996. The error
bar at 4 um for NaCl is the result of a precision
study presented below. By extrapolation the lower
size limit of detection for NaCl was 0.2 vm while
that of latex was 0.15 pro. In the case of NaCl, there
is good agreement with the results of Crider and
Strong.7
The results for microbiological aerosols are
summarized in Table 1. Among six common linearly
transformable functions, a best least-square fit was
obtained with the function:
• 04X
Y = 14.8e"
(2)
The slope of the response curve, or the sensitiv-
ity, was nearly 3/2 in the cases of organic and in-
organic aerosols (see Equation 1).
Table 1. Summary of Microbiological Mass Response
Data from the FID
Microbiological
Aerotiol
Mean Volume
(um3)
Mean Area
Under the
Curve (cm2)
Relative
Standard
Deviation
S, mapoeaoens
0.4 + 0.4
15.1
15%
P. vulgaris
4.1 + 3.3
19.0
18%
8. oereus
15.0 + 7.7
23-0
20%
S, aeveviaiae
90.1 + 74.5
420.0
25%
FID Precision
A see of twenty output signals was measured each
day for six days for a 4 p NaCl particle. Analysis of
Variance6 revealed a 16.8 percent analytical error and
2.2 percent sampling error. Due to errors such as the
planimetric data reduction, it would appear that a
figure of 5 to 15 percent would be representative of
the repeatability of the FID. The small sampling error
indicates that good control was achieved over genera-
tion and delivery of the aerosol to the experimental
chamber. The 4 um data point itv Figure 11 has been,
enclosed by 2 a error bars.
MFID Output Signals
Data was collected from the MFID by connecting
the amplifier to the lowest electrode (Electrode 1),
and recording a series ot signals for an aerosol of a
given size and species. The amplifier was then
connected to the next electrode (Electrode 2) and the
process repeated. The results of such a study are
shown in the hand-traced data of Figure 12 for a 4 pm
NaCl particle and Servatia Mareescens.
Each MFID electrode develops a part of the total
signal seen by a conventional FID. At first glance,
however, there does not appear tr> he any relationship
between the MFID data and the FID data for the species
shown. ,
The MFID data was uniquely characterized by the
presence of bipolar features in the wave forms. These
signals predominated at the fringe electrodes furthest
f\
I
'!\
Figure 12. Tableau of Representative Output Signals
from Collection Electrode Segments 1 to
13 of the MFID, as Produced by Individual
Particles from Separate Aerosols of NaCl
and S. maroesaens (hand traced from
original data)
from the flame. They appeared to develop continuously
from the start of the waveform. A closer inspection of
the MFID data revealed the presence of a second signal
component which was most apparent at those segments
directly across from the flame. This component can be
seen most clearly, in the case of the Serratia
mavcesoens aerosol, at Electrodes 2 to 4, as a trailing
second peak. A similar signal can be observed for the
4 ym NaCl aerosol at Electrodes 2 to 6—although not
as clearly,
DISCUSSIOH
Of particular interest during these studies was the
selectivity of the FID, i.e., the extent to which the
waveform of the output current pulse could be uniquely
related to composition. Of the tested electrode
geometries, only from the parallel plate electrodes
were signals obtained with discrete information which
might serve potentially as a "fingerprint" for identi-
fication. These signals were composed, generally, of
a leading spike with one or two trailing shoulders.
Organic and biologic aerosols produce pulses with two
shoulders while certain inorganic aerosols like NaCl
produce pulses with marked diminuation or disappearance
of one or both shoulders. Of possible practical signi-
ficance were certain biological aerosols like yeast
(Saoaharomyces cereoisiae) and a mold (.Aspergillus
ntgev) which produced unique waveforms. Presence of
these aerosols in the air might be detected by an FID
equipped with a suitable pattern recognition system.
The electrodes of the FID, described above, form
a parallel plate capacitor with the flame centered
between the plates. As a particle is consumed 1n the
flame and the resultant charge carriers are separated
by the field, there occurs, effectively, a transient,
localized change in the permittivity of the medium
between the electrodes. The result of such an "upset"
event is a transient, fluctuation in the field and, also,
in the steady state charge accumulated at each electrode,
or plate. During, or following this process, the
created charge carriers then arrive at their respective
5-7
-------
electrodes. The resultant output current pulse would
be the algebraic sum of these two effects. If one
takes a typical mobility of positive ions in ambient
air as 1 cm2/volt-sec,8 and allows for the effect of
tha elevated temperatures within the FID enclosure
upon the mean speed, then at 500 VDC and a flame-to-
plate distance of approximately 7 to 8 mm, a time-of-
flight of less than 3 milliseconds is reasonable.
This is consistent with the duration of the FID pulse.
The net current from the positive electrode of
the FID can be described, for the simple one dimension-
al case, by the expression:
I(t)~ = Aco ^ - Aq~b E(o,t)n(o,t)
where I(t)~ is the net output current at any time "t",
"A" is the area of the collection electrode (cm2), Eo
is the permittivity (coul/volt-cm), E(o,t) is the
electric field component normal to the positive elec-
trode surface at time "t", b~ is the mobility of the
negative charge (jarriers (cm2/volt-sec) and n~ is the
number concentration at the surface of the positive
electrode (cm-3). The first term refers to the current
associated with the changing field at the electrode
surface. The second term arises from the neutraliza-
tion of the gaseous charge carriers as they arrive at
the electrode surface. Consistent with the principle
of electroneutrality, the net current at the collec-
tion electrode, X(t) , must be equal to that at the
polarizing electrode, I(t)~.
If one calculates the variation of the electric
field with time at the surface of the electrode, then
the first term in the above equation can be plotted
as shown qualitatively in Figure 13.
FIELD-INOUCED CURRENT
¦COMPOSITE CURVE
-GASEOUS ION CURRENT
the field-induced current predominates at the peripheral
electrodes while the ion current is restricted to the
central set of electrodes. This is reasonable since one
should expect the charge carrier current to be greatest
at those electrode segments in a direct line of flight
from the flame.
It should be noted that an enormous amount of data
is produced by the MFID from a single particle. The
sum total of all these data could, itself, serve as a
fingerprint for particle composition.
BIBLIOGRAPHY
1. Ohline, R. W. , "Hydrogen Flame Ionization for the
Detection and Sizing of Organic Aerosols," Anal.
Chem., 37, No. 1, 93 (1965).
2. Ohline, R. W., Thall, E., and Oe.y, P. H., "General
Considerations Concerning Atmospheric Aerosol
Monitoring with Hydrogen Flame Ionization Detector,"
Anal. Chem., 41, No. 2, 302 (1969).
3. Frostling, H., and Lindgren, P. H., "A Flame Ioniza-
tion Instrument for the Detection of Organic
Aerosols in Air," J. Gas Chromatography, 4, 243
(1966).
4. Crider, W. L., and Strong, A. A., "Flame Ionization-
Pulse Aerosol Particle Analyzer (FIPAPA), "Rev. Sai.
Inst,vim., 38, 1772 (1967).
5. Bolton, H. C. McWilliam, I. G., "Pulse Characteris-
tics of the Flame Ionization Aerosol Particle
Analyzer," Anal. Chen., 44, No. 9, 1575 (1972).
6. Altpeter, L. L., Jr., Pilney, J. P., Rust, L. W.,
Senechal, A. J., Overland, D. L., "Recent Develop-
ments Regarding the Use of a Flame Ionization
Detector As an Aerosol Monitor," in "Symposium on
Fine Particles", Minneapolis, Minnesota, May 28-30,
1975. Academic Press, New York, 1975 (in press)
7. Berglund, R. N., and Liu, B. Y. H., "Generation of
Monodlsperse Aerosol Standards," Env. Sai. Tech-
nology, 7, No. 2, 147 (1973).
8. Von Engel, A., Ionized Gassae, 2nd Edition, Oxford-
at-the-Clarendon Press, (1945).
Figure 13. Sketch of Component Current Output
Signals from a Parallel Plate FID
Also shown in Figure 13 is a qualitative sketch
of the charge carrier current (dotted line), which is
the second term in Equation 4. The arrival of the
gaseous charge carriers at the positive plate would
be delayed because of the time of flight factor.
Although qualitative, the presentation in Figure
13 is consistent with actual data from both the FID
and the MFID. Thus, the composite curve of the over-
all current in Figure 13 can be compared with the FTD
data of Figure 6 or 7. The bipolar waveform of
Figure 13 for the field-induced current can be com-
pared with those observed in the MFID data of
Figure 12. In the MFID data of Figure 12 these bipolar
waveforms were observed most at the peripheral elec-
trodes, i.e., those above and below the flame. Many
of the electrodes closest to the heart of the flame
showed waveforms with a second peak. This is readily
observed for S.m., less so for NaCl. Consistant with
the previous discussion, the second shoulder could he
assigned to the gaseous charge carrier current. Thus,
5 5.7
-------
MODIFYING FACTORS IN METAL TOXICOLOGY
David H. Groth, M.D.
National Institute for Occupational Safety and Health
1014 Broadway
Cincinnati, Ohio 45202
Introduction
One of the missions of the National Institute for
Occupational Safety and Health is the setting of maxi-
mum permissible levels or standards for toxic chemicals
in the working environment. These levels of exposure
must not have a detrimental health effect on the worker
during his lifetime, which might include 45 years of
exposure at his job. Some of the data utilized to
arrive at these levels is derived from retrospective
epidemiological studies on humans. Several different
population groups are studied and effects are compared
and correlated with estimated exposure levels.
What is generally found to be the rule is that not
all people exposed to the same concentrations of sus-
pected toxic materials contract the same chronic
disease; that the incidence of chronic diseases in two
different groups of people is not the same, although
exposures are similar; and that the incidence of
chronic diseases in two different groups will vary des-
pite the fact that concentrations of the suspected
etiologic agents are the same in their tissues and
fluids. In other words, different people respond dif-
ferently to the same exposure levels of toxic sub-
stances and the biological tissue and fluid concentra-
tions of the toxic substances do not always reflect
the gtate of health of the individual. In addition,
complete analyses of the substances to which these
people are exposed are rarely done, and therefore, we
do not know how comparable the exposures really are.
As a consequence, epidemiological data can be used to
identify toxic environments, but in cases where the
disease is produced only after several years of ex-
posure, e.g., cancer, a specific chemical can rarely
be implicated.
In order to arrive at some objective approach to
solving this problem, animals are exposed to various
levels of the toxic substance, and the lowest level
that produces a significant effect is the level below
which a standard is set. Yet in many instances there
is a big difference between the results from the
animal experiments and the human epidemiological data,
not only in terms of dose, but in terms of the disease
produced. It becomes apparent that a more precise
method for setting standards is needed. In order to
develop that method, the variables that effect the
toxicology of metals must be considered.
Variables Effecting Toxicity
The following factors have been shown in animal
experiments and can be deduced from animal experiments
and human exposures to have an effect on the toxicity
of chemicals: genetic, metabolic, sex and age differ-
ences, time of day of exposure, the concentration of
the toxic substance, the duration of exposure, the
route of administration, and the relative concentra-
tions of other interacting chemicals (whether in the
diet or in injections or inhaled). When the exposure
is to inhaled particles, then the particle properties,
e.g., size, shape, density, solubility, surface and
subsurface chemistry and surface area influence the
toxicity. In the case of metals, the compound in which
it occurs will also effect the toxicity.
In the presentation I will not attempt to illus-
trate how all these variables effect the toxicology
of metals, but will,select two areas of interest to me,
and 1 hope to you: the areas of interacting chemicals
and particle properties.
Interacting Chemicals
A metal must Interact with other chemicals in the
biological system to produce its toxic effect. The
specific chemical or chemicals with which it reacts to
produce its toxic effect is rarely known. However, it
is known that in some instance other chemicals given
simultaneously will produce synergistic, additive or
antagonistic effects. In recent years much research
has been performed to show how essential dietary ingre-
dients, principally inorganics, can prevent the toxicity
of certain metals. This approach to the study of the
toxicology of metals has several benefits. Experimental
results help us to understand: (1) why some people are
more susceptible than others, (2) the mechanism of
action of the toxic metal and (3) suggest means of pre-
venting diseases (by supplementing the diet with essen-
tial nutrients). To illustrate these points, examples
of how several essential elements modify and prevent
the toxic effects of lead, cadmium, mercury and nickel
will be presented.
Lead
For many years toxicologisis have regarded dietary
lead as only moderately toxic based upon animal experi-
ments. The environmentalists, however, believed that
lead poisoning in children occurred at much lower expo-
sure levels than those required in animals. This dis-
crepancy was resolved in 1973 when Mahaffey and Goyer*
published their work showing that when the calcium con-
centration in the rat diet was reduced from 0.7% to
0.1% (a level which is closer to the 0.05% of human
diets), the toxicity of lead Increased tenfold. The
criteria for the toxic effect were the concentration
of urinary delta-aminolevulinic acid, renal weight
and the concentration of lead in the femurs and kidneys.
This now puts lead into the category of a more highly
toxic substance, particularly in population groups
where the calcium concentration in the diet is as low
as 100 ppm. Other investigators have found that die-
tary iron and zinc can also reduce the toxicity of
lead.
Cadmium
Schroeder has shown that rats fed either 0.6 ppm
or 5 ppm cadmium develop high blood pressure within
12 months, and that rats made hypertensive with in-
jected cadmium can be successfully treated with a zinc
chelate^. Cadmium Injections produce testicular
atrophy and lethal effects on placentas and fetuses ,
which can be prevented by the simultaneous injection of
zinc . Schroeder® also found that people dying of
hypertension had a higher ratio of cadmium to zinc in
their kidneys than did people dying from accidents and
other diseases. Cadmium also effects copper metabolism.
When copper is present in the rat diet at a concentra-
tion similar to that in the human diet (2.6 ppm) only
1.5 ppm dietary cadmium is necessary to reduce plasma
ceruloplasmln . The element, however, which produces
the greatest protective effect at low doses is the
essential element selenium. Simultaneous Injections
of selenium will prevent testicular necrosis*®, toxemia
1
6-1
-------
of pregnancy, placental necrosis and maternal and fetal
death1*, and teratogenic effects produced by cadmium.
Hypertension induced by 2.5 ppm Cd in the drinking
water can be prevented by simultaneous administration
of only 0.9 ppm Se1^. Selenium compounds are more
efficient in protecting from cadmium toxicity than
are zinc compounds. Whereas only equimolar amounts
of selenium are needed, zinc is usually required at a
molar concentration 50 times that of cadmium.
Mercury
As early as 194,1^ in an experiment to test the
efficacy of possible anti-cancer compounds, it was
found that an organic compound containing mercury and
selenium was non-toxic even though kidney concentra-
tions of selenium were extremely high. More recently
Parizek showed that the simultaneous injection of equi-
molar amounts of selenium protected rats from the
LDioo dose of meicury1^. Then in 1969 Eybl, et. al.1^
showed in studies with radioactive labelled mercury
that selenium and also tellurium could dramatically
increase the biological half-time of mercury in
tissues. In 1973 Groth, et. al.16 reported that the
simultaneous administration of selenium in drinking
water prevented the chronic nephritis produced in rats
by mercuric chloride in a 22-month experiment. Al-
though the toxicity of mercury was reduced by selenium,
the concentration of mercury in the tissues was in-
creased. When mercury was given alone at 50 ppm in the
drinking water, the concentration of Hg in the liver
and kidneys was 2.4 ppm and 100 ppm respectively, where-
as when selenium was given with the mercury, the con-
centration of mercury in the livers was increased 125-
fold to 300 ppm and in the kidneys 5-fold to 500 ppm17.
It was found that the increased concentrations could
be explained on the basis of the formation of a rela-
tively stable Hg-Se complex which could be seen as
particles in macrophages in various tissues including
the livers, lymph nodes, spleens, lungs, and kidneys,
and also as intranuclear inclusion bodies in the renal
proximal convoluted tabular cells . The particles
were approximately 50A in diameter and aggregated into
masses that were easily visible by light microscopy.
Microprobe analyses of the particles showed that the
elements Hg and Se were present in a molar ratio of
1:1. It was further reported that these particles
and intranuclear bodies were still present in the
tissue 8 months or more after cessation of exposure
and it was postulated that this form of Hg was probably
the form that is responsible for the very long half-
time of mercury in animal tissue. Since then studies in
in rats have shown that inorganic selenium protects
from methyl mercury poisoning and that tuna fish (which
contained 3-5 times as much selenium as Hg) also pro-
tected from methyl mercury poisoning1®. In addition,
tellurium and the combination of selenium and tellur-
ium have been found to protect from mercury-induced
chronic nephritis in rata1^.
Nickel
Sundermann under contract from NIOSH has shown
that particles of Mn metal can inhibit the production
of malignant tumors in rats resulting from the in-
jection of particles of IU3S2 •
Implicit in this research is the theory that toxic
metals produce their toxic effects by interfering with
biological processes in which the essential elements
are key factors, and that the absolute concentration
of the toxic metal in the diet is not as important aa
the ratio of the concentration of the toxic metal to
the concentration of the essential elements. It is
conceivable that for every toxic metal there are one
or more essential elements that are antagonists. Con-
versely, for every essential element, there might be
one or more toxic metals that are antagonists. It was
recently demonstrated that rats existing on diets
deficient in magnesium developed leukemias and lymph-
omas ' . The possibility, therefore, exists that
some toxic metal that replaces or interferes with
magnesium in biological processes is responsible for
the induction of leukemia and lymphomas.
The concentrations of essential elements mentioned
above vary 10-fold or more in various food products.
Since different population groups consume different
foods, it can be expected that the toxicities of the
metals will also vary similarly.
This Information is also of great value in de-
signing animal experiments for the purpose of estab-
lishing standards of exposures for humans. Unfor-
tunately, most toxicological experiments in rats are
conducted with commercial rat diets that are heavily
supplemented with essential elements. Compared to
the average 14-day human diet compiled by the U.S.
Dept. of Agriculture , the commercial rat diets con-
tain 10 times as much Cu, 42 times as much Fe, 42
times as much Mn, 12 times as much Mg and 18 times as
much Ca1^.
Recent studies in our laboratory using a rat diet
formulated to contain the same concentrations of essen-
tial vitamins and minerals as the aforementioned human
diet have shown it to be satisfactory for growth and
normal levels of serum and tissue magnesium, calcium,
copper and iron. When rats on this diet were injected
intratracheally with Be(0H)2, which has been show to
produce cancer in rats on a commercial diet in this
laboratory, no tumors developed. This lack of pul-
monary carcinogenesis in rats is probably a more
correct reflection of the incidence of lung cancer in
the workers in the beryllium extraction industry^.
Particle Properties
Since the inhalation of particles is a major means
by which individuals are exposed to metals, a few
comments will be addressed to how particle properties
effect toxicity.
For a relatively insoluble particle to have an
effect on the lung, it must reach the alveolar region
of the lung. To do this a non-fibrous particle must
be < 10 y and a fiber < 2 v in diameter.
A non-fibrous particle will be transported in
several ways. In the case of quartz, the particles
are translocated to lymphatics where the quartz
stimulates an excess production of collagen and the
resulting disease is called silicosis. In the case of
fibers, such aa chrysotile,~the fibers less than 10 u
in length are also transported to lymphatics, but
apparently in moderate doses do not stimulate fibrosis
at those sites, The fibers that,are greater than 10 v
cannot be translocated by macrophages and remain in
the alveoli where they stimulate the production of
collagen on the alveolar walls to cause asbestosis, a
form of interstitial fibrosis.
In these two diseases the excess production of
collagen is the common factor. The difference in
particle shape and size determine the anatomical re-
tention site for the particle and the location of the
lesions. This results clinically in two different
diseases, namely, silicosis and asbestosis.
Another more recent discovery, which will probably
have a greater impact on sampling procedures and toxi-
cological experiments, is that the "volatile" metals,
e.g., As, Cd, Pb, and Tl, are concentrated in fly-ash
particles of respirable size. Natusch^ has shown that
1-2 u particles in fly ash from coal-fired power plants
2
6-1
-------
contain between 3 to 18 times more of the toxic metals
Pb, Tl, As, Se, Cd and Sb than do the greater than 40 u
particles. It has also been demonstrated^ that these
metals are concentrated on the surfaces of these par-
ticles. Evidently the more volatile elements vaporize
and recondense on the surfaces of the iron and silicon
oxide particles. Therefore, the mass concentration of
these volatile elements in these particles is directly
related to the surface area of the particles.
Since the chronic toxic effect of relatively in-
soluble particles might be solely due to the surface
chemicals, it is conceivable that relatively inert
particles coated with a monomolecular layer of volatile
metals will have the same effect as the solid particle
of the same volatile metal. If this is true, then a
JLOO-fold difference in dose-response for the same vola-
tile metal in the two different preparations could exist.
Conclusion
Many variables effect the toxicities of metals.
From the standpoints of monitoring and modifying our
environment and preventive medicine, the principal
variable that is most amenable to study and applica-
tion is the interactions of chemicals. Variations in
concentrations of interacting essential elements in
diets do effect the toxicities of metals. Since at
least one element (selenium) reduces the toxicity of
some metals, while at the same time promotes the reten-
tion of those metals in tissue, tissue concentrations
of the toxic metal cannot always be expected to reflect
the health status of an individual.
Particle properties, e.g., size, shape and surface
chemistry, influence the toxic responses. Volatile
toxic metals have been recently found to concentrate
on the surfaces of particles emanating from coal-fired
power plants, resulting in much higher concentrations
of these toxic metals in the respirable particles. The
toxic effects of these types of particles remain to be
determined.
References
1. Mahaffey, K.R. and Goyer, R.A. J. Lab. Clin. Med.
82:92, 1973.
2. Six, K.M. and Goyer, R.A. J. lab. Clin. Med. _7£:
128, 1972.
3. Schroeder, H.A. Amer. J. Physiol. 207:62, 1964.
4. Schroeder, H.A. and Buckman, J. Arch. Environ.
Health 14:693, 1967.
5. Gunn, S.A. , et. al. Am. .T. Path. _42^:685, 1963.
6. Parizek, J. J. Reprod, Fert. ]_:263, 1964.
7. Ferm, V.H. and Carpenter, S.J. Lab. Invest. 18:429,
1968.
8. Schroeder, H.A. J. Chron. Dis. _18:647, 1965.
9. Campbell, J.K. and Mills, C.F. Proc. Nutr. Soc.
_33:15A, 1973.
10. Mason, K.E., _et. _al. Anat. Rec. 148:309, 1964,
11. Parizek, J., £t. _al. J. Rep. Fert. JL6_:507, 1968.
12. Perry, H.M. and Erlanger, M.W. J. Lab, Clin. Med.
83:541, 1974.
13. Gunsberg, S.B., et. al. J. Pharm. Exptl. Therap.
71:239, 1941.
14. Parizek, J. and Ostadalova, I. Experientia 23:
142, 1967.
15. Eybl. V., et. al. Arch. Toxikol. ^5:296, 1969.
16. Groth, D.H., e^t. al. Trace Substances in Environ-
mental Health VI:187, 1973.
17. Groth, D.H., et. al. Interactions of mercury,
cadmium, selenium, tellurium, arsenic and beryllium.
Chap. 4 in the book Effects and Dose-Response Rela-
tionships of Toxic Metals, G.F. Nordberg, Ed. A.S.P.
Biological and Medical Press. (In press)
18. Ganther, H.E., _et. al. Science 175:1122, 1972.
19. Sunderman, F.W., ^et. _al. Contract No. HSM 99-72-
24.
20. Battifora, H.A., et. al. Arch. Path. 86:610, 1968.
21. Bois, P., et. al. Cancer Res, 29:763, 1969.
22. Zook, E.G. and Lehmann, J. J. Assoc. Offic. Agr.
Chem. 48:850, 1965.
23. Bayliss, D.L. and Lainhart, W.S. NIOSH Internal
Report.
24. Natusch, D.F.S. and Wallace, J.R. Science 183:
202, Jan. 18, 1974.
25. Baker, J.E., et_. al. Proc. 10th Annual Conf.
Microbeam Analysis Soc., Aug. 11-15, 1975.
3
6-1
-------
THE DEPOSITION OF Pb-CONTAINING PARTICLES FROM THE LOS ANGELES ATMOSPHERE
C. I. Davidson, S. V. Bering, and S. K. Friedlander
W, M. Keck Laboratories
California Institute of Technology-
Pasadena, California 91125
Introduction
and in greater detail by Sehmel.
Particle deposition from the atmosphere in the
absense of precipitation (dry deposition) is controlled
by complex transport processes. Most previous work in
this area has been confined to laboratory experiments,
although some field work has been reported. Lundgren
and Paulus have meas'ured total mass distributions and
dustfall in a suburban area. Chamberlain and Gregory
et al. have released spores over natural^grasslands and
observed their deposition. Sehmel et al. have examined
the deposition of a polydispersed rhodamine-B aerosol
under field conditions. The deposition of Pb-containing
particles from^the atmosphere hgs been examined by
Peirson et al. and by Servant.
The .particle deposition rate determines the
residence time of particles in an urban atmosphere and
provides i&fofmation on the effects of a source, on
surrounding regions. Depositing particles may damage
plants or harm animals feeding on the plants if toxic
elements are present in the aerosol. A knowledge of
deposition also has application to tracer studies,
where one type of particle or gas may be used as a
tracer for another if their deposition rates are
similar.
For small particles not influenced by gravity and
assuming 13 « e_, equation (2) reduces to (except very
close to the surface):
dz
dC
-UZ—
* dz
-ku
dC
* d(log z)
(3)
where u^ is the friction velocity in cm/sec and k is ^
von Karman's constant equal to 0.4 (after Chamberlain ).
Log £ refers to the natural logarithm of height.
For large particles, equation (2) becomes:
-v C
s
(4)
Based on the above equations, the concentration
profile C(z) can be used as an indicator of deposition.
For example, Figure 1 shows the condensation nuclei
concentration as a function of height above a sparse
field of 65 cm high Avena Satlva (wild oat grass).
These points are the average of 5 experiments, with
most variations less than 5% between runs. Particles
larger than about 25 Angstroms are detected by the
condensation nuclei counter (9), and sedimentation can
be neglected for most of thesi particles.
The deposition rate depends on the properties of
the aerosol, the air flow patterns, and the nature of
the deposition surface. The major mechanisms of
particle transport Include sedimentation by gravity,
inertial impaction, turbulent deposition, and diffusion.
There are also several minor mechanisms, such as
electrostatics and affects of temperature and humidity.
These mechanisms are reviewed by Hidy.
This paper treats deposition of Pb-containing
particles from the atmosphere, by examining airborne
size distributions and fluxes to flat, horizontal
surfaces. From the airborne size distribution, the
deposition rates can be predicted and compared with the
flux determined experimentally.
Throughout this paper, particle diameters (dp) are
equivalent aerodynamic diameters for spheres of unit
density.
Variation of Concentration with Height
The basic relation between airborne concentration
C in ng/cm^ and flux of depositing particles J in
ng/cm^gec is given by:
vg(z)
-J/CU)
(1)
where v is known as the deposition velocity, In cm/sec.
The govfrnlng equation for the flux is:
J » - (D + e)
dC
dz
V C
s
(2)
where D and e_ are the Brownian and eddy diffusivities,
respectively, in cm /sec, v^ is the gravitational
settling velocity in cm/sec, and x. is the height in cm
above the surface.
To determine v^ from equations (1) and (2), it is
necessary to combine and Integrate these equations.
This integration has been performed by Chamberlain2 and
Far above the surface, dC/dz is small in the
region where the eddy diffusion coefficient is large.
As the surface is neared, £ decreases causing a larger
dC/dz. It is assumed that the flux J_ in equation (3)
is constant with height. Since dC/dz is small except
T
T
400 -
300 -
2
o
o"
2
=>
O
DC
O
UJ
>
§ 200
X
o
C
X 100
HEIGHT OF VEGETATION
X
.20 .40 .60
C(Z)/C (4 METERS)
u*'43 CM/SEC
1,00
Figure 1. Airborne concentration of condensation nuclei
as.a function of height above the ground In a residen-
tial area of Loa Angeles, June 1974, C(4 maters) »
AS,000 fortlSIe8 , Two CNC's were used to make these
cmJ
measurements (General Electric CNC, and Environment One
Model Rich J00). A hot wire anemometer was used to .
determine u .
6-3
-------
near the surface, eddy diffusion must transport the
particles rapidly down to the surface, and the rate-
limiting step in the overall deposition process must
be at the surface itself.
If the points in Figure 1 are plotted as concen-
tration versus log .z, a line with a slope of 2300
particles/cm^ results. The friction velocity calcu-
lated from the windspeed profile was 43 cm/sec during
these runs. From equations (1) and (3), the calculated
deposition velocity was 0.9 cm/sec for condensation
nuclei at a height of 4 meters.
Assuming that the mass of Pb deposited on each
plate originates in the same portion of the size
spectrum, the values of Pb deposition in Figure 2
should be proportional to the airborne concentration of
Pb in that portion of the spectrum. Deposition, hence
concentration increases near the roof surface. Dust on
the roof was analyzed and found to contain a high con-
centration of Pb. Thus resuspension of roof dust is
probably responsible for this increase in Pb deposition
below 150 cm. The deposition is relatively constant
above 150 cm, showing that the influence of roof dust
Is negligible above this height.
For larger particles (and a different surface), a
concentration profile different from Figure 1 may be
expected. Figure 2 illustrates this point. In July
1975 measurements of total Pb mass depositing on flat
teflon plates were made. The plates were located at
heights of 1 cm, 11 cm, 58 cm, 148 cm, and 350 cm
above the roof of the W. M. Keck Laboratories at
Caltech. Each plate was arranged to measure the flux
of Pb to the top and bottom surfaces. Two similar
experiments were run, each with sticky and tionaticky
teflon surfaces. The sticky surfaces were composed of
Tem-R-Tape (Connecticut Hard Rubber Company) with Pb
content below detectability. The results of the two
experiments agreed closely. Values shown on Figure 2
are the averages for both sticky and nonsticky surfaces
(averages of 4 values). The nonsticky surfaces
generally had slightly greater deposition values, but
the difference (<20%) is not significant. The
analytical procedure was the same used previously by
the authors, employing flame atomic absorption spectro-
scopy.
350
300
250
s
o
U."
o
£ 200
UJ
>
o
oo
<
150
X
o
UJ
X
100
50
PLATES FACING
UPWARD
PLATES
FACING
DOWNWARD
_L
»0 20 30 40
Pb DEPOSITION, ng/CM2 DAY
50
Figure 2. Total deposition of Pb on teflon plates at
various heights above the roof of Keck Laboratories,
Caltech.
Airborne Size Distributions of Pb
The mass median diameter (MMD) of airborne Pb-
containing particles is about 0.5um (Lundgren11, Lee
et al.i2),it is not possible to predict the deposition
from only the MMD, however, since portions of the size
distribution that have little mass but high deposition
velocities may dominate the deposition mass. Thus the
entire size distribution must be accurately determined.
Peirson et_ al. have measured the overall deposi-
tion velocity of Pb on a flat surface by dividing the
total flux by total airborne concentration, resulting
in v < .18 cm/sec. Their measurements were made near
Lake®Windermere in England. A similar measurement by
Servant ® in residential Toulouse, France, yielded
.13 cm/sec. In the present work, a value of .29 cm/sec
was obtained. This higher value may be due to differ-
ences in size distributions, but it is important to
note that all of these values are of the same order.
The measured atmospheric distribution of Pb with
respect to size is the result of deposition processes
and dilution acting on the initial distribution. The
source curve In Figure 3 shows a Pb size distribution,
normalized to the total mass, which may be expected at
the tailpipe exit of an automobile run on leaded fuel.
The shape of the curve is based on the data of
Huntzicker et al.10, Lundgren11, Hirschler et al. and
Habibi ¦*•'. TotTT Huntzicker and Lundgren reported sub-
micron peaks in the Pb size distributions at sites far
from the source, showing that considerable submicron
mass must be present at the tailpipe. Hirschler found
Pb particles ranging from .Olym to several millimeters
at the tailpipe. (For the present study, a lOOOvm or
1 mm upper limit is assumed.) Habibi also made measure-
ments at the tailpipe and found 57% of the Pb mass in
particles greater than 9ym with some particles as large
as 300-3000ym.
The receptor curve is the distribution far from the
source, normalized to eliminate the effects of dilution.
As coagulation and other growth mechanisms move the
small particle peak to the right, sedimentation rapidly
depletes the larger particles. Some loss of small
particles due to diffusional deposition and large
particles due to turbulent deposition and impaction also
occurs. Since the large particle loss is greater than
the loss of small particles, the normalized receptor
curve has a large submicron peak.
Figure 4 shows experimental data on two Pb size
distributions: Habibi's source distribution at the
tailpipe, and the present receptor study at Caltech in
July 1975, more than 300 meters from the nearest well-
traveled street. The source distribution was determined
using a 1 cfm Andersen impactor and a .14 cfm Monsanto
impactor. The sampling procedure is described in detail
by Habibi. The receptor distribution was measured
using a modified 1 cfm Andersen impactor.
For these experiments, it was important to sample
large Pb particles accurately, since these particles
have high deposition velocities. The Andersen impactor
6-3
-------
^/A log dp
r>
cm iRrr / /oiFFUSlOMAL
SOURCE/ / DEPOSITION
fi TURBULENT x
DEPOSITION
coagulation and
GROWTH
SEDI . N
MENTATIONN
RECEPTOR
1000
Figure 3, Hypothetical normalized mass distributions
of Pb-containing particulate at the source and
receptor. Maximum particle diameters are assumed to be
l,000utn and lOOvim* respectively.
—/Alogdp
0.2
SOURCE
o.l
100
1000
Op.f"
1975. Only the last two experiments were isokinetic.
The importance of isokinetic sampling is evident when
the percentage of total Pb mass greater than 10pm is
compared for these runs: 2%, 4%, 14%, and 14% respec-
tively.
Calculation of Pb Deposition
An electron microprobe study was conducted to
measure Pb particles deposited on a flat carbon surface
over a two-week period in January 1975, and on copper
over a four-week period in July 1975. Analysis showed
Pb-containing particles as large as 15pm true diameter
(37pm aerodynamic diameter for a density of 6 g/cm3).
This supports the idea that gravitational settling is
important in Pb deposition. In addition, deposition
velocity curves for a smooth surface show v = v for
a windspeed of 2.2 m/sec and d > 0.2pm. ~® ^
V
A calculation of Pb deposition for each stage of
the impactor based on Vg " Vg is shown in Table 1. The
total calculated deposition of 40 ng/cm^day can be
compared to an observed teflon plate deposition of
33 ± 1 ng/cm day (average of sticky and nonsticky sur-
faces) at the same height as the impactor. This value
is higher than the Pb deposition data of Figure 2,
which show 24-hour day deposition, since the teflon
plates exposed during the impactor runs from 8AM to 8PM
received only the large daytime doses.
The calculation of Table 1 is also supported by
comparing the upward and downward plate deposition of
Figure 2. A much larger value of deposition on the
upward plate suggests that sedimentation is the
dominant mechanism.
Figure 4. Experimentally-determined size distributions
of Pb particulate at the source (tailpipe of an auto-
mobile, Habibi ).and receptor (roof of Keck Labora-
tories, Caltech, far from a major roadway).
was therefore positioned on its side facing into the
wind to sample isokinetically. A new impactor orifice
was constructed so that at a flow rate of 13.2 liters
per minute, the inlet air speed approximately matched
the average windspeed expected during the experiment.
Two top stages were constructed with calculated cutoffs
of 26.1pm and 37.0ym. Sticky teflon substrates were
used to minimize bounceoff. The impactor was located
155 cm above the Keck Laboratories roof, which is about
15 meters above ground. Resuspended roof dust should
not be a problem, according to Figure 2.
The experiment consisted of five 12-hour runs,
each from 8AM to 8PM. The windspeed was monitored
continuously and wind direction checked every hour.
The wind direction was found to be consistently from
the south, and very few changes in impactor direction^
were needed once each run began. According to Watson ,
changes in direction of ± 45° will cause a maximum
error of 21% for 37pm particles, and 13% for 12p»
particles for windspeed equal to inlet air speed. The
true sampling error may be greater than this, however,
since the average windspeed of 1.4 m/sec was 80% of the
inlet air speed. Both windspeed and wind direction
measurements show that the mass on the upper stages
will be a lower limit to the true mass.
Four impactor runs were made at the receptor site,
in November 1972, February 1974, May 1975, and July
Table 1
Calculation of Deposition of Pb-Contalning Particles
Impactor
Stage
d ,Mm
min.
37.0
26.1
10.2
vg,cm/sec ^
ave per stage C,ng/m
min 4.1
2.9
0.8
67
32
89
J,ng/cm day
-24
8
2,3,..AF .01 many values
1126 _2
Total 40
Teflon plate 33
The Andersen impactor was calibrated at 13.2
LPM using polystyrene latex and polyvlnyltoluene
monodispersed particles from Dow Chemical Company,
Indianapolis, Indiana. A millipore total filter run
in parallel with the impactor (although not iso-
kinetic) yielded 1117 ng/m3, in agreement with the
irapactor data. Values of £ are ± 10%.
The values of Vg were calculated using sedimen-
tation velocity formulas of Fuchs (16). for the
geometric mean particle diameter on each impactor
stage,except stage A. The flux J was calculated from
equation (4) for each individual stage, and summed
for stages 2 through AF (after-filter). Although
Sehmel's deposition velocity curves show Vg>v^ for
the small particles on the after-filter, the flux of
these particles is negligible because Vg is small.
Conclusions
Calculations based on the size distribution
measurements of Pb-containing particles, indicate that
deposition of Pb is controlled by sedimentation. Even
though 86% of the receptor mass is in particles smaller
than 10pm, these particles do not play a significant
role in the deposition process.
5.-3
-------
Isokinetic sampling was essential for these
measurements, since the deposition calculation depends
critically on the airborne concentration of large
particles.
These calculations apply only to a smooth surface.
For rough surfaces where impaction on roughness
elements becomes important, a higher rate of deposi-
tion would be expected. This is because the same
large particles influenced by gravity are also effi-
ciently deposited by impaction.
Since all portions of the size spectrum do not
deposit equally, smaller measured Pb fluxes in remote
regions (Servant^ , Huntzicker et_ al. , and Davidson
et al.^Jdo not imply proportionally smaller airborne
concentrations. Once the large particles with short
atmospheric residence times are deposited near urban
areas, the smaller, more stable material may remain
airborne for considerable periods of time.
Acknowledgment
The authors gratefully acknowledge S. Garcia for
her technical assistance. This work was supported by
EPA Contract RFP #CI 74-0134 and by NIEHS Grant T01-
ES00004-14. The contents do not necessarily reflect
the'views of the Environmental Protection Agency.
References
1. Lundgren, D. A., and Paulus, H. J., "The Mass
Distribution of La'rge Atmospheric Particles —
And How it Relates to What a High-Volume Sampler
Collects," Paper 73-163, APCA Annual Meeting,
Chicago, June 24-28, 1973.
2. Chamberlain, A. C., "Transport of Lycopodlum
Spores and Other Small Particles to Rough Surfaces',1
Proc. Royal Soc. (London), 296, 45(1966).
3. Gregory, P. H., Longhurst, and Sreeramulu,experi-
mental data from The Microbiology of the Atmo-
sphere , P. H. Gregory, Leonard Hill Books, Ltd.,
London 1961, p. 78.
10. Huntzicker, J. J., Friedlander, S. K., and
Davidson, C. I., "Material Balance for Automobile-
Emitted Lead in the Los Angeles Basin," Env. Sci.
& Tech., 9, 448 (1975) .
11. Lundgren, D.A., "Atmospheric Aerosol Composition
and Concentration as a Function of Particle Size
and Time," J. APCA, 20_, 603 (1970).
12. Lee, R. E., Jr., Goranson, S.S., Enrione, R. E.,
and Morgan, G. B., "National Air Surveillance
Cascade Impactor Network II. Size Distribution
Measurements of Trace Metal Components," Env. Sci.
& Tech. , 6, 1025 (1972) .
13. Hirschler, D.A., Gilbert, L. F., Lamb, F. W., and
Niebylski, L. M., "Particulate Lead Components in
Automobile Exhaust Gas," Ind. Eng. Chem., 49, 1131
(1957).
14. Habibi, K., "Characterization of Particulate
Matter in Vehicle Exhaust," Env. Sci. & Tech. ,
223 (1973).
15. Watson, H. H., "Errors due to Anisokinetic
Sampling of Aerosols," Ind. Hygiene Quart..
March 1954, p. 21.
16. Fuchs, N. A., The Mechanics of Aerosols. Pergamon
Press, MacMillan Co., New York, 1964, Chapter 2.
17. Davidson, C. I., Huntzicker, J. J., and
Friedlander, S. K., "The Flow of Trace Elements
through the Los Angeles Area: Effect on Non-
Urban Areas," presented at the 167th National
Meeting of the ACS, Division of Env. Chem.,
Los Angeles, April 1-5, 1974.
4. Sehmel, G. A., Sutter, S. L., and Dana, M. T.,
"Dry Deposition Processes," Batelle N.W. Labora-
tory Paper 1751 Pt 1, 1973, p. 43.
5. Peirson, D. H., Cawse, P. A., Salmon, L., and
Cambray, R. S., "Trace Elements in the Atmo-
spheric Environment," Nature, 241, 252 (1973).
6. Servant, J., "Deposition of Atmospheric Lead
Particles to Natural Surfaces in Field Experi-
ments," presented at the Conference of Atmosphere-
Surface Exchange Processes, Richland, Washington,
September 1974.
7. Hidy, G. M,, "Removal Processes of Gaseous and
Particulate Pollutants," Chapter 3 of Chemistry
of the Lower Atmosphere, S. I. Rasool, editor,
Plenum Press, New York 1973.
8. Sehmel, G. A., "Particle Eddy Dlffusivities and
Deposition Velocities for Isothermal Flow and
Smooth Surfaces," J. Aerosol Sci.. 4_, 125 (1973).
9. Instruction Manual for the Environment One Corp.
Model Rich 100 CNC.
4
6-3
-------
Cd-HPTAKE BY CEPEALS AMD VEGETAPLES
A. Dewit1, R. De Jaeoere2, D.L. Massart3
1 Belgian military force, Poval School of Health Service,
Cy of Brussels, Kroonlaan 145, B-1050 Brussel (Belqium)
2 Vrije Universiteit Brussel, Department of Plant Pbvsiolooy,
Paardenstraat 67, B-1640 St. Genesius Rode (Belgium)
3 Vrije Universiteit Brussel, Pharmaceutical Institute
Department of Analytical Chemistry and Bromatology,
Paardenstraat 67, B-1640 Sint Genesius Rode (Belgium)
Summary
t< critical study of the analytical and the plant
culture,.methods was carried out for a study of
Cd-uptake by plants. Eventually it was chosen to
study the physiological asoects on water-cultured
whole plants. The amount of cadmium taken up in
function of the time, the temperature, the concen-
tration of Cadmiumsalt, the nature of the anion
and the pH of the solution was determined. Other
investigated aspects are the Influence of Calcium,
Zinc and complete nutrient solutions on the cad-
mium uptake. Soil cultures were also started to
compare the cadmium contamination of the soil with
the cadmium contamination of the comestihle part
of those plants.
Introduction
Cadmium is known as a major environmental pollu-
tant and 1s a potential danger for man when in-
gested or inhaled. Several studies have been
described about high cadmium concentrations in air,
water and soil near areas of heavy automobile traf-
fic and near Zinc smelters. The cadmium can be
taken up by plants from soil and water or, 1f pre-
sent as an aerosol, precipitate on the plants.
This can, of course, lead to ingestion by man.
It was our purpose to study the mechanisms of Cd-
uptake by plants and the factors governing them.
In this communication, we present the methods used
to achieve this and our first results.
Analysis
A critical study of the analysis of Cd was made,
using radioactive tracers. Six different destruc-
tion methods have been studied, namely 1) tfry
ashing in a muffle furnace at 450 to 500°C, 2) Wet
ashing in open vessels with a mixture of HCIOk/HKOi
(1:1), 3) Wet ashing 1n open vessels with a mixture
of HNOi/HiSOv/HjOj (3:3:1), 4) Wet ashing in.pres-
sure bottles with a mixture of HClOn/HNOj (1:1),
5) Low temperature ashing and 6) SchBniger destruc-
tion. More than 98 % yield was obtained with each
of these methods except 1. Method 2 was used 1n
routine in most experiments, 1h conjunction with
direct atomic absorption. When very low concentra-
tions (<0,1 ppm) are present, this method is no
longer sufficient and a preconcentration step is
necessary. Two possibilities were investigated :
extraction with a chelating agent or with liquid ion
exchangers. Concerning the former, more than 98 %
extraction was obtained with the following extraction
systems : saturated dithizone in methylisobutylketone
(MIBK) from a pH 4-6 solution, 1 % 2-mercaptobenzothi-
azol in MIBK from a pH 6-7 solution, MIBK from a pH
6-11 solution containing 0,05 % sodiumdiethyldithlo-
carbamate solution, MIBK from a pH 5 solution contai-
ning 1 % ammonlumpyrrol1d1nedithiocarbamate. To avoid
the difficulty of adjusting the pH of the destruction
mixture within rather narrow limits, a method was de-
veloped 1n which Cd is extracted with more than 99 %
yield in the pH range 3-11 using a mixture of dithizo-
ne and sodiumdiethyldithlocarbamate. Even then, the
addition of a base can lead to unacceptable blank
values when very low Cd concentrations have to be
measured. This can be avoided by using a procedure
1n which, after the plant has been subjected to low
temperature ashing, the ash is dissolved 1n 1 M HC1
and extracted directly Into a liquid ion exchanger,
such as Alamine-336. This method has already been
published.1
Culture methods
Water Cultures
Seeds of wheat (Tritlcum aestivum) were germinated
for 24 hr on white sand and then on gauze for 2-3
days, in aerated 0,5 mM CaS0» solution. The plants
were then transferred to plastic sieves containing
25 plants each and grown another 3 days on 0.5 mM
CaSOn. -
They were grown at 25°C, first in the dark (on gauze}
and then in the light (on the plastic sieves).
After this, the roots were washed with distilled
water and the plants used for Cd-uptake experiments.
For each treatment three plastic sieves (75 plants)
were used.
Results
The main conclusions which can be drawn from the
experimental water culture results are the following-.
1. Most of the Cd-uptake Is passive, but part of 1t
is active. This can be shown by following the up-
take at two different temperatures (25°C and 3 °C)
l
-------
2. The active uptake shows a dual pattern : one
mechanism seems to be operative at low Cd-concentra-
tions (<0,5 ppm) and the other at higher Cd-concen-
trations (>0,5 ppm).
3. The amount of Cd taken up at low concentration
(<10 ppm) by wheat-seedlings reaches a maximum va-
lue after a time depending both on Cd-concentratlon
and temperature. The experimental data fit a Lang-
tnuir adsorption-isotherm. Me therefore postulate
that the limiting step in the Cd-uptake is an adsorp-
tion on some cell-components. Calculations show
that all the available adsorption sites are occupied
by 38 ug Cd/g dry weight.
4. When working with high Cd-concentrat1ons (>10 ppm)
this maximum value 1s not reached : Cd-uptake increa-
ses regularly with time. This phenomenon is coupled
to a continuous K^loss. This probably means that at
such Cd-concentrat1ons the physiological functions
are so deeply disturbed that the plants slowly die.
5. For a given Cd-concentrat1on in the solution, the
velocity of uptake becomes constant after + 3 h.
During the very first phase of uptake, the~velocity
varies, partly because the plants undergo a metabo-
lic "shock" when transferred from the CaSO,, to the
Cd-splution.
6. The Cd-uptake depends on the presence of calcium
and zinc, among others. The effect of Zinc is 3
times higher than that of Calcium.
Soil experiments
These studies are conducted 1n a greenhouse. Plants
are grown from seed 1n 10 liter pots filled with a
sandy'loam soil. Treatments are Initiated when the
plants are 3 weeks old and 10 to 15 cm high. They
receive weekly 1 liter of a complete nutrient solu-
tion (containing the six major elements) to which Cd
1s added. The four treatments Investigated are the
following : 0 ; 0,1 ; 1 and 10 mg Cd/1 solution.
Each treatment is replicated 6 times and we start
with 6 plants in each pot. During the first four
weeks, after starting the treatments, five plants are
collected and analysed. The 6th plant is harvested
at the moment it is ready for consumption.
The first results for radish (Raphanus satlvus L.)
and lettuce (Lactuca satlva L.) show a continuous
Cd-uptake during 3 (lettuce) to 4 (radish) weeks.
After this period, the Cd content decreases markedly.
Because the K-content 1s constant during the first
weeks and decreases also at the moment Cd decreases
we think, that when the Cd-concentrat1on in plants
attains to a certain level (which can be different for
each plant), the state of health of these plants 1s
disturbed. The study of other cations in these
plants and the results we are calculating for other
vegetables can maybe lead to a clearer insight Into
these experiments.
References
1 A. Dewit, R. Duwljn, J. Smeyers-Verbeke, P.L. Mas-
sart (1975), Bull. Soc. Chim. Beiges, Vol. 84 (91)
2
6-4
-------
SOME CONSIDERATIONS ON MONITORING OF TRACE METALS IN ESTUARIES AND OCEANS
DOUGLAS A. SEGAR AND ADRIANA 1. CANTILLO
National Oceanic and Atmospheric Administration
Atlantic Oceanographic and Meteorological Laboratories
15 Rickenbacker Causeway
Miami, Florida 33149
Summary
The trace metal chemistry of the coastal waters of
New York and New Jersey in the vicinity of the Hudson-
Raritan river discharge has been studied as part of the
National Oceanic and Atmospheric Administration (NOAA)
Marine Ecosystems Analysis (MESA) New York Bight pro-
ject. Sampling and analysis procedures which minimize
contamination and analysis time were developed for the
determination of several dissolved trace metals. The
geographical and short term temporal variations of
trace metal content appear to be large. However,
intensive sampling in a restricted geographical area
does reveal the existence of coherent cells of water
which contain anomalously high metal concentrations.
The geographical location of these cells suggests they
are caused by the river discharge influence and by the
sewage sludge or dredge spoil dumping. An appreciable
fraction of the metal present in the dissolved phase in
New York Bight is not determined by historically
preferred analytical techniques. The extreme vari-
ability of metal concentrations necessitates extensive
sampling programs if the processes affecting metal
introductions, transport, and removal are to be ad-
equately described. Shipboard instrumentation is under
development which will fulfill this requirement by
continuous real time horizontal profiling of trace
metal concentration from a moving vessel. This equip-
ment may ultimately be modified to perform continuous
unattended monitoring from a buoy or platform.
Introduction
Many metallic elements are known to be released to
the estuarine and coastal environment as contaminants
with the waste products of civilization. These same
metals are also present in the natural airborne and
waterborne products of continental weathering and
volcanism that reach the estuaries and oceans. Many of
these elements such as Hg and Pb are extremely toxic to
living organisms. Other elements such as Cd, Cu, Mn,
and Zn are essential to the functioning of living
organisms but may be inhibitory or toxic if present in
high concentrations or in certain chemical forms.
Considerable effort has been directed in recent
years to the problem of assessing the impact of con-
taminant trace elements on estuarine and coastal
ecosystems. However, little has been learned concern-
ing the distribution, speciation and control mechanisms
of dissolved trace elements in coastal waters. Most of
the recent pagers have been concerned with relatively
few analyses. Inevitably this has restricted
observations deduced from the data to generalisations
concerning the metal distribution over large geograph-
ical areas2-8, observations of the long term (days and
months) temporal variation at a small number of loca-
tions®"^", or attempts to define the loss or gain from
solution of elements during transport by rivers to the
ocean.1'1"13
The National Oceanic and Atmospheric Administra-
tion (NOAA) Marine Ecosystems Analysis (MESA) New York
Bight Project was recently initiated as a coordinated
multi disciplinary research program designed to enable
us to understand more about the impact of man's act-
ivities on the coastal ocean, particularly the coastal
waters adjacent to the New York Metropolitan area.
These waters receive large quantities of contaminants
from river discharge, atmospheric fallout, and ocean
dumping of sewage sludge, dredge spoil and chemical
wastes. One of the important questions concerning
man's impact on this region is the role of trace metals
in the substantial modifications of marine ecology that
have been observed during the past several decades.
Carmody et al.15 have shown that anomalously high trace
metal concentrations are found in the sediments of New
York Bight in areas where the benthic fauna is markedly
impoverished. This correlation suggests the possibility
of a cause and effect relationship. However, it is
precisely this type of cause and effect relationship
which cannot be proven without a detailed understanding
of the chemistry, biochemistry, and geochemistry of
contaminant elements in coastal ecosystems.
As the first stage in a comprehensive study of the
trace metal biogeochemistry of New York Bight our
initial efforts were aimed at the most difficult analy-
tical problem, the determination of the distribution
and speciation of dissolved trace metals. This problem
is important especially as metals in the dissolved
phase have a greater availability to organisms than
those in solid phases and are therefore potentially
more important.
Sampling and Analysis Techniques
Obtaining a contamination free sample of sea water
for trace metal analysis is not a trivial task.1® Most
sampling systems, pumps and bottles, will contaminate
the sample with certain metals. In order to obtain
contamination free samples a new design of sampling
bottle, the "top drop" Siskin bottle, has been employ-
ed. Use of this bottle was shown to be superior to
other sampling methods.16 Sample collection with these
bottles was carried out on a rosette multi-sampler
which was mounted around, and about 0.5 m above, the
sensor head of an Inter Ocean CSTD system. This ar-
rangement served two purposes. As the sampling bottles
are not hung on a hydrowire or triggered by messenger
the potential contamination associated with this means
of operation^S j_B eliminated. In addition, and perhaps
more important, the system permits samples to be taken
not at randomly selected or at standard depths, but at
depths carefully selected to obtain complete coverage
of the major water types, as indicated by parameters
read from the CSTD. The CSTD system is designed so
that any two of the parameters conductivity (salinity),
temperature, depth, pH, oxygen concentration, and
transmissivity can be plotted out against any other of
these parameters on an XYY' recorder, during lowering
of the rosette sampler. The remaining parameters may
also be monitored on a printed paper recorder. Al-
though this capability has not yet been fully utilized,
it should enable us to optimize sampling in order to
adequately describe the water column with the fewest
possible samples, particularly in deeper waters of the
outer continental shelf. Several NOAA ships currently
have litis capability, although most of the shipboard
systems consist of an interface of a rosette multi-
sampler with a standard STD, without the water quality
sensors of the Inter Ocean system. The analysis of sea
water samples for trace metal concentrations by tradit-
ional techniques is tedious, difficult, and error prone.
However, we have developed simple, direct analysis
techniques for total dissolved Fe, Mn, Cd, Cu, Zn, Cr
and Ni in sea water'''. These methods do not require
any prechemistry to be performed on the sea water and
employ flasteless atomic absorption spectrophotometry
6-5
-------
with a Perkin Elmer HGA 2X00 heated graphite atomizer.
Several hundred samples a month may be analyzed for
these elements with a single atomic absorption spectro-
photometer. Even this analysis process will soon be
completely automated.
The collection and analysis procedure for dissolv-
ed trace metal determinations is simple. The sea water
sample is filtered through an 0.4/1 Nuclepore filter and
collected in a precleaned 1 £ linear polyethylene bottle
Filtration is carried out immediately on board ship by
coupling a low pressure inert gas supply to the top of
the Niskin "top drop" and a snap on fitting containing
the filter to the sampler drain cock. The sample is
not transferred to another container for filtering and
does not contact the open ship's atmosphere. This
procedure minimizes contamination. The filtered sea
water is acidified with 1 milliliter of concentrated
silica distilled HNO3 per liter of water and returned
-to the laboratory. Analysis for the metals listed
above requires less than 5 milliliters of sample.
Larger samples are collected to minimize storage
bottle surface area to volume ratio, and to permit
parallel analyses by solvent extraction to be carried
out as required.
Dissolved Trace Metals in New York Bight
As a preliminary to a more detailed study of the
trace metal biogeochemistry of New York Bight a series
of cruises were carried out to establish the general
features of geographical and temporal variability of
dissolved metals in the water column. Seven cruises
were carried out within the apex of New York Bight over
the period April to November 1974 at approximately 26
day intervals to coincide with an ERTS (Earth Resources
Technology Satellite) overpass. On each cruise vert-
ical profiles of salinity, temperature, oxygen, pH,
nitrate, nitrite, phosphate, silicate, total suspended
load, particle size distribution, suspended total
organic carbon and nitrogen, suspended carbohydrates
and suspended proteins were obtained at 25 stations
(fig. 1) from Niskin bottle collected samples, in
addition to the samples for dissolved trace metal
analyses. Continuous vertical profiles with the Inter
Ocean salinity, temperature, depth, oxygen, pH, and
transmissivity probe were obtained synchronously. The
Nuclepore filters from the filtration for trace metal
analysis have been retained and will ultimately be
analyzed to determine the inorganic composition of the
particulates.
A single additional cruise was carried out in
February-March 1975 on which the same analyses were
performed on samples collected from a series of 64
stations from an area of the New York Bight extending
out to the edge of the continental shelf, approximately
100 nautical miles offshore, and 70 miles south and
east-northeast of New York.
Although the data obtained from these cruises have
not yet been fully evaluated, a number of important
observations have been made which point the direction
in which future research and monitoring efforts should
be directed.
The geographical distribution of each of the
metals analyzed is very non-uniform. The sampling
density in the apex (stations approximately 2-5 nautical
miles apart) is marginal for observations of the detail
of this patchy distribution. Individual high or low
concentration values observed at one depth and one
station could possibly be due to sampling or analytical
error. However, it is believed that most if not all
such observations of anomalies are real, as in many
instances the validity of one high value is supported
by several other values at adjacent stations and depths
(figs. 2, 3). It is apparent that even with the
relatively closely spaced sampling grid that we have
used, we have achieved only marginal resolution of the
major features of the trace metal distributions of the
40
30
. B'
»C'
20
D*—
40°,
20
24
E -
74°00'W
01 ^ _ 10 ^ 20
NAUTICAL MILES
Figure 1. station locations in the apex of New York
Bight. Transect designations refer to
vertical sections in Figure 3.
New York Bight. The New York Bight is an extremely
heterogeneous system with river discharge, sewage
sludge, dredge spoil and acid waste dumping all pro-
viding point sources for contaminant element input and
with a complex pattern of tidal and nontidal currents.
We might expect other coastal zones to exhibit less
heterogeneity. However, it is in just such situations
as the New York Bight that we have the greatest need
to understand not only the total impact of man on the
ecosystem but also the relative individual contribu-
tions of contaminant sources. Without such detailed
knowledge, cost and benefit effective environmental
management decisions such as location of dumpsites
cannot be made. The implication is clear that if we
are to understand the trace metal geochemistry in
areas such as the New York Bight Apex we must have the
ability to determine the size, the origin, and the
fate of masses of water with anomalous trace element
concentrations.
Our present data clearly identify the Hudson-
Raritan river outflow as a major contributor of dissolv-
ed metals, particularly Mn, to the New York Bight
apex. However, mixing curves indicate that much of the
Mn and probably the other metals contributed by this
source may be lost from solution soon after introduc-
tion to the Bight. Dumping of either sewage sludge or
dredge spoil appears to be the source of appreciable
quantities of Zn and perhaps other metals. However,
metals released to solution from such sources are
almost certainly not conservative and are lost from
solution to active particle surfaces and to organisms.
A preliminary assessment of the temporal varia-
tions of trace metal concentrations in the Apex region
was made by occupying one station at both the begin-
ning and the end of each cruise and by comparing data
from consecutive cruises. Variability at any station
and depth is appreciable and often erratic. These
variations are presumably related to the small scale
geographical variability, and temporal variations of
the contaminant chemical inputs to the system both by
the river and by dumping. However, even though the
station which was sampled twice on each cruise was
selected to be the most responsive to tidal and river
flow variations, repeated observations several days
apart usually fell close to each other when compared
to the differences in concentration observed between
stations located in different regions in the Apex. In
2 6-5
-------
\)
74°00 W
73°30 W
Figure 2. Distribution of total dissolved zinc in
_ surface waters (lm) of New York Bight
apex, august 21-24, 1974.
addition, when concentrations at several stations
and/or depths from different regions of the apex (e.g.
stations from near the Hudson-Raritan outflow, sta-
tions from near the dumpsites etc.,) are averaged,
clearly defined seasonal trends are observed. This
suggests that although more detailed synoptic sampling
is required to understand the trace metal biogeo-
chemistry in an area such as the New York Bight,
considerable information could be gained by collecting
time averaged samples or continuous data at about the
same number of stations as were occupied in the present
study. Time averaging of the samples over periods of
between one day and several weeks would probably be
necessary to smooth out the variability so that the
major geographical and seasonal distributions could be
observed. Such samples would probably best be collect-
ed from fixed platforms or buoys with automated equip-
ment.
Trace Metal Speciation
Trace metals in sea water may exist in a diver-
sity of chemical forms. These include simple cationic
and anionic species, inorganic complex ions, organic
complexes, metallo-organic compounds, inorganic and
organic colloids, and various macro solid phases.
Many of these chemical forms are such that they will
pass through a 0.4 m filter and so by convention will
be called "dissolved" even though they may not be
truly in solution. The various chemical forms will
have different properties with respect to their avail-
ability for uptake by organisms or sorbtion by other
particles. Unfortunately very little is known about
the nature of metal species in sea water or the nature
of the influence of species on the biogeochemistry of
the elements.
Most analytical techniques historically used for
trace metals dissolved in sea water involve a separa-
tion step which requires that the element be in a
specific ionic form. Solvent extraction, ion and
Ligand exchange, colorimetric procedures, and polar-
jgraphic analyses all require that the element be in
Lonic form to be separated or detected. Only neutron
ictivation analysis determines the total metal in the
sample without prior separation and this technique can
se used for only a very small number of elements in
sea water. The various analytical techniques used
listorically cal1 r separations under different
:onditions of .id temperature, and with complex ¦-
ition reactions of different stability constants. As
cn
tr
UJ
h-
UJ
a.
u
o
V)
sr.
UJ
t—
LU
2
CL
UJ
a
tJH N f
44
44
36
•
•
45 y
yr
/AI| -
53 m
-
E'
23
24
25
Figure 3.
Vertical distributions of total zinc in
New York Bight Apex. August 21-24,
1974.
3
6-5
-------
the various chemical forms of any element in sea water
may be critically affected by such parameters, each
different technique will determine different fractions
of the total element present. This could be one of
the major reasons why intercalibration trials18 and
literature data agree so poorly on the concentrations
of trace elements in sea water. Poor agreement would
result if significant fractions of the element in sea
water were in chemical forms not readily brought into
ionic form during normal analysis procedures.
The analysis technique that we have developed,
using flameless atomic absorption spectrophotometry
directly on sea water, determines a value close to the
total trace metal concentration in the sample. Very
volatile metal-organic compounds may escape before
atomization and not be included in the analysis. In
addition, differences in vaporization and atomization
rate between chemical forms of the elements in the
sample that do not equilibrate with the standard spike
and those that do, might lead to an error. However,
both of these errors are certainly small and probably
negligible.
We have attempted to establish the fraction of
the various trace metals whose total concentrations we
have determined in New York Bight that are determin-
able by traditionally used procedures, by subjecting
some of these same samples to analysis by a much used
solvent extraction technique. The metals were extract-
ed with ammonium pyrollidine dithiocarbamate (APDC)
into methyl iso-butyl ketone (MIBK) using
spiked samples to check the recovery of the procedure
and to act as standards. Analysis was carried out by
flameless atomic absorption spectrophotometry. Al-
though very few samples have yet been analyzed by this
method, it is already clear that a significant pro-
portion of the Fe and Cu that passes an 0.4 ^filter
is notrextractable, even though these samples were
stored at pH-1 before analysis, and recovery of spiked
additions was quantitative. In fact the extractable
percentage of total metal in some instances is very
small (<10%) and it appears that the percentage extract-
able generally decreases with distance from the region
of the Hudson-Raritan estuary discharge {Table 1). In
the river discharge the Fe and Cu can be extracted
almost quantitatively. The small percentage that is
extractable in the offshore samples is apparently not
related primarily to dumped materials, as several
samples from different open ocean areas which have
been analyzed by the two techniques show similar
results. Further research is required to define the
distribution of extractable and nonextractable frac-
tions of metals in the oceans and to investigate the
physico-chemical nature of the nonextractable material.
It is apparent from these observations that a
major fraction or fractions of some elements in sea
water may not be determined by traditional techniques.
The diversity of species of metals in sea water has
been known for some time"^' ^ but appears to have
been largely ignored in recent years. It is likely
that different separation and determination methodolog"
ies even those based upon the same techniques but with
different analytical conditions, will not give the
same results for trace metal concentrations in sea
water. Therefore, it is imperative that monitoring
programs for trace metal concentrations in different
parts of the oceans be carried out using a single
standardized technique in order that data may be
compared. Direct injection flameless atomic absorp-
tion is rapid, simple, and determines the total metal
present. Therefore, this may be the most appropriate
technique for monitoring, at least until we learn more
about the nature of the trace metal species in "solu-
tion" .
Table 1.
Percentage of total dissolved metal (direct infection
flameless atomic absorption) which is determinable
by solvent extraction (APDC/MIBK) atomic absorption
(Nov. 1973) .
Station
Cu
Fe
Raritan Bay
110
94
6
100
99
8
110
68
12
120
68
13
76
38
22
33
65
25
50
55
Continuous Profiling or Monitoring
Carefully designed laboratory experiments under
simulated conditions may be performed to establish the
nature of the major biogeochemical reactions occuring
during processes such as estuarine mixing and waste
dumping. However, the conditions cannot duplicate
exactly those that are encountered in the environment
itself.
In order to study and understand the factors
controlling geographical and temporal variability of
trace metal concentrations in an area such as the New
York Bight where sporadic dumping and river discharge
occur, it is essential that sampling be carried out on
a much more intensive scale than ever previously
attempted for trace metal analysis. Because of the
cost and time involved in sampling and analyzing large
numbers of samples for trace metals, such a sampling
program cannot be carried out at present.
This apparent dilemma would be solved, if it were
possible to monitor the metal concentration of sea
water in real time during a sampling cruise. The
sampling plan could then be modified continuously to
optimize the observation of transient phenomena both by
minimizing the number of samples taken and by ensuring
the maximum possible synopticity of observation.
Currently under development in NOAA is a shipboard
sampling and analysis system which would at least
partially fulfill the necessary requirements. Water
will be continuously sampled from a vessel under-way
through a teflon hose attached to a towed fish. In-
itially this fish will travel only at a single depth
but eventually it may be designed to be depth con-
trollable, in order to obtain vertical profiles under-
way. Water will be pumped on board the ship by a high
volume pump located on the ship itself. Subsampling of
the sea water stream will take place through take off
points in the teflon hose before the water has passed
through the pump. Subsampled water will be collected
in bottles and will also be passed into a four channel
Auto Analyzer for inorganic micro-nutrient analysis,
into a flow cell salinometer, and into a flameless
atomic absorption spectrophotometer through a series of
sampling valves. The flameless atomic absorption
spectrophotometer will not be operated in a true con-
tinuous mode but will perform one analysis every 1 to 3
minutes depending on the element of interest. Stand-
ardization of the instrument will be carried out by
spiking the sea water samples directly in the atomizer
at regular intervals by means of a separate injection
valve system. Initially only one element will be
determined at a time. However, later developments
are possible with either additional spectrophotometers
or multi element analysis atomic spectroscopy systems
expected to become commercially available soon. It
is expected that with this shipboard system much can be
4
6-5
-------
learned about the nature of the chemical interactions
taking place in the coastal zone, particularly those
associated with transient events such as dumping.
References
1. D. DryssenC-Patterson, J. U. I and G. F. Weichart,
In E. P. Goldberg "A Guide to Marine Pollution",
Gordon and Breach, N. Y. 1972, 41-58.
2. H. L. windom and R. G. Smith, Deep Sea Res.,
19 (1972) 727.
3. J. F. Slowey and D. W. Hood, Geochim. et Cos-
mochim. Acta 35 (1972) 72.
4. p. c. Head, J. Mar. Biol. Assoc. U. K., 51
(1971)891.
5. J. Butterworth, P. Lester and G. Nickless,
Mar. poll. Bui; 3 (1972) 72.
6- A. Preston, Nature 242 (1973) 95.
7. M. I. Abdullah, L. G. Royle and A. W. Morris,
Nature, 235 (1972) 158.
8. v. N. Sankaranarayanan and C. V. G. Reddy,
Current Sci., <42 (1973)223.
9. P. Foster and A. W. Morris, Deep Sea Res.,
18 (1971) 231.
10. A. w. "Morris, Nature, 233 (1971) 427.
11. P. C. Head and J. D. Burton, J. Mar. Biol.
Assoc. ft. K., 50 (1970) 439.
12. v. Sundaraj and K. Krishnamurthy, Current Sci.,
41 (1972) 315.
13. H. J. Windom, K. C. Geek and R. Smith, South
East Geol., 12 (1971) 169.
14. J. H. Steele, A. D. Mclntyre, R. Johnson,
I. G. Baxter, G. Topping and H. D. Dooley
Mar. Poll. Bull., 4 (1974) 173.
15. D. A. Segar and G. A. Berberian, In, H. B.
Mark and J. S. Mattson "Water Quality: The
Chemical Point of View." Marcel Dekkar, N. Y.,
In Press.
16. D. A. Segar and G. A. Berberian, In, T. R. P.
Gibb "Analytical Methods in Oceanography."
Advances in Chemistry Series, American Chemical
Society, Wash., D. C., In Press.
18. P. G. Brewer and D. W. Spencer, Technical
Report, Woods Hole Oceanographic Institute,
Reference No. 70-62, (1970).
19. P. G. Brewer, D. W. Spencer and C. L. Smith,
Atomic Absorption Spectroscopy, ASTM STP 443,
American Society for Testing and Materials,
(1969) 70.
20. K. Kremling and H. Petersen, Anal. Chim. Acta,
70 (1974) 35.
21. D. A. Segar, Int. J. Environ. Anal. Chem., 3
(1973) 107.
22. J. F. Slowey and L. M. Jeffrey, Nature, 214
(1967) 377.
23. D. W. Hood, Environ. Sci. and Technol., 1
(1967) 303.
5
6-5
-------
HEAVY METAL CONCENTRATIONS IN MARINE ORGANISMS AND SEDIMENTS
COLLECTED NEAR AN INDUSTRIAL WASTE OUTFALL
by
Raymond R. Emerson, Dorothy F. Soule and Mikihiko Oguri
Allan Hancock Foundation, University of Southern California
Kenneth Y. Chen and James Lu
Environmental Engineering Program, University of Southern California
Los Angeles, California
Abstract
Sediments and organisms from the bottom and water
column were analyzed by atomic absorption for cadmium,
chromium, copper, iron, manganese, nickel, lead and
zinc. Concentration levels in the sediments were re-
duced in correspondence with the distance from the out-
fall, but concentration levels in the tissues of the
organisms did not correspond similarly. Tissue con-
centrations were mostly below sediment concentrations
by an order of magnitude, but varied independently of
phylogenetic affinities. Uptake and regulatory capa-
bilities of marine organisms may be related to feeding
strategies and environmental contamination levels.
Introduction
The ability of marine organisms to concentrate
and regulate, or their ability to reject, the uptake
of contaminants from their environment has been demon-
strated to vary widely between different species irre-
1 ?
pective of their phylogenetic affinities.
Different concentration levels of contaminants by
marine organisms may be caused by a variety of biotic
factors, some of which would include feeding strate-
gies, age, and environmental conditioning.3'4'5 The
concentration of contaminants such as heavy metals by
these organisms would also be dependent upon the
dynamics and the concentration levels of contaminants
in the physical environment.6 Heavy metal concentra-
tion would more readily be evidenced in populations
exposed to low levels of environmental contamination J
while regulation of heavy metals is more likely to
occur in those organisms living in highly contaminated
conditions such as Los Angeles Harbor.
The purpose of this study was to determine the
concentration of heavy metals in the dominant marine
organisms collected near an industrial waste outfall
in Los Angeles Harbor and to determine if the organ-
isms concentrated the contaminants in relation to their
distance from the outfall.
Materials and Methods
Benthic marine organisms and sediment samples
were collected from five stations located at different
distances from the outfall (Figure 1). Samples were
taken with a stainless steel 0.1 cubic meter Campbell
grab; samples were screened, placed in gallon jars
with seawater and returned to the laboratory. Speci-
mens were sorted and placed in a clean container of
fresh seawater for 24 hours to allow the organisms to
purge themselves pf sediments. The organisms utilized
in this study were limited to those species that could
be collected in sufficient biomass to obtain concen-
trations of the contaminants above existing detection
limits. An attempt was made to obtain similar species
from each station, although in most cases the diver-
sity and number of specimens at each station was
highly variable. Similar species were obtained in
abundance at only two or three of the five stations;
thus a comparison of contaminants in the selected organ-
isms with environmental gradients of heavy metals was
limited in most cases. Fish were collected from the
study area at near bottom depths with an otter trawl.
Ten species of vertebrates, six species of fish
and sediment samples were analyzed for cadmium, chro-
mium, copper, iron, nickel, manganese, lead and zinc.
Clean Laboratory Technique: In preparation for
the analysis of sediments and tissues it is essential
that all labware undergo established cleaning procedures
to control background contamination levels.® The lab-
ware was cleaned in one of three 1000ml teflon beakers
which had previously been cleaned with two changes of
concentrated analytical reagent-grade HNO3 for two per-
iods of three days each at a temperature of about 80°C.
Following the acid wash the beakers were thoroughly
rinsed in deionized quartz-distilled water.
All teflon and quartz glass labware was cleaned
in the manner outlined for the 1000ml beakers. All
subsequent cleanings consisted of subjecting the lab-
ware to two changes of concentrated HNO3 anc' a single
change of 1% NBS-double distilled HNO3 (National Bureau
of Standards double distilled quality). After cleaning,
the labware was rinsed in deionized quartz-distilled
water and wrapped wet in Saran Wrap. All clean labware
was handled only with clean non-talced polyethylene
gloves. Polyethylene vials were cleaned with 6N HC1
instead of concentrated HNO3 which degrades polyethyl-
ene. All procedures were performed in a standard labor-
atory fume hood.
Heavy Metals Analysis: Tissue samples were dried
at 103°C for 24 hours. The sample was weighed and
transferred to a 20-ml teflon beaker into which was
pipetted 2 ml of a 2:1:1 solution of deionized quartz-
distilled water, H2SO4 and HNO3 (NBS quality) and
heated at about 150°C. The digestion was monitored for
a clear, light yellow-colored solution which indicates
total digestion. The sample was collected in a tared
20 ml polyethylene vial and diluted up to 10 ml for
each 0.5 gms of tissue. Samples were analyzed by
atomic absorption. Reagent blanks were run with each
group of samples. Sediment samples were prepared and
analyzed for heavy metals as detailed by Chen and Lu.9
Results
Heavy metal content of the sediments was highest
at stations #1 and #3 nearest the outfall, and sta-
tions #2, #4 and #5 were less contaminated (Table 1).
Heavy metal concentrations in the organisms did not
correspond to sediment concentrations and were highly
variable between species. Synergistic relationships
between heavy metals, along with differences in uptake
and regulation by each species cannot be assessed from
this limited study; however, some general trends be-
tween tissue and sediment heavy metal concentrations
are evident.
6-7
-------
Invertebrates
gaster aggregata which was unusually high at 163 ppm.
Iron: The concentration of iron in the inverte-
brates ranged from 204-338 ppm (Table 2) while sediment
concentrations were two orders of magnitude higher, at
approximately 20,000 to 30,000 ppm. The difference in
the levels of iron in the sediment and in the tissues
suggests that a considerable degree of regulation is
occurring in these organisms or the trace metal is
associated with the sediment in a chemical form which
is unacceptable to the organism.
Copper: The concentration gradient of copper in
the sediment ranged from 39-148 ppm. Tissue levels
were quite variable between species at each station,
ranging from 26-334 ppm. Copper was accumulated above
sediment concentrations in some species. Those organ-
isms evidencing accumulation were Nassarius fossatus
(gastropod), Callianassa calxforniensis (ghost shrimp),
and an unidentified pelecypod.
Zinc: Invertebrate tissue concentrations of zinc
ranged from approximately 60-100 ppm, while sediment
concentrations ranged from 98-335 ppm. The gastropod
Nassarius fossatus and the polychaete Sigambra tenta-
culata accumulated zinc at the lower sediment concen-
trations,»wh'ich occurred at station #5. Other species
found in sediments with greater concentrations of zinc
maintained zinc levels to 100-200 ppm below the concen-
tration levels of the more contaminated sediments.
Cadmium: Low levels of cadmium were present in
both tissues and sediments. Tissue concentrations were
lower than sediment concentration in all cases except
in the polychaete sigambra tentaculata collected at
station #3 which was about double the sediment concen-
trations.
Chromium: Concentrations of chromium in the sedi-
ment ranged from 69-158 ppm. Tissue levels ranged from
2-32 ppm with the exception of cirriformia (polychaete)
collected at station #2 which was unusually high at
75 ppm.
Manganese: Tissue concentrations of manganese
were about the same order of magnitude as chromium and
ranged from 2-34 ppm, although sediment concentrations
were higher on the average by a factor of about 4. This
suggests a greater degree of regulation of manganese
by most species, or of some unknown dynamics which miti-
gates against concentration, sigambra collected at
station #5 was exceptionally high at 106 ppm.
Nickel: Tissue concentrations of nickel were sim-
ilar to concentration levels of chromium and manganese,
although tissue levels of nickel were only slightly
less than sediment concentrations. Sediment concen-
trations ranged from 21-40 ppm while tissue concentra-
tions were more varied and ranged from 2^-36 ppm.
lead: Lead was present in all species in only
trace amounts and ranged from approximately-only 1-3
ppm. Sediment concentrations were considerably higher
and varied from 63-270 ppm.
Vertebrates
Six species of fish were collected near the Indus-
trial waste outfall and the muscle tissue was examined
for heavy metal concentrations. Concentrations of
heavy metals In the fishes were an order of magnitude
lower than the concentrations of heavy metals found in
the benthic Invertebrates with the exception of the
nickel (Table 3). Nickel concentrations in the verte-
brates ranged from 2-36 ppm while fish muscle tissue
ranged from 10-20 ppm with the exception of cymato -
Discussion
In general, the ratios of heavy metals in the
sediment and in the invertebrate organisms was roughly
similar. Higher levels of heavy metals in the sedi-
ments may alter the magnitude of these ratios, but the
process of regulation in these organisms would tend to
modify the variability between different species under
more contaminated environmental conditions. Accumula-
tion of heavy metals was shown by a few species and
may indicate a potential for these organisms to toler-
ate greater levels of environmental contamination.
The concentration of heavy metals in the inverte-
brates did not correspond with concentrations of
heavy metals in the sediments in relation to distance
from the outfall. The heavy metal concentrations in
those organisms collected at more than one of the sta-
tions, which included Sigambra tentaculata, Macoma
nasuta, and Cirriformia luxuriosa, did not correspond
in contamination levels to the sediments.
Major phylogenetic groups were also highly vari-
able in that the four species of polychaetes did not
demonstrate any similarity 1n heavy metal content.
This could be explained in part by the differences in
the modes of feeding as cirriformia luxuriosa feeds by
means .of elongated grooved tentacles while the lumbrin-
erids with well developed jaws are active burrowers
and seek out larger particles of organic material.
Notomastus is also an active burrower, without jaw
parts, and may be limited to smaller particles of
organic materials, sigambra, which 1s also an active
burrower, exhibits predatory capability, having a
characteristic everslble proboscis.
The diversity of feeding strategies represented
in other major phylogenetic groups would also account
for the highly variable levels of heavy metals In
these organisms. The crustacean calllanassa caii-
fomiensis constructs a mud burrow within which it
digs up the sediment and winnows out the organic
detritus. The bivalve mollusc Macoma nasuta 1s a
selective deposit feeder which collects detritus from
the surface by means of siphon currents. Particles
are drawn into the body cavity where they are sorted
after which the larger particles are accumulated and
periodically forced out the excurrent siphon.
Additional feeding strategies are represented by
the cerianthld or burrowing anemone which may be con-
sidered a carnivore and selective deposit feeder.
Flagellar movement transports organic material to the
tip of the tentacles whereupon it is transferred by
the tentacles to the mouth, while small prey swimming
into the perimeter of the tentacles may be captured by
the tentacles with the aid of nematophores.
The ascidian ciona intestihalia is a filter feeder
collecting plankton from the surrounding water medium.
While the heavy metal concentrations of dona and the
sediment are not comparable 1t should be noted that
heavy metal levels were within the concentration
ranges of those invertebrates more directly dependent
upon sediment quality. The Internal organs had higher
levels of chromium, copper. Iron, lead and zinc than
the external tunic which had higher levels of cadmium
and nickel. Concentration levels of manganese were
similar 1n the internal organs and tunic.
When considering the low heavy metals concentra-
tions 1n the fish It should be noted that only the
muscle tissue was analyzed and that other tissues such
as the bone and skin may concentrate or regulate
6-7
-------
differently the levels of heavy metals in these organ-
isms.
The feeding strategies of the fishes are not well
known, although they are primarily planktonic feeders
rather than benthic, and not directly affected by con-
tamination levels occurring in the sediment and ben-
thic organisms.
Literature Cited
Graham, D.L. (1972). Trace metal levels in intertidal
mollusks of California. Veliger. 14(4):365-372.
Pringle, B.H., et al. (1968). Trace metal accumulation
by estuarine mollusks. ASCE Sanitary Engineering
Div. J. 94(SA3 Paper 5970) :455-475.
- Bryan, G.W. and L.G. Hummerstone (1973). Adaptation
of polychaete Nereis diversicoior to estuarine
sediments containing high concentrations of zinc
and cadmium. J. Mar. Biol. Assn. U.K. 53:839-857.
Leatherland, T.M. and J.D. Burton (1974). The occur-
rence of some trace metals in coastal organisms
with particular reference to the Solent Region.
J. Mar. Biol. Assn. U.K. 54:457-468.
Phelps, D.K., R.J. Santiago, D. Luciano, and N.Irizarry
(1969). Trace element composition of inshore and
offshore benthic populations. Nat. Symposium
Radioecology. 2nd, Ann Arbor.
Goldberg, E.D. (1957). Biogeochemistry of trace metals,
p. 345-358. hi Joel W . Hedgpeth (ed.) Treatise
on marine ecology and paleoecology. Vol. 1.
Ecology. Geol. Soc. Amer. Mem. 67.
Brooks, R.R. and M.G. Rumsby (1965). The biogeochem-
istry of trace element uptake by some New Zealand
bivalves. Limno. Oceanog. 10:521-527.
Patterson, C. (1974). Interlaboratory lead analyses of
standardized samples of seawater. Mar. Chem.
2:69-84.
Chen, K.Y. and C.S. Lu (1974). Sediment compositions
in Los Angeles-Long Beach Harbors and San Pedro
Basins. Iji Marine Studies of San Pedro Bay,
Calif. Part 7. Allan Hancock Foundation and Sea
Grant Program, Univ. So. Calif. 177 pp.
¦
J
.li ' Jt
SAH PEDRO BAY
6000
3000
PACIFIC OCEAN
Figure 1. Map of Los Angeles and Long Beach Harbors showing the location
of the industrial waste outfall and the sites from which the
samples were collected.
3
6-7
-------
Table 1. Concentrations of heavy metals 1n sediment samples from Los Angeles Harbor
Element concentration # 1n |
ppm (dry wt)
Station
Cd
Cr
Cu
Fe
Mn
Ni
Pb
Zn
1
5.72
159
128.0
32570
550
34.8
261
335
2
3.94
122
85.5
30190
468
21.3
126
168
3
4.95
144
148.3
30720
515
40.1
270
235
4
3.96
123
92.2
26390
415
31.4
139
174
5
3.26
69
39.4
19430
398
26.9
75
98
# Analy'ses are means of duplicate determinations.
Table 2. Concentrations of heavy metals in invertebrates from five benthic stations
Station Species
Cd
Cr
Element concentration# in ppm (dry wt)
Cu Fe Mn Ni
Pb
Zn
1
Ciona Intestinalis (tunic)
3.0
7.0
54.8
208.0
25.6
31.7
0.4
85.0
1
c. intestinalis (internal org.)
1.2
13.3
72.7
312.8
25.5
13.2
1.1
157.1
2
Cirriformia luxuriosa
1.8
75.3
119.6
338.2
34.4
9.8
2.2
67.6
2
Lumbrineris sp.
1.3
10.2
29.6
233.3
16.1
3.5
0.7
92.6
2
Pelecypod (internal org.)
1.0
4.1
315.9
217.1
21.7
2.2
0.5
65.0
3
Macoma nasuta
1.0
19.1
140.7
266.1
34.3
6.5
2.0
75.9
3
Notomastus tenuis
0.6
33.0
46.1
180.8
13.5
7.8
1.2
59.4
3
Pelecypod (internal org.)
2.6
26.3
81.9
289.0
15.1
4.5
0.9
114.9
3
Sigambra tentaculata
8.3
9.3
127.1
228.5
7.7
3.8
1.3
113.7
4
Callianassa califomiensis
0.7
3.4
207.7
222.7
8.8
3.2
0.1
67.8
4
Cerianthus Sp.
0.9
2.4
26.4
204.8
2.8
3.1
0.4
105.8
4
Macoma nasuta
0.9
22.2
79.8
249.4
15.3
19.2
1.3
99.5
4
Sigambra tentaculata
1.3
10.9
124.1
226.5
5.71
18.7
0.4
249.2
5
Callianassa califomiensis
1.9
5.7
216.4
210.7
21.4
2.1
0.7
76.0
5
Cirriformia luxuriosa
1.0
4.8
48.2
235.3
9.0
13.1
0.4
57.1
5
Nassavius fossatus
2.3
12.1
334.7
319.1
7.6
36.2
0.9
213.9
5
Sigambra tentaculata
2.2
21.6
149.9
294.7
106.6
6.6
1.0
133.2
# Analyses are means of duplicate determinations.
4
6-7
-------
Table 3. Concentrations of heavy metals in vertebrates (fishes) collected from the study area
Species
Cd
Element concentration# in ppm
Cr Cu Fe Mn
(dry wt)
Ni
pb
Zn
Anchoa compressa
0.7
0.8
1.
1
30.3
2.
. 1
13.6
trace
47.
,3
Cymatogaster aggregata
0.8
2.0
1.
5
49.3
2.
.3
163.9
0.6
56.
.0
Engraulis raordax
0.9
2.5
25.
1
70.9
3.
.7
14.1
0.3
43.
,8
Genyonemus lineatus
1.2
0.3
10.
9
33.1
1.
.0
14.5
0.2
17.
,8
Phanerodon furcatus
0.4
0.6
1.
3
9.8
0.
9
9.0
trace
17.
9
Seriphus politus
0.9
0.7
1.
8
29.7
0.
.5
18.5
0.1
23.
.7
# Analyses are means of duplicate determinations.
5
6-7
-------
ANALYSIS OF FOOD FOR RESIDUES OF PESTICIDES
Jerry Burke and Bernadette McMahon
U.S. Food and Drug Administration
Washington, D.C.
Summary
Analysis of food commodities for residues of synthetic
organic pesticides may be done for different reasons,
but greater demands are placed on the methodology and
the chemist in analyzing a variety of commodities for
multiple residues without benefit of treatment history.
Analytical methodology developed and used by the U.S.
Food and Drug Administration in compliance and monitor-
ing programs will form the basis for summarizing an
approach to the analysis of food for organic pesticide
residues. Emphasis Is on multiresitoe methods for
organochlorine and organophosphorus insecticides,
their applicability to various sample types, scope of
coverage for chemicals, and requirements for analytical
methodology in^a regulatory program. Important consid-
erations necessary for reliable residue analyses are
covered in relation to the unit operations: extrac-
tion, -cleanup, determination, and confirmation of
residue identity.
Analyses of food commodities for residues of
pesticides are done for different purposes; e.g.: (1)
to follow the residue decay in a plant or animal after
treatment with a known rate or amount of a particular
pesticide, (2) to help elucidate plant or animal
metabolism of a particular pesticide, (3) to monitor
for pesticide residues, and (4) to enforce previously
established pesticide residue limits or tolerances.
No matter what the ultimate use of analytical
results, their reliability is dependent on the quality
of the total analytical process and the degree of
attention to detail with which this process is applied.
The specific requirements demanded of any analytical
approach are dictated by the ultimate purpose of the
work. In any case, we must presume that the universal
requirement is the correct answer, whether that answer
be a simple Yes or No as to a residue's presence or a
much more difficult determination of the "exact" iden-
tity and quantity.
This presentation will focus on multiresidue
analyses for the purpose of monitoring and regulatory
surveillance using as examples the programs and meth-
ods with which we are most familiar, those of the U.S.
Food and Drug Administration (FDA). This agency's
responsibilities in the pesticide area are to monitor
the United States domestic and imported food supply
for pesticide residues and to enforce tolerances for
these residues in foods. Tolerances have been estab-
lished for approximately 250 pesticides on a large
number of food, products. Tolerances for many chemi-
cals may exist for one food item; for example,
residues of about 90 pesticides are permitted on
apples. A tolerance may include metabolites of the
pesticide. In addition, the need exists to determine
residues on products where there is no tolerance for
their presence. To meet these needs, it is necessary
to use multiresidue procedures, capable of simultane-
ous detection and measurement of several residues in
a single sample and applicable to a wide range of
commodity types. Ho other course is practical, given
the thousands of possible residue/commodity combina-
tions and the fact that samples are analyzed without
the benefit of sample treatment history.
Analyses of foods for these purposes are typi-
fied by the need to cope with: (l) a large number of
commodities of varied physical and chemical composition;
(2) potential residues of a large number of pesticides
and alteration products comprising a wide range of
chemical properties', (3) residues resulting from direct
application, translocation from soil, or ingestion by
animals; and (4) the need based on legal and/or toxico-
logical considerations for different limits of quanti-
tation for different residues. It is believed by many
marginally acquainted with residue analysis that there
is one method capable of detecting and measuring all
pesticides in all commodities. With the aforementioned
variations, it will never be possible to devise a single
method satisfactory for all purposes. It has been
possible, however, to extend the applicability of some
methods to encompass a number of different sample types
and pesticides. To develop methodology with comprehen-
sive multiresidue capability, applicable to a variety
of sample types, it is desirable to choose a sound
analytical method and "stick with it" while expanding
its capability through additional research. Validation
of its accuracy and intra- and inter- laboratory test-
ing can then follow. Decisions based on residue data
resulting from food surveillance analyses are of such
public health and economic significance that the suit-
ability of the analytical method and reliability of the
results cannot be taken for granted. We caution that
such a large investment in research and validation in
any one analytical scheme may make it difficult to
justify changing to use of different methodology. Fur-
thermore, when a method has been highly developed, it
is possibly too time consuming and costly to use unless
the objective is truly in line with the scope of the
method. Careful consideration should therefore be
given to future needs and method adaptability before
choosing a method for extensive development and use.
The total analytical process may be divided into
two major parts: (1) that which precedes carrying out
the analytical method: sampling, sample preparation,
and sample storage, and (2) the analysis itself, which
typically consists of the operations of extraction,
cleanup, determination, including confirmation of
residue identity, carried out in a prescribed sequence.
Sampling, Sample Preparation and Storage
Sampling and the associated matters of sample
preparation and storage are extremely important areas
that can be touched upon only briefly in this discus-
sion. It should be evident that the most sophisti-
cated and accurate analysis cannot make up for ill-planned
sampling or errors introduced during sample preparation
and storage. Sampling must be designed and carried
out in a way that will answer the questions to which
the residue analytical program, is directed. Therefore,
a carefully thought-out and precise statement of
objectives of the program is prerequisite. "Guide-
lines on Sampling and Statistical Methodologies for
Ambient Pesticide Monitoring" has been prepared by the
Federal Working Group on Pest Management.^
Practical matters such as sample containers and
shipment of samples to the laboratory must be decided
upon. In general, it is desirable to analyze samples
as soon as possible. Containers should be of a con-
struction that will not contaminate the sample with
substances leached from the walls or caps, that will
not allow migration of the chemicals sought into the
container, and that will hold the sample without
leakage. Glass containers are the most universally
1
7-1
-------
acceptable. Closures should avoid rubber or plastic
seals that could contaminate the sample with materials
detectable in the analysis.
Sample preparation involves both the selection of
the appropriate part of the commodity for analysis and
the mechanics of preparing a homogeneous mix from which
to take the analytical portion. Sample preparation for
residue surveillance or tolerance enforcement programs
may be directed by pesticide regulations of the govern-
ment or policies of the organization conducting the
program. In the United States, pesticide tolerances
established by the U.S. Environmental Protection Agency
usually apply to the raw agricultural product as it is
shipped in interstate commerce.3 Pesticide residues
requiring tolerances on processed foods and feeds (e.g.,
dried citrus pulp) are considered "food additives" and
included in a different regulation series.4 Some regu-
lations specify the portion of the commodity to which
the tolerance applies. For example, the U.S. tolerance
for DDT in corn is 3.5 ppm on "sweet corn (determined
on kernels plus cob after removing'any husk present
when marketed)".3 With foods having no formal toler-
ance for the pesticide(s) in questions, the FDA has
chosen to analyze the "edible portion"; with corn the
kernels only would be analyzed. Directions employed
by the FDA for preparing a large variety of agricul-
tural commodities for analysis are contained in the
FDA Pesticide Analytical Manual, Volume 1.5
Care is necessary to ensure that the gross sample
is homogeneous before taking the analytical sample.
Liquids should be thoroughly mixed and solids chopped
to a small particle size with thorough mixing before
the analytical sampling. The choice of various mechan-
ical chopping and grinding devices used will usually be
dependent on factors such as size and type of samples,
i.e.,'fruits and vegetables or oil seeds and grains,
and experience of other laboratories with similar
commodities. It is often desirable and usually requir-
ed in regulatory programs to retain a reserve portion
of the sample for replicate analysis or for examination
by another laboratory. Subsequent analysis of the
stored portion should be preceded by thorough remixing
and reincorporation of any liquid which may have
separated.
Sample storage of perishable commodities at about
-10°C or below is a generally acceptable practice.
Consideration should be given to the effects of storage
on residue stability.^ If questions exist, control
samples fortified with known quantities of the pesti-
cide(s) of interest should be held at the same storage
conditions and analyzed at appropriate intervals. As
previously mentioned.it is desirable to minimize the
time of sample storage, and to use containers which
will ensure sample integrity without adding contamin-
ants.
Analytical Methodology
The large literature on residue methodology for
the synthetic organic pesticides dates from the mid
1950's. Interest in detection and measurement of resi-
dues of pesticides in foods no doubt has ushered in a
new realm of analytical chemistry with its specialized
needs in reagents, instrumentation, analytical methods,
and chemist training and experience. No attempt will
be made here to review this literature. Instead we
will give a brief overview of the present state of resi-
due analytical methodology regularly used by the FDA
and- emphasize aspects of the methodology which we
believe are of maximum importance. The points,to be
made about these methods are generally applicable to
most residue methods despite differences in purpose and
structure.
2
Multiple residue analytical methods used by the
FDA in regulatory programs and a variety of related
information are included in Volume I of the FDA Pesti-
cide Analytical Manual.^ Volume II of this Manual
contains methods for individual pesticides for which
tolerances have been established. A booklet entitled,
"Guidelines on Analytical Methodology for Pesticide
Residue Monitoring", prepared by the Federal Working
Group on Pest Management, discusses many methodology
considerations that are vital in a residue analysis
program.''' These guidelines can be recommended to both
the beginning and experienced residue analyst.
Of the more important chemical classes of pesti-
cides (organochlorines, organophosphorus, carbamates,
chlorophenoxy acids, triazines), analytical methodology
is furthest advanced for the organochlorines. Multiple
residue methods came into use in the late 1950's with
Mills' outline of the classical approach still in use.®
Thin layer chromatography (TLC) provided the means of
simultaneous detection and semi-quantitation of about
a dozen organochlorine compounds. Although much improve-
ment, expansion, interlaboratory testing, and actual
use of the methodology has occurred in the intervening
years,the basic approach remains the same.9 Gas
chromatography (GLC) with electron capture detection
has for several years been the primary means of detect-
ing and measuring organochlorine residues.
Nonpolar Compounds
Tables 1 and 2 outline this analytical scheme,
which' has been used by the FDA since the early 1960's
to analyze for residues of organochlorine and other
nonpolar pesticides and industrial chemicals in most
of the tens of thousands of food samples collected for
surveillance and tolerance enforcement.
Table 1. Procedure for organochlorine (nonpolar)
pesticides in fatty products
1. 25-100 g sample (to provide 3 g fat)
2. Extract fat from sample: various means
3. Partition £ 3 g fat/petroleum ether with acetoni-
trile (4 times)
4. Transfer residues from acetonitrile + water to
petroleum ether (2 times}
5. Cleanup by adsorption chromatography on Florisil.
Choice of 2 elution systems, 3 eluants each (ethyl
ether/petroleum ether or methylene chloride/ace-
tonitrile/hexane combinations)
6. Concentrate eluates to definite volumes
7. Determine identity and quantity of residues by
GLC with EC and/or element-selective detectors
8. Perform additional cleanup if necessary
9. Confirm identity and quantity of residue
Table 2. Procedure for organochlorine (nonpolar)
pesticides in nonfatty products
1. 25-100 g sample
2. Blend with acetonitrile or acetonitrile + water
3. Filter; measure volume of filtrate
4. Transfer residues from acetonitrile + water to
petroleum ether; measure volume
5-9. Same as Figure 1
This methodology is applicable only to residues of non-
polar compounds, but close examination of its primary
features can serve as an outline of important points
that should be considered in using any multiresidue
method. Extraction, cleanup and determination consti-
tute unique procedures within the total method; yet
each step is inter-dependent on the others for its
effectiveness and/or accuracy. Inter-relationships,
upon which the ultimate reliability of the analytical
7-1
-------
results depend, will be pointed out during this discus-
sion.
Extraction. To deal with different product types
the extraction procedure is varied dependent on the
physical-chemical nature of the sample. Samples are
broadly classified as nonfatty or fatty. Those in the
fatty category range from about 2 to 100# fat or oil,
such as low-fat fish ta vegetable oil. Fat must be
completely extracted from the sample in order to make
an accurate determination of the lipid-soluble organo-
chlorine residues. For example, thoroughly ground
fish tissue is dehydrated with sodium sulfate and the
fat extracted by blending with hexane or petroleum
ether. Fat is extracted from milk with ethyl ether-
petroleum ether after treatment with alcohol and sodium
oxalate thus freeing the fat for solvent extraction.
Nonfatty products are broadly classified as high or low
moisture, such as fresh vegetables or hay. Most resi-
dues of organochlorine, organophosphorus and probably
carbamate pesticides are essentially completely extrac-
ted from fresh fruits and vegetables with acetoni-
trile. Acetonitrile alone has been found
incapable of completely extracting residues from dry
products.; the lower the moisture content, the poorer
the extraction. A mixture of 35# water/acetonitrile is
an effective extractant for samples with less than about
75% moisture content.Because the analytical scheme
for nonfatty products is less time consuming than for
fatty materials it is tempting to use the shorter pro-
cedure with samples of relatively low fat content.
This has been shown to be satisfactory for some
products, e.g., eggs.1^ The sample size is reduced so
that only one or two grams total fat are present and
the sample treated as a nonfatty item. Low fat fish
are also sometimes handled in this manner, A caution
is in order with fish; literature values for fat
content can be misleading since the amount of fat may
vary with season and physical condition of the fish.
Residues of pesticides translocated from the soil to
aerial plant parts or adsorbed from soil into root
cropB are not completely removed by the usual proce-
dures such as blending with acetonitrile. ' ' The
laboratory may need to resort to the Soxhlet extraction
with chloroform-methanol to deal with this type resi-
due.15 Selection of the extraction procedure for any
of a variety of sample types is a critical test of
analyst experience and judgment. Analyses are often
required for samples not clearly fitting into a partic-
ular category. Laboratories are encouraged to select,
to the extent possible, extraction procedures which
have documented validity, and where necessary to conduct
their own short studies to avoid the use of procedures
which would result in generation of questionable resi-
due data. Experiments to determine the recovery through
the analytical method of pertinent chemicals, added to
the analytical sample just prior to the extraction
step, are absolutely necessary to check method perfor-
mance. Such recovery studies do not test the effective-
ness of the extraction in removing actual residues,
however. A study with radiolabeled pesticides "grown
in" to the biological specimen is the only way to make
an absolute evaluation of an extraction procedure.16
A practical and good test of extraction efficiency can
be gotten by analysis of samples containing actual
residues using both an exhaustive extraction and the
analytical extraction procedure. Soxhlet extraction
with chloroform-methanol has been proven effective
with difficult-to-remove residues and is widely used
as a comparative exhaustive procedure.
Cleanup. The "cleanup" procedure for isolating
residues from other substances coextracted from the
sample varies somewhat for fatty and nonfatty foods.
Extracts of both sample types are cleaned up by
column chromatography with an active adsorbent,
Florisil, a synthetic magnesium silicate. Fats and
3
oils are partitioned between petroleum ether and
acetonitrile prior to chromatography to eliminate most
of the lipid material. Extracts of plant material and
fats are placed on the column in the hydrocarbon solu-
tion to which they were transferred from water-acetoni-
trile. Sample sizes must be chosen so that the amount
of extractives placed on the Florisil column does not
exceed the column's capacity to retain them. A maxi-
mum of three grams of fat or oil is treated by the
initial partitioning, because a larger sample size
inhibits the partitioning of pesticides to acetonitrile
and results in more than the 250-300 mg of lipid
extractives that the Florisil column can tolerate. The
Florisil column accomodates the extractives from about
75 g of fresh green vegetable but the higher concen-
tration of solid materials in dehydrated products
forces tne choice of an accordingly smaller sample
size. Overloading the Florisil column with excess
extractives results in inadequate cleanup and/or pre-
mature elution of the residues. Presence of even a
small amount of polar solvent like acetonitrile' in the
extract solution placed on the column will cause the
same effects.
The complete cleanup process requires several
different manipulations of the sample solution,each
of which is a potential source of loss of the chemi-
cals sought.1''' Recoveries through the transfer from
aqueous acetonitrile into petroleum ether and from a
petroleum ether solution of fat into acetonitrile
depend on partitioning characteristics of the chemicals
sought and the physical parameters of the partitioning
manipulation.In both cases the two solvents must
be given sufficient contact by vigorous mixing for 30
seconds to one minute. It is desirable to hold the
separatory funnel in a horizontal position to maximize
surface area at the interface.^ Published partition-
ing values are helpful in predicting pesticide "behavior
in the two solution transfer operations. Partition-
ing 'behavior of a chemical in the two different
partitionings is exactly opposite. Compounds readily
transferred from petroleum ether to acetonitrile,
such as lindane, are prone to low recovery in the
transfer from aqueous acetonitrile to petroleum ether
unless the two phases are very well mixed. Hexachloro-
benzene and mirex are easily extracted from aqueous
acetonitrile into petroleum ether but only about 70
per cent of eitVier is recovered through the petroleum
ether to acetonitrile partitioning.21
Evaporation of solutions is a potential source of
residue loss if done in open vessels and if solutions
are allowed to go to dryness.22 As a general rule
volatility of the compound increases with decreasing
GLC retention time. Losses are particularly severe
when less than about 5 micrograms of tne pesticide is
present and when the solution is free of sample extrac-
tives which tend to retard loss. Kuderna-Danish
evaporators with special reflux condensers and micro
evaporating vessels with similar condenser Columns are
recommended for evaporation of-solvents which are vola-
tile at steambath temperatures.
The adsorbent column is the source of much analyst
concern in sample cleanup procedures. This is where
most of the sample coextxactives are removed and
important separations of pesticides are made, and it
is well known that adsorbents vary in activity from
lot to lot with consequent effect on both cleanup and
separations. Just as premature elution of pesticides
or unsatisfactory cleanup may be caused by the condi-
tion of the sample solution going onto the column,
late elution of pesticides may occur because the parti-
cular lot or activation batch of adsorbent is
excessively retentive. This can have an Insidious
effect because chemicals expected in the last elilate
from the column may be only partially eluted; for
7-1
-------
example dieldrin, when the Florisil column is succes-
sively eluted with 6% and 15% ethyl ether/petroleum
ether. Unless prior recovery experiments have been
done and e'lution fractions checked with volumes in
excess of that normal for the analysis, low results may
be reported with no obvious indication of a problem.
When multiple residues are sought and when the column
chromatographic step is designed to separate the pesti-
cides into groups, much attention must be given to the
detail of the chromatography to insure that expected
performance is consistently maintained. Each laboratory
should establish for itself a procedure for assuring
reliable behavior of the chromatographic adsorbent used.
This will include purchase according to particular
specifications, record keeping of lot numbers, dates
received, dates and type of activation, activation by a
particular procedure, use of proper storage containers
and conditions, checking for contaminants, and regular
testing with chemicals indicative of variations in
adsorbent "behavior. The use of Florisil by our labora-
tories in a comprehensive multiresidue analysis has
been improved greatly by attention'to these details.
A test based on the adsorption of lauric acid from hex-
ane solution is a valuable adjunct in the evaluation
of the relative adsorptivity of each lot.23 Compensa-
tion for lot to lot differences is made by adjusting
the weight of Florisil used in the column.
"Cleanup of fatty and nonfatty foods by these pro-
cedures is satisfactory, with some exceptions, for GLC
and TLC determination of residues at the levels
required in food analysis. Individual samples may
present particular problems. With fats and oils the
second elution from the Florisil column sometimes may
contain excessive lipids. Electron capturing substan-
ces extracted from some plants, such as carrots, are
not completely removed by Florisil. Supplemental
cleanup such as acid or base treatment or additional
column chromatography is done at the analyst's discre-
tion, with consideration of the effect these treatments
will have on the residues indicated or suspected. >2<+
Determination. GLC with electron capture (ECD)
and alkali flame (AFD) or flame photometric detection
(FPD) is used for the determination of organochlorine,
organophosphorous and some sulfur-and nitrogen-contain-
ing compounds that are recovered by this methodology.
Microcoulometric and electrolytic conductivity detec-
tors providing specific detection of halogen or nitrogen
are used primarily to substantiate findings by the ECD.
The ECD, which has made modern pesticide residue
analysis possible because of its sensitive response to
organochlorine, requires a number of precautions in its
use. Because it is not specific, non-pesticidal
substances extracted from the sample and impurities
present in solvents and reagents may elicit responses.
For example, several of the widely used and distributed
phthalate ester plastizexs(non-halogenated compounds)
are readily detected by the ECD-GLC systems used for
pesticide residue analysis.25 It is imperative that
solvents be distilled from all glass apparatus to
eliminate electron-capturing substances.Solvents
especially treated for use wit'h ECD-GLC are available
commercially. Technical and spectroscopic grades in
most cases are not satisfactory. Solid reagents such
as the filter aid Celite require thorough washing and
checking before use. ECD systems may have a limited
linear response range which can be affected by detector
cell design, radioactive source and its condition,
applied voltage, and degree of contamination by column
liquid phase or sample extractives. ' The analyst must
thoroughly acquaint himself with the performance charac-
teristics of his particular system and develop routine
maintenance procedures. Sellable performance must be
regularly ascertained by the chromatography of standards
ana preparation or response vs quantity plots. Problems
should be corrected without delay.
Quantitation of residues found present by GLC re-
quires careful comparison of peak size of the residue
and a known quantity of reference standard chromato-
graphed shortly before or after the sample. Separation
of the peak of interest from other peaks, availability
of reference standards of known purity, optimum and
stable performance of the GLC column and detector
systems, accurate and reproducible injection of sample,
and care and reproducibility in peak measurement are all
essential ingredients of accurate quantitation by GLC.
Qualitative interpretation of GLC curves necessitates
a decision by the analyst as to whether or not additional
separations or tests to confirm residue identity are
needed and if so, what. For example, heptachlor epoxide
and oxychlordane have identical retention times on
methyl silicone GLC liquid phases but are separatee on
dtethylene glycol succinate (DEGS). These compounds can
also be separated prior to GLC on the Florisil column
by using eluants comprised of bexane, methylene chlor-
ide, and acetonitrile rather than mixtures of petroleum
ether and ethyl ether.Polychlorinated biphenyls
(PCB), widely used industrial chemicals which may
contaminate some foods, are recovered along with organo-
chlorine pesticides. Gas chromatograms of these
multicomponent chemicals consist of 10-20 peaks in the
retention time range of the common organochlorine pesti-
cides. Separation of PCB from pesticides is necessary
for reliable identification and measurement of each.
Column chromatography on silicic acid is a common
approach.29 Treatment with acid and base followed by
column chromatography on Florisil will eliminate the
DDT group leaving the stable PCB for further analysis.30
To verify a finding based on tne presumptive infor-
mation provided by GLC, in particular the non-speciric
ECD, the identity of residues which are illegal, signif-
icant for some reason, or unusual should be confirmed
by separate and unrelated procedure(s). Confirmatory
tests should also be employed occasionally, e.g, every
5 or 10 samples, even when findings are considered
routine. Confirmatory procedures are most valuable when
tney are capable of supporting the original quantitation
as well as identification of the residue. The particu-
lar proceaure(s) for confirmation of the initial
finding must be decided upon by the analyst, again
placing a heavy demand on his knowledge and experience.
Confirmatory procedures readily available to the resi-
due laDoratory Include the formation of characteristic
derivatives Dy chemical reaction or ultraviolet photo-
lysis, TLC> element selective GLC detectors, partition-
ing values, and GLC columns giving different retention
times. An appropriate combination of tests based on
different physico-chemical parameters of the chemical
in question should be used In order to maximize the
certainty of identification.31 GLC coupled with mass
spectrometry offers the most sound residue confirmation
and identification and should be utilized when the more
readily available procedures do not.provide sufficient
certainty In identification.
The parameters of the GLC step most often used
with this method are designed to provide maximum multi-
residue capability for the separation and reliable
quantitative measurement of tne majority of the pesti-
cides In the organochlorine class. 32 These conditions
are also satisfactory for many other chemicals including
a number of organophosphorus compounds." Column liquid
pnases regularly employed range from the nonpolar metnyl
silicones (DC-200 or 0V-101) through the fluoramethyl
silicones (OF-1 or 0V-210) to the polar DEGS. The
methyl silicones are excellent liquid phases for a large
variety or chemicals. The largest number of chemicals
are chromatographed on six foot columns at aoout 200°C
with liquid phase loadings of 5-10%.>* Lower and higher
temperatures are needed for tnose compounds which elute
4
7-1
-------
in or very close to the solvent front or have exces-
sively long retention times at 200°C. detention time
values relative to reference compounds have been col-
lected for a large number of chemicals on several
columns at standardized operating conditions.^
Related Considerations. As used in FDA programs,
sensitivities of the ECD and phosphorus selective AFD
and FPD are adjusted to give one-half scale recorder
deflection for one nanogram of heptachlor epoxide and
two nanograms of parathion,respectively.
Based primarily on experience with analysts' abil-
ity to measure chromatographic peaks and inspection of
chromatograms from many food types, a peak of 10% full
scale recorder deflection is defined as the lowest
quantifiable response. This is 0.2 and 0.4 nanogram
for heptachlor epoxide and parathion,respectively.
Sample size ana cleanup are designed to permit the
injection of the equivalent of 20 mg nonfat or 3 mg fat
for routine GLC analysis. With this weight injected
and minimum quantifiable response, tne limit of quanti-
tation is 0,01 .jppm for heptachlor epoxide and 0.02 ppm
for parathion in nonfatty foods. With fats the quanti-
tation limit is slightly below 0.1 ppm for heptachlor
epoxide and slightly above 0.1 ppm for parathion.
Tabulated response values at standardized GLC conditions
quickly allow estimation of the limit of quantitation
for a large number of chemicals.5
This somewhat arbitrary but practical means of
establishing the limit of quantitation has provided a
way whereby all laboratories using the methodology can
generate quantitative data with a uniform and defined
value for the lowest finite value reportable. The
sample weight and the effectiveness of the cleanup must
be carefully considered along with the sensitivity of
the detection system in arriving at the desired limit
of quantitation. Above all, the limit of quantitation
should fit the purpose of the analysis. A group of
laboratories such as those of the FDA, conducting a
food surveillance program, must strive to adhere to the
limit of quantitation for tne analyses in all labora-
tories. A single laboratory carrying out a particular
study should maintain a consistent quantitation limit
throughout the study.
Seeking the smallest detectable pesticide residue
level in food surveillance is not usually necessary or
prudent. Tolerance enforcement and general monitoring,
with but few exceptions, do not require the methodology
to be overextended in reaching extremely low residue
levels. In a similar vein the advantages, disadvan-
tages, and pitfalls of using ultra-sensitive detection
systems with concurrent reduction in size of the analy-
tical sample should be examined carefully before
choosing the small sample-high sensitivity approach.35,36
Reliable interpretation of GLC chromatograms from
the analysis of foods is dependent on many factors.
Obviously the methodology must be sound and the GLC
column and detector systems must be functioning at top
performance. Not so obvious perhaps is the fact that
the analyst must be knowledgeable and experienced in the
art and science of residue analysis by GLC. He must be
aware of the capabilities and limitations of the method-
ology and unwilling to reach unwarranted conclusions,
yet judgments he must make. He should be knowledgeable
in the residue chemistry of pesticides, in the perfor-
mance capabilites of tne GLC instrumentation, in the
performance of the methodology with different sample
types, and in the behavior of many chemicals through
all steps of the method.
To meet the last need we have developed and
compiled information on the behavior of over 250 pesti-
cides, metabolites, and industrial chemicals through
the methodology outlined in Tables 1 and 2. This work
is continuing with several chemicals evaluated each year
according to a defined experimental protocol. The analy-
tical characteristics information is published as several
tables in the Pesticide Analytical Manual Vol. 1.5 For
FDA internal use, the same data have been computerized
for easy up-dating. Included in the computer listing
for each chemical are: common or trade name and FDA
identification number, GLC retention times relative to
aldrin and to parathion on two columns, quantity for
one-half scale recorder deflection for GLC with ECD
and/or AFD and recovery information through the method-
ology for fatty and nonfatty foods. About 175 of the
chemicals checked are completely or partially recovered.
The relative retention time and response data and knowl-
edge of chemicals recovered as well as those which are
not recovered are vital to good qualitative and quanti-
tative interpretation of GLC curves. This information
also benefits the overall evaluation of residue survey
data.
Polar Compounds
The general approach that has been used with the
basic multiresidue scheme just described can also be
applied to methods now available for dealing with multi-
ple residues of pesticides of other chemical types.
Once a basic scheme has Deen chosen for concentrated
effort, studies should be conducted to: (1) assure tnat
adequate extraction of residues from samples is being
achieved; (2) determine the behavior of additional
compounds through the various steps of the method, and
maintain an up-to-date compilation of such data; (3)
improve or expand the method where necessary without
limiting its applicability.
Table 3 outlines a general method for organophos-
phorus pesticides and polar metabolites in plant
materials whicn has been developed along lines similar
to the methodology for organochlorine and other nonpolar
chemicals.37 Sample extraction of nonfatty foods with
acetonitrile is employed by both methods and the GLC
with either AFD or FPD is applied as it is with the
methodology outlined in Tables 1 and 2.
Table 3. Procedure for organophosphorus (polar)
pesticides in nonfatty products
1. 25-100 g sample
2. Blend with acetonitrile or,acetonitrile + water
3. Filter; take aliquot of filtrate
4. Transfer residues from acetonitrile + water to methy-
lene chloride (3 times)
5. Cleanup by adsorption chromatography: charcoal +
Celite + MgO; elute with 5015 acetonitrile/benzene
6. Concentrate eluate to definite volume
7. Determine identity and quantity of residues by GLC
with detector selective to phosphorus: AFD or FPD
8. Confirm identity and ifuantity of residue.
This procedure with methylene Q.hloride extraction of
the aqueous acetonitrile sample extract and charcoal-MgO
Celite column cleanup with acetonitrile/benzene elution
probably recovers,a wide polarity range of pesticide
chemicals including organochlorines and carbamates.
However, the cleanup, which is adequate for GLC with
the phosphorus selective AFD or FPD, is insufficient
for the ECD. Further development of the method to
include a number of important nitrogen containing
compounds has so far Deen inhibited because traces of
acetonitrile remaining in the final concentrated solu-
tion produces an overwhelming solvent peak with the
nitrogen specific electrolytic conductivity GLC detector.
This method has been in use in FDA's surveillance pro-
gram since 1972 for the analysis of a few thousand
fruit and vegetable samples.
5
7-1
-------
A method which detects the dinitrophenyl ether
derivative of aromatic methylcarbamates by ECD-GLC also
utilizes the acetonitrile extraction of plant material.
The carbamates are extracted from the acetonitrile
extract into methylene chloride after salting out the
water and washing the acetonitrile with petroleum
ether.38 cleanup is done by coagulating plant matter
with an aqueous solution of H3PO4. and NH^Cl added to the
methylene chloride extract. Residues of the parent
carbamates are reacted with fluoro-1,2-dinitrobenzene
to form the respective dinitrophenyl ether. This meth-
od has not yet undergone extensive development or
practical use in a food surveillance program, therefore,
its "ruggedness" and need for refinement remain in
question. It appears to have potential as a general
procedure for aromatic carbamates not having an amino
group. However, we feel it preferable for multiresidue
methods to detect and measure the actual pesticide or
metabolite residue rather than a derivative.
' Multiple residue methods have been available for
some time for the chlorophenoxy acid herbicides.
Although used to some extent in food surveillance pro-
grams, the relative low mammalian toxicity of these
herbicides and infrequent findings of significant resi-
dues have minimized use and testing of analytical
methods. The method used by the FDA for these compounds
makes use of the same acetonitrile extraction of plant
material used for the organochlorine, organophosphorus,
and carbamate pesticides.39 Likewise the triazine
herbicides have received little attention in surveil-
lance. Recent developments in multiple residue method-
ology and expected improvements in nitrogen selective
GLC detectors should lead to more attention to this
class of herbicides.40
The methods discussed for organochlorines, organo-
phosphorus, and carbamates respectively have been sub-
jected to interlaboratory collaborative study under
auspices of the Association of Official Analytical
Chemists (AOAC). The methodology for nonpolar chemicals
(Tables 1 and 2) has been the subject of several studies
and has AOAC official status for 14 organochlorine and
5 organophosphorus pesticides in fish, eggs, dairy prod-
ucts, vegetable oils, and a variety of fruits and
vegetables. Based on one collaborative study of each,
the AOAC recognizes the method for organophosphorus
pesticides (Table 3) as official for parathion, para-
oxon, carbophenothion, carbophenothion oxygen analog,
and EPN and the method for carbamates as official for
aarbaryl, propoxur, carbofuran, and carbanolate.41,42
A good estimate of interlaboratory quantitative
capability of the methodology in Tables 1 and 2 can be
gotten from the large number of results from the several
collaborative studies leading to official AOAC status.41
The average of all interlaboratory coefficients of vari-
ation for the 19 pesticides included is 14 percent. The
reported values compared to the true value for the un-
knowns analyzed in these studies averaged 99 percent
(range 66-112) for organochlorine pesticides and 90
percent (range 81-97) for organophosphorus pesticides.
It is interesting that very similar observations with
this methodology were made in a continuing interlabora-
tory quality assurance program conducted by the FDA.43
In the analysis of 59 unknowns, each by about 16 labora-
tories, the average of coefficients of variation was 15
percent, while recoveries of the unknowns averaged 95
percent (range 76-119). In all the AOAC and FDA studies
the best and poorest interlaboratory coefficients of
variation for a single pesticide were 5 and 30 percent,
respectively.
Future Outlook
As we look to the future, we want to encourage
greater attention to the reasonable expansion and
refinement of basically sound and practical multiresidue
methods rather than a proliferation of slightly different
approaches to analysis.44 Following some period of actual
use and informal interlaboratory evaluation, selected
well-defined methods would be subjected to controlled
collaborative testing under auspices of a recognized
organization such as the AOAC. Among the benefits that
we believe would result are: residue data less subject
to question, improved communication among organizations
and countries concerning analytical methods, and a better
basis for introducing pesticide residue control programs
and training analysts in the emerging countries. The
shift that is well underway in many countries from organo-
chlorine to less persistent chemicals needs to be
paralleled by further development of multiresidue method-
ology for their more polar and generally more difficult-
to-determine residues. High speea liquid chromatography
will undoubtedly play a key role, for example with the
aromatic carbamates. The question of residues that are
somehow "bound" to the biological substrate, their signifi-
cance, and associated analytical questions will require
considerable attention.45 Questions of immediate practi-
cality deserve attention, such as the possibility of
substituting methanol or acetone for acetonitrile to
permit the use of a nitrogen selective GLC detector, thus
expanding an existing method to another class of chemi-
cals. And we should not forget that inexpensive and
uncomplicated, yet refined, TLC techniques could be used
to provide acceptable screening for residues in areas
where lack of facilities and high cost prohibit instru-
mentation such as GLC.46 We have in two or three instances
referred to the need for analyst experience and knowledge
for the generation of reliable residue data. This point
cannot be over-emphasized. Analyses of foods for "trace"
residues of toxic organic chemicals are of such complexity
and the findings, both positive and negative, are of such
significance that this responsibility must go only to
dedicated and competent residue analytical chemists, fully
supported in this endeavor by their organizational manage-
ment.
References
1. Lykken, L., Residue Rev. 3_, 19 (1963)
2. Fed. Working Group on Pest. Migmt., Guidelines on
Sampling and Statistical Methodologies for Ambient
Pesticide Monitoring. Washington, D.C., Oct., 1974
3. 40 U.S. Code of Federal Regulations: part 180
4. 21 U.S. Code of Federal Regulations: part 121
5. Food and Drug Administration, Pesticide Analytical
Manual, Vols. I and II. Second edition, 1968.
Revised at irregular intervals. Available on request
from the Food and Drug Administration, Office of the
Associate Commissioner for Compliance, 5600 Fishers
Lane, Rockville, MD 20852
6. Kawar, N.S., et al., Residue Rev. 48, 45 (1973)
7. Fed. Working Group on Pest. Mgmt., Guidelines on
Analytical Methodology for Pesticide Residue Monitor-
ing. Washington, D.C.
8. Mills, P.A., J. Assoc. Offic. Agric. Chem. 42, 734 (1959
9. Burke, J.A., Residue Rev. 34, 59 (1971)
10. Burke, J.A., et al., J. Assoc. Offic. Anal. Chem. 54,
142 (1971) ~"
11. Watts, R.R., J. Assoc. Offic. Anal. Chem. 54, 953 (1971
12. Porter, M.L.. et al., J. Assoc. Offic. Anal. Chem.
52, 177 (1969)
6
7-1
-------
13. Bertuzzi, P.F., et al., J. Assoc. Offlo. Anal. Chera.
50, 623 (1967)
14. Wessel, J.R., J. Assoc. Offic. Anal. Chem. 52, 172
(1969)
15. Wheeler, W.B., et al., J. Agric. Food Chan. 15, 227
(1967)
16. Wheeler, W.B. and Frear, D.E.H., Residue Rev. 16,
86 (1966)
17. Chiba, M., and Morley, H.V., J. Assoc. Offic. Anal.
Chem 51, 55 (1968)
18. Furman, W.B., and Fehringer, N.F., J. Assoc. Offic.
Anal. Chem. 50, 903 (1967)
19. Porter, M.L., et al., J. Assoc. Offic. Anal. Chem.
50, 644 (1967)
20. Bowman, M.C.* and Beroza, M., J. Assoc. Offic. Agric.
Chem. 48, 943 (1965)
21. Bong", R., J. Assoc. Offic. Anal. Chem. 58, 557
(1975)
22. Burke, J.A., et al., J. Assoc. Offic. Anal. Chem.
49, 999 (1966)
23. Mills, P.A., J. Assoc. Offic. Anal. Chem. 51, 29
(1968)
24. Young, S.J.V., and Burke, J.A., Bull. Environ.
Contam. and Toxicol. 7, 160 (1972)
25. Stallings, D.L., et al., Environ. Health Perspec.
No. 3, 159 (1973)
26. Burke, J.A., and Giuffrida. L., J. Assoc. Offic.
Agric. Chem. 47, 326 (1964)
27. Giuffrida, L., et al., J. Assoc. Offic. Anal. Chem.
49, 8 (1969;
28. Mills, P.A,, et al., J. Assoc. Offic. Anal. Chem.
55, 39 V1972)
30. Trotter, W.J., J. Assoc. Offic. Anal. Chem. 58,
461 (1975)
31. Elgar, K.E., Advan. Chem. Ser~ No. 104, 151 (1971)
32. Burke, J.A., J. Assoc. Offic. Agric. Chem. 48, 1037
(1965) ~~
33. Watts, R.R., and Storherr, R.W., J. Assoc. Offic.
Anal. Chem. 5^ 513 (1969)
34. Bostwick, D.C., and Giuffrida, L., J. Assoc. Offic.
Anal. Chem. 51, 34 (1968)
35. Gunther, F.A., Ann. N.Y. Acad. Sci. 160, 72 (1969)
36. Gunther, F.A., Pure Appl. Chan. 21, 355 (1970)
37. Storherr, R.W., et al., J. Assoc. Offic. Anal. Chem.
54, 513 (1971)
38. Holden, E.R., J. Assoc. Offic. Anal. Chem. 56, 713
(1973) """
39. Yip, G., J. Assoc. Offic. Anal. Chem., 54, 966
(1971)
40. Ramsteiner, K., et al., J. Assoc. Offic. Anal. Chem.
57, 192 (1974)
41. Horwitz, W. (ed.), Official Methods of Analysis of
the Association of Official Analytical Chemists, 12th
ed., 1975
42. Changes in Official Methods, J. Assoc. Offic. Anal.
Chem. 58, 397 (1975)
43. Burke, J.A. and Corneliussen, P.E., To be published,
J. Environ. Qual. Safety, 1975
44. Frehse, H., To be published, Pure Appl. Chem., 1975
45. Conference on Bound and Conjugated Pesticides Resi-
dues, June 23-25, 1975, Vail, Colo. Proceedings to
be published in ACS Symposium Series
46. Getz, M.E., "Methods in Residue Analysis" in
Pesticide Chemistry, Vol. 4t Tahori, A.H. (ed.),
11971) Gordon and Breach, New York, pp. 43-63
7
-------
THE DETERMINATION OF PESTICIDE RESIDUES IN AIR
James N. Seiber and James E. Woodrow
Department of Environmental Toxicology
University of California
Davis, California 95616
The atmosphere plays a major role in the movement,
dispersal, and eventual fate of many pesticides. The
methodology and relative merits of various procedures
available for determination of pesticides in this
medium are discussed, with particular emphasis on sam-
pling methods appropriate for monitoring ambient con-
centrations.
Introduction
There has been a growing realization in recent
years that the atmosphere is a major transport route
and reservoir for pesticide residues. To a large
extent this realization has resulted, as with our
knowledge of residue distribution in soil, water, and
biota, from the spectacular advances made in the
development of selective and sensitive detection
techniques so familiar to all residue chemists. But,
while some analogies exist in the measurement of small
concentrations of pesticides in air and condensed
media, each represents a distinct set of problems, and
they have followed quite separate paths of historical
development. To be certain the art of residue analysis
applied to airborne pesticides has lagged significantly
behind that applicable to condensed samples; the
reasons for this lag may be due to the slowness in
comprehending what has only recently become common
knowledge regarding pesticide volatilization mechan-
isms, for lack of legal mandates regulating permissable
levels in air, from the fact that pesticides are not,
with a few exceptions, applied intentionally to the air
(though we have realized for some time that drift is an
inescapable consequence of any application situation),
and from the peculiar problems associated with the
analysis itself. In the course of this brief review of
analytical procedures for pesticides in air, we shall
adopt the point of view that, while a significant start
has now been made in this analytical area, substantial
improvements are to be expected in the future. There
can be little doubt that such improvements will indeed
take place, given the room for improvement and the
apparently boundless ingenuity of residue chemists.
Wheatley1 has summarized the routes of entry of
pesticides to the atmosphere. They Include the inter-
mittent sources associated with drift during appli-
cation, wind erosion of soil, and the burning of
agricultural wastes. Important continuous sources
include volatilization (vaporization) of residues from
leaf, water, soil, and other surfaces, mechanisms which
can hardly be overemphasized as major routes of intro-
duction2. As much as half of agriculturally-applied
chlorinated insecticides may be lost from soil by
evaporation3 * the half-life from evaporation of dis-
solved DDT in surface water is only 3.7 days5; and
Spencer ejt aj..6 have estimated that the majority of
lindane and parathion deposits would disappear by
vaporization in less than one day from mature citrus
tree leaves. These examples were selected out of their
modifying contexts to dramatize the magnitude of this
source. Comprehensive treatment of volatilization
processes may be found in the reviews of Wheatley^,
Hartley3, Hamaker7, and Spencer and Cliath8.
Pesticides may residue in the atmosphere both as
vapors and associated with particulate matter (both
liquid and solid aerosols). Our lack of understanding
of the relative importance of each in ambient situa-
tions may be traced directly to the difficulties
associated with separate analysis of the two forms.
Among the few estimates of distribution to date,
Beyermann and Eckrich9 found the fraction of vaporized
insecticides(lindane, dieldrin, DDD, DDT) to be several
times greater than the fraction bound to dust parti-
cles, from analysis of 15 ambient air samples taken in
Germany; and, in a study of residues of organochlorine
and organophosphorus insecticides from four cities in
the U.S., distribution proportions varied consider-
ably10. These data, limited as they are, are consis-
tent with the expectation that higher molecular weight,
less volatile compounds are more likely
to be associated with particles than lower molecular
weight, more volatile chemicals.
Extremes in behavior are represented by the salts and
esters of 2,4-D; Grover et aJ.11 found no vapor drift
with the essentially non-volatile dimethylamine salt,
while 25-30% of the volatile butyl ester applied was
collected as vapor drift in a short time following
treatment. Relatively clean air (low particulate
matter) tends to favor distribution to the vapor form;
organochlorine compounds in sea air taken near Bermuda
were apparently almost exclusively in the vapor form12,
while half or more of the atmospheric DDT from a pollu-
ted urban environment may be associated with particu-
late matter13. Certainly, the adsorptive nature of the
chemical in question could greatly affect the distri-
bution once equilibrium is attained. No firm data is
available regarding the relative distribution of the
more polar pesticides, but some particle association
would be expected providing the air contains sufficient
solid matter to form an extensive adsorptive surface.
The amounts of pesticides found to date in the air
are susceptible to no useful quantitative generaliza-
tions. They may easily vary by orders of magnitude
depending on the sampling site, season, sampling
method, and investigator. The model of Woodwell et^
al.^ led to calculation of a peak average atmospheric
concentration of DDT of 84 ng/m3 in 1966. This value
is, in fact, within the range frequently observed in
air taken over land but away from application
sites9*10'15. Observed values for samples of sea air
are, however, much lower12'16'17, leading one to
question the distribution calculations, or the calcu-
lated atmospheric time constant of DDT of 3,3 years.
The monitoring program of the U.S. Public Health
Service and Environmental Protection Agency in 1970-
1971, the most comprehensive study of its type carried
out to date, gives an idea of the types and frequency
of occurrence of pesticides expected in ambient air
samples (Table 1). DDT and its relatives, dieldrin,
a- and yBHC, and diazinon were frequently encountered;
a number of other compounds (aldrin, 8-BHC, 2,4-D
esters, heptachlor, malathion, parathion, toxaphene,
and trifluralin) were more localized and seasonal in
occurrence. Maximum levels observed during the sam-
pling period were 6833 ng/m3 (malathion, 1971, Pennsyl-
vania site), 2291 ng/m3 (thiodan, 1970, Kentucky site),
and 1684 ng/m3 (toxaphene, 1970, Mississippi site).
The majority of the reported values were in the 1-10
ng/m3 range. Consideration of these results, and data
from more limited monitoring programs9'10*1s~2i, leads
to the following generalization: Pesticides most
likely to be found in atmospheric samples taken far
from application sites are those which possess a
1
7-2
-------
relatively high degree of chemical stability, have a
reasonable volatility from surfaces {enhanced by low
adsorptivity, low water solubility, and a high vapor
pressure) and are currently or have been in the recent
past used in sizeable quantities. These conditions
match best with the organochlorine Insecticide class,
and they are indeed the most frequently reported. One
can't help but wonder, however, how much this frequency
is affected by availability of sensitive analytical
methods for this pesticide class, and that perhaps
aerial contamination by other heavily used pesticide
groups is not found simply because proper analytical
tools are lacking.
Table 1. EPA Division of Pesticide Community Studies
Monitoring Results, 1970-197115
Samples
Locations
Residues Occurring
in >50% Samples
Residues Occurring
at all Sites
All-1
1970
882
30
p,p'-DDT
p,p'-DDE
Dieldrin
o,p'-DDT
Ct-BHC
Y-BHC
Diazinon
p,p'-DDT
o, p' -DDT
Dieldrin
p,p'-DDE
1971
968
45
p,p'-DDT
p,p'-DDE
Dieldrin
o,p'-DDT
a-BHC
¦y-BHC
p,p'-DDT
Dieldrin
analysis experiment.
Static Sampling Methods
Methods in which no mechanical concentration of
the air precedes determination, i.e. static methods25,
have frequently been employed to determine pesticide
levels in air. Some of the first such methods were
based on analysis o£ atmospheric precipitation. For
example, low concentrations of several organochlorine
insecticides were observed in samples of rainwater
taken from the British Isles26-28. And Peterley29
calculated the accumulation of DDT residues in Antarc-
tic snow by analysis of snow-melt.
Most static sampling methods rely on contact of
moving air with a surface which is frequently modified
to stabilize impacted material. For example, Rise-
brough et_ al^. 16 used nylon nets coated with glycerine
to determine background levels of some organochlorines
in trade winds sampled at Barbados, and residues of
some organochlorine materials were determined in Sweden
by fallout on SE 30-impregnated nylon nets30. Seba and
Prospero17 estimated that under typical trade wind con-
ditions (average velocity 25 km/hr) on the order of a
million m3 of air could be contacted by 3 m2 of nylon
mesh in 24 hr, and assumed that 50% of the aerosol
particles above several microns diameter would be
retained. Nylon chiffon (0.5 m2) saturated with ethy-
lene glycol and suspended in a wooden frame provided a
convenient noise- and external power-free device for
sampling air in homes'
Side by side comparison of
Methods of Sampling and Analysis
The steps involved in any pesticide analysis for
monitoring purposes include (1) sampling the environ-
ment of interest in a valid and representative way,
with suitable checks and replicates, (2) selective and
efficient removal of the pesticides of interest from
extraneous matter, and their placement in a form
amenable to further examination, (3) cleanup and frac-
tionation of the resulting samples in such a way that
the chemicals of interest become concentrated and
separated into most easily determined groups (4) the
actual detection and quantitative measurement, and (5)
adequate confirmation of results. These are elementary
considerations for chemists concerned with analysis of
soil, water, plant and animal tissue but they have
received surprisingly little attention in connection
with air monitoring. For example, detailed pesticide
sampling protocols for crops, soil, water, and^other
commodities are reasonably well established ' , while
quidelines for air sampling are practically non-
existent. It is not to be expected that a single air
sample obtained from a rooftop in a downtown area
during midday could possibly reflect the pesticide
level in an entire community, yet this type of extra-
polation is often made. Furthermore, given the rudi-
mentary nature of some sampling devices, it is shocking
to find that few studies Included provision for repli-
cation and compositing. As a matter of fact, air
concentrations have frequently been reported using
methods of completely unknown efficiencies. Adequate
experimental design should certainly include embodiment
of spatial and time factors in a thoughtful sampling
grid, an accurate record of all pertinent meteoro-
logical conditions, sufficient replication for statis-
tical treatment of results and validation of the method
with check and spike, runs. While we will concentrate
in this review on the mechanics of the actual methodo-
logy, our comments will be predicated on the assumption
that a proper design accompanied the sampling and
this device with two ethylene glycol impingers showed,
after four hours exposure (2.5 m3of air processed by
the impingers), about four times as much DDT in a
single nylon cloth as in the two impinger solutions
combined. These static systems appear quite suitable
for comparing levels from one location to another.
Their sampling efficiency and retention ability for
pesticides as small particles and vapors is not known,
and of course they do not allow for direct determina-
tion of concentration. It seems, however, that with
the simple addition of a constant velocity air mover
(i.e. a fan) in their path, and careful calibration
against impinger techniques, they could yield quantita-
tive information which would be quite suitable for many
purposes.
Static sampling methods of, more restricted use
include plastic drop sheets for drift studies and
determination of vegetation damage following drift32.
Modern in situ spectroscopic methods developed for air
pollutant analysis33 have not been adapted to pesticide
determination; they could find advantageous use with
the more volatile low molecular weight fumigants in
some field situations.
Dynamic Methods
Four basic components constitute a generalized
scheme for dynamic air sampling2s:
Air
Inlet System
Collection
Device
Air Flow
Measurement
Air Mover
and Exit
The key to successful deployment of this methodology
lies in the proper selection of the collection device.
Reported methods for sampling pesticides in air roughly
fall into categories, defined by the air volume pro-
cessed per unit time, of "low-volume samplers" (10-30
1/min, approximately 0.3-1 ft3/min) and "high-volume
samplers" (0.2-1 m3/min). Our discussion of collection
devices will follow this somewhat artificial division.
Particulate matter is generally sampled by filtra-
tion, retained particles constituting tha a«!iple fat
analysis, while with vapors a concentration atepi by
7-2
-------
solution absorption or surface adsorption, is an integ-
ral part of the sampling step. Thus collection devices
may consist of two components, often in tandem in the
air stream, capable of separately collecting particu-
lates and vapors. The term "sampling efficiency"
refers to the ability of the component(s) of the
collection device to sample, extract (in the case of
absorbents and adsorbents), and retain the chemicals of
interest. Determination of sampling efficiency
involves spiking inlet air with known amounts of chemi-
cals as dusts and/or vapors to determine recovery.
Failure to determine this parameter under all condi-
tions of use, and with all the chemicals of interest
can negate several months or years of laborious field
work. Sampling efficiency is distinct from the percent
recovery familiar to residue chemists determined in the
course of spiking experiments with water, soil, fruits,
etc. It is a much more difficult value to obtain since
it is influenced by the variables of time, flow rate,
temperature, humidity, and the like associated with the
sampling process. Additional considerations regarding
sampling and post-sampling workup have been presented
elsewhere 3 \
Table 2. Some media for low volume dynamic air sam-
pling of pesticides and related chemicals.
Air Flow
Chemicals
Medium
1/min
Sampled
Ref
Solvents
DMF
15
CH
20
DMF
12
fenitro
35
Water
10.3
DDVP
36
EG
10.3
CH, OP
36
EG
23-28.5
CH, OP,
34,37
carb, herb
HG
29
CH, OP
10
EtOAc
2
DDVP
38
Adsorbents
Alumina
29
CH, OP
10
Chrom 101
3
HCB, HCBD
39
Chrom 102
8-10
CH, OP, herb
40
Florisil
12
fenitro
35
Sil gel
10-14.5
2,4-D
11,41
Coated solids
ODS/Chrom
PE on Sil Gel
PG on SS nets
4-20
1.4-28
16-190
CH
CH, OP, herb
CH, OP, herb
42
43
9,44
Studies air monitoring program in 1970-71. The method,
briefly, was as follows"*5 :
Two impingers, each charged with 100 ml of pesti-
cide-quality ethylene glycol and with a glass cloth
filter preceding the impinger, were operated simul-
taneously for 12 hr at 28.5 1/ml using a special
sequential sampler10. Two additional impinger-filter
combinations were run simultaneously for a second 12-hr
period. The hexane extracts of the four filters and
400 ml of glycol were then pooled to give a sample
representing 80 m3 of air for the 24-hr period. The
extract was then subjected to Florisil column cleanup,
the eluate being collected in two or three fractions
which were analyzed by gas chromatography using elec-
tron capture and flame photometric detectors. Detec-
tion limits ranged from 0.2 ng/m3 for a-BHC to 167
ng/m"
ng/m3
for perthane, and were commonly in the 0.5-2.0
range for most organochlorine and organophos-
phorus compounds, and in the 1-50 ng/m range for 2,4-D
and 2,4,5-T esters.
The validity of some aspects of the impinger
method may be open to question. The glass cloth filter
does not provide a clean separation of particulate-
bound pesticides from vapors even if analyzed sepa-
rately from the impinger contents. Small particles
undoubtedly pass through the filter, and Miles et al.36
showed that some compounds (diazinon, DDVP) introduced
as vapors were partially collected on glass fiber fil-
ters. Furthermore, once trapped on the filter, chemi-
cals may revolatilize or degrade. The latter may be
particularly troublesome with easily oxidized pesti-
cides, such as those of the thiophosphate class. The
ability of a single ethylene glycol-charged impinger to
efficiently extract and retain pesticide vapors during
the 12 hr sampling period varies considerably with the
chemical. In one series of tests efficiencies ranged
from a low of 20% (aldrin) to 90% (parathion)'
Effi-
ciencies for most of the 23 pesticides checked were in
the 50-80% range. Both incomplete extraction and
revolatilization contribute to less than quantitative
retention. Hydrolytic or oxidative degradation was not
observed in any appreciable amounts. Greatest losses
are observed with the more volatile constituents of
mixtures, as expected; for example, the two most
rapidly-eluting major gas chromatographic peaks of
technical chlordane ("Diels-Alder adduct" and hepta-
chlor) were sampled with 39 and 38% efficiencies during
a 12-hr vapor spiking run, while efficiencies ranged
from 67-84% for the six later-eluting major peaks .
Abbreviations: Carb, carbamates; CH, chlorinated
hydrocarbons; Chrom, Chromosorb; 2,4-D, 2,4-dichloro
phenoxy acid and esters; DDVP, dichlorvos; DMF,
dimethyl formamide; EG, ethylene glycol; EtOAc, ethyl
acetate; fenitro, fenitrothion; HCB, hexachlorobenzene;
herb, herbicides; HCBD, hexachloro-1,3-butadiene; HG,
hexylene glycol; 0DS, octadecylsilyl; OP, organophos-
phates; PE, polyethylene, PG, polyethylene glycol; SS,
stainless steel.
Several devices have been used to collect pesti-
cides during low volume, dynamic air sampling [Table 2;
for additional references prior to 1970 see Miles et_
a^.36]. Of these glycol-filled Greenburg-Smith impin-
gers have been the most thoroughly tested and widely
used for ambient monitoring. Miles et_ al_.36 charac-
terized the sampling efficiency of the small impinger
filled with 25 ml of ethylene glycol toward both
vapors and dusts of several pesticides. A later study
employed a larger version of the impinger filled with
100 ml of hexylene glycol; an alumina adsorption tube
was added to trap compounds not retained in the filter,
and a glass cloth filter preceded the impinger to
remove dust10. This basic sampling sequence, modified
by Enos et^^l-37 was used in the U.S. PHS-EPA Community
Method recoveries from extraction of the glycol
through the modified Mills cleanup were greater than
85% for the pesticides tested by Thompson1*5. An alter-
nate extraction-cleanup scheme, using methylene
chloride extraction and silica gel column cleanup and
fractionation, gave method recoveries of 75-100% for
most of the 25 chemicals checked1*7. A major advantage
of the micro silica gel procedure is its allowance for
collection of carbamate insecticides, compounds not
recovered from the Florisil cleanup method. Using
either cleanup method the glycol impinger system may be
employed with some confidence for determining a rela-
tively wide range of pesticides at levels in excess of
0.1 ng/m3, with efficiencies considering both sampling
and post-sampling workup of approximately 40-70%. Its
major disadvantages include no provision for separate
analysis of vapors and particulates, the possibility of
degradation of labile compounds on the filter, restric-
tions of air flow to ca 30 1/min, and the need for
relatively expensive and fragile glassware and an ex-
pensive (ca $1600) sequential sampling device for each
sampling site.
There have been several attempts to develop alter-
nate sampling methods (Table 2) but none have received
-------
the rigorous testing attention devoted to the glycol
impinger. Inorganic adsorbents such as Florisil,
silica gel, alumina, and charcoal may be useful for
specific chemicals. They do not, however, promise
broad utility because of their propensity to change
activity when exposed to air, their potential for pro-
moting surface-catalyzed degradation, and difficulties
in obtaining complete post-sampling extraction of
strongly adsorbed chemicals. Liquid phase-coated
solids seem less likely to suffer from these potential
disadvantages but may contribute problems of a differ-
ent type from their solubility in the same solvents
used in the post-sampling extraction and cleanup of
pesticides of interest.
Organic solids of high surface area, such as the
porous polymeric resins of the Chromosorb 100, Poropak,
and XAD series offer a compromise. These materials
have shown considerable utility for sampling volatile
constituents of biological interest1*8 and hydrocarbons
of interest in the pollution field1*9. It is not sur-
prising to find that they satisfactorily trap vapors of
a variety of pesticides"0 and related chemicals3 . For
example, one gram of Chromosorb 102 trapped £a 0.1-10
Ug quantities of pesticides ranging in volatilities
from lindane, trifluralin, and diazinon, to DDT and
methoxychlor in as much as 10 m3 air volumes with 80%
and greater efficiencies'*0. The air volume:medium
weight ratio resulting in pesticide sampling efficiency
of greater than ca 50% defines a parameter (hereafter
referred to as simply the vw ratio) which may be useful
in comparing sampling media and conditions. It may be
calculated for Chromosorb 102 as follows:
vw ratio = Sampling rate (tn3/min) x Sampling
duration (min) x 1/media weight (g)
- 0.01 m3/min x 1000 min x 1/lg
- 10 m3/g
By contrast the vw ratio for a single ethylene glycol
impinger operated under the usual conditions (28.5
1/min, 12 hr) is ca 0.2 m3/g. The ratios tell us, for
one thing, that if both sampling media have been
reduced to equivalent analytical background levels by
pre-sampling cleanup, and equivalent allquots of final
extracts are taken for analysis, then the analytical
reagent background from ethylene glycol will be nearly
50 times that from Chromosorb 102. Furthermore, if it
is assumed that both media absorb background materials
from air in proportion to their mass, air-derived back-
ground could be correspondingly greater for ethylene
glycol than for Chromosorb 102. Since detection limits
are set by background50 one might expect to attain
lower detection limits with Chromosorb 102 than with
ethylene glycol, even with the great uncertainty in-
herent in the assumptions. The vw ratio may find
further utility in choosing proper amounts of sampling
media in scaling up to higher flow rates. .
The data of Beyermann and Eckrich9" l"t developed
with a film of polyethylene glycol spread on stainless
steel nets are amenable to a similar calculation.
Their parameters for 90% sampling efficiency lead to
calculation of a vw ratio of 4 m /g, again an order of
magnitude higher than for the glycol impinger.
The use of solid or liquid-coated solid sampling
media offers flexibility in sampler design. For
example, a 1.9 cm (od) x 12.7 cm glass tube containing
4 g of 60/80 mesh Chromosorb 102 is compatible with an
air flow of 8-10 1/min1,0. Higher flow rates can be
accommodated by an increase in particle diameter, a
larger diameter container, or a combination of the two.
Beyermann and Eckrich9 employed a tandem arrangement of
a 5 cm diameter glass filter and ten 5 em diameter
coated nets. A flow of 200 1/min was attainable
through this system, though it should be noted that the
sampling efficiency was low at this flow rate.
Many of the efforts aimed at developing monitoring
methods for pesticides in ambient air have pursued a
high-volume sampling approach using solid or liquid-
coated solid media. The rationale for this approach is
simply that a larger sample should allow detection and
measurement of smaller pesticide concentrations than
possible with low-volume systems, and that a shorter
sampling duration, if needed for a particular study,
would be possible. A high-volume method for sampling a
variety of pesticides employed glass beads coated with
3 ml of cottonseed oil22'51. Air was sampled at 280
1/min, approximately ten times the rate of the large
glycol impinger. Collection efficiencies were high for
aerosols and exceeded 70% for the vapors of eight pes-
ticides during 2 hr sampling tests. Despite the high
vw ratio (11 m3/g) there has been some question about
electron capture background from traces of cottonseed
oil ( and its oxidation products) not removed by post-
sampling partition cleanup. Perhaps a more inert
liquid, such as paraffin oil or a glycol, would over-
come this problem while preserving the merits of this
basic approach.
In fact, paraffin oil was used successfully as a
liquid phase (5%) on 20/30 mesh Chromosorb A to monitor
concentrations of trifluralin and its breakdown
products in the air near treated fields52. The coated
phase (30 g) was held in a filter-holder for low-volume
sampling, or in the neck of a commercial high-volume
sampler. Sampling efficiencies for vapor spikeB of 1
and 25 ug were in excess of 40% in runs at 1 m3/min
flow rates and 2 hr duration (vw ratio « 4 m'/g based
on 30 g of coated phase, and 80 m3/g based on paraffin
oil). Trifluralin is more volatile than many of the
common pesticides of monitoring interest, and an even
better performance might be expected for compounds such
as DDT and its relatives. A potential drawback of this
system—in fact any system which has an inorganic
absorbent as part of the sampling medium—is the ten-
dency for oxidative breakdown of labile compounds on
its surface. We have found that, while parathion is
sampled with high efficiency, significant amounts of
paraoxon (up to 30% of the trapped material) are formed
on the medium surface1*6.
XAD-4, a porous polystyrene resin commercially
available in 20/50 mesh particle sizes, appears to be
a suitable replacement for coated Chromosorb A. 3
Grams of prepurlfied XAD-4 gave >50% sampling effi-
ciency for parathion vapors in 5 hr tests at 1 mVmln
(vw ratio 10 m'/g), and less than 10% breakdown to
paraoxon. We are in the process of investigating its
sampling capabilities towards other compounds; based on
the efficiency of a similar resin, Chromosorb 102, the
prospects appear good.
Bidleman and Olney12'5 3 described an elegant
application of a polymeric trapping agent to a high-
volume sampling approach which well illustrates the
advantage of a high vw ratio. In their system two 10
x 5 cm cylindrical plugs of polyurethane foam (density
0.021 g/cm3) were held in aluminum containers through
which air was pulled at 0.5-1 m'/min by a Hurricane
pump. An 8 x 10 in. glass fiber filter preceded the
first plug, and a 10 ft. section of flexible hosing
allowed placement of the probe away from the pump ex-
haust. Calibrating recoveries were in excess 8f 90% on
the first plug for tri-, tetra-, and pentachlorobi-
phenyls (PCB's) sampled for approximately 24 hr. The
consequence of the exceptionally high vw ratio (90
m3/g) is apparent in the small quantities of PCB's (0.5
ng/ms), DDT (0.01 ng/m3), and chlordane (0.01 ng/m')
which could be accurately quantitated in ambient air
samples. This is at least an order of oa&aitude
4
7-2
-------
improvement in detection limits over the glycol
impinger system.
Recommendations and Conclusions
The use of solids as trapping media opens alterna-
tive approaches to the conventional ethylene glycol
impinger for sampling pesticides in ambient air.
forous polymers particularly offer advantages in design
flexibility, broad-spectrum applicability, high sam-
pling volume capacity (reflected in their vw ratios),
and apparently low detection limits. It should be
emphasized, however, that no rigorous and systemic
testing of these materials has yet been completed, and
it would thus be premature to give an unqualified
endorsement to their use at this point in time.
For one thing, the possibility of accomplishing
monitoring objectives with low-volume, rather than
high-volume, sampling through solids has not received
sufficient consideration. One could envision use of a
small diameter tube containing a gram or so of porous
polymer through which air was sampled at ca 10 1/min,
providing a 10 m3 sample in less than 20 hr. Sampled
pesticides could be removed by simply rinsing the
polymer with a small volume of solvent, followed by
concentration to ca 0.1 ml and analysis of aliquots (10
til) on gas chromatographs equipped with C1-, P-, S-,
and N-selective detectors. Suitable detectors (flame
photometric, alkali-flame, and conductivity) are avail-
able for which 1 ng or less gives a measurable res-
ponse, corresponding to a theoretical detection limit
of 0.1 ng/m3. This same sample size is also compatible
with many mass spectrometric confirmation techniques.
Thus most of the objectives one might envision for a
monitoring program could be realized: The use of
small, perhaps portable, and inexpensive pumps would
make replication possible with a small capital outlay;
personnel requirements would be low, since many traps
could be prepared simultaneously for field deployment,
and very few operations would be needed in the labora-
tory; and at this flow rate a filter ahead of the vapor
sampler might be expected to give a minimum of release
of particulate-bound pesticides to the vapor trap, so
that estimation of proportions of vapor and particulate
pesticides would be possible. The exclusive use of
selective detectors is proposed so that cleanup and
fractionation could be bypassed; if this proves not to
be feasible, a microcolumn fractionation1*7 could be
inserted in the analytical scheme.
Alternatively, the high volume approach would not
tax analytical instrument detection limits and would
allow for shorter sampling durations. Commercial high
volume samplers, already in service in most air moni-
toring stations, could process 120 m3 of air in two
hours. Thus several samples could be collected each
day at the same station, serving as a repl1 '¦".tion of
sorts.
The advantages and shortcomings of these and other
alternatives should be thoroughly examined before a
uniform protocol for monitoring pesticides in air
becomes finalized.
References
1. Wheatley, G.A. Pesticides in the atmosphere. In:
Environmental Pollution by Pesticides. C.A.
Edwards, editor, New York, Plenum Press, 1973
Chapter 10.
2. Freed, V.H. , R. Haque, and D. Schmedding. Vapori-
zation and environmental contamination by
DDT. Chemosphere jL:61 (1972).
3. Hartley, G.S. Evaporation of pesticides. In:
Pesticide Formulations Research. R.F. Gould,
editor, Washington, American Chemical
Society, Adv. Chem. Ser. f56:U5, 1969.
4. Lloyd-Jones, C.P. Evaporation of DDT. Nature
229:65 (1971).
5. Mackay, D., and A.W. Wolkoff. Rate of evaporation
of low-solubility contaminants from water
bodies to atmosphere. Environ. Sci. Technol.
2:611 (1973).
6. Spencer, W.F., W.J. Farmer, and M.M. Cliath. Pes-
ticide volatilization. Res. Revs. 49:1
(1973).
7. Hamaker, J.W. Diffusion and volatilization. In:
Organic Chemicals in the Soil Environment.
C.A.I. Goring and J.W. Hamaker, editors, New
York, Marcel Dekker, 1972, Chapter 5.
8. Spencer, W.F., and M.M. Cliath. Vaporization of
chemicals. In: Environmental Dynamics of
Pesticides. R. Haque and V.H. Freed,
editors, New York, Plenum Press, 1975, pp.
61-78.
9. Beyermann, K. , and W. Eckrich. Trennung des
Insecticidgehaltes der luft in den aerosol-
gebundenen und den gasformigen anteil. Z.
Anal. Chem. 269:279 (1974).
10. Stanley, C.W., J.E, Barney, M.R. Helton, and A.R.
Yobs. Measurement of atmospheric levels of
pesticides. Environ. Sci. Technol. .5:430
(1971).
11. Grover, R., J. Maybank, and Y. Yoshida. Droplet
and vapor drift from butyl ester and
dime thylamine salt of 2,4-D. Weed Science
20:320 (1972).
12. Bidleman, T.F., and C.E. Olney. Chlorinated
hydrocarbons in the Sargasso Sea atmosphere
and surface water. Science 183:516 (1974).
13. Atkins, D.H.F., and A.E.J. Eggleton. Studies of
atmospheric washout and deposition of gamma-
BHC, dieldrin, and p,p-DDT using radio-
labelled pesticides. AERE-R6528 1AEA/SM-
142/32. 1970. Quoted by C.P. Lloyd-Jones,
Atmos. Environ. 6/.283 (1972).
14. Woodwell, G.M., P.P. Craig, and H.A. Johnson. DDT
in the biosphere: where does it go? Science
174:1101 (1971).
15. Yobs, A.R., J.A. Hanan, B.L. Stevensen, J.J.
Boland, and H.F. Enos. Levels of selected
pesticides in ambient air of the United
States. (024-Pesticide Chem.), Division of
Pesticide Chemistry, 163rd Meeting, ACS,
Boston, Massachusetts. April, 1972.
16. Risebrough, R.W., R.J. Huggett, J.J. Griffin, and
E.D. Goldberg. Pesticides: Transatlantic
movements in the northeast trades. Science
159:1233 (1968).
17. Seba, D.B., and J.M. Prospero. Pesticides In the
lower atmosphere of the northern equatorial
Atlantic Ocean. Atmos. Environ. .5:1043
(1971).
18. Tabor, E.C. Pesticides in urban atmospheres. J.
Air Pollut. Contr. Ass. 15:415 (1965).
19. Tabor, E.C. Contamination of urban air through
the use of pesticides. Trans. N.Y. Acad.
Sci. Ser. IX. 28:569 (1966).
20. Abbott, D.C., R.B. Harrison, J.O'G. Tatton, and J.
Thomson. Oeganochlorine pesticides in the
atmosphere. Nature 211:259 (1966).
5
7-2
-------
21. Jegier, Z. Pesticide residues in the atmosphere.
Ann. N.Y. Acad. Sci. 160:143 (1969).
22. Compton, B., P.D. Bazydlo, and G. Zweig. Field
evaluation of methods of collection and
analysis of airborne pesticides. Summary
report to U.S. Environmental Protection
Agency, EPA Contract CPA 70-145, SURC Project
L1073-05. May, 1972.
23. Garber, M.J. Statistical evaluation of results
and sampling procedures. In: Analytical
Methods for Pesticides, Plant Growth Regula-
tors, and Food Additives. G. Zweig, editor,
New York, Academic Press, 1963. Volume I,
Chapter 19.
24. Federal Working Group on Pest Management. Guide-
lines on sampling and statistical methodolo-
gies for ambient pesticide monitoring.
Washington, D.C. 1974.
25. Stern, A.C., H.C. Wohlers, R.W. Boubel, and W.P.
LowryFundamentals of Air Pollution. New
York, Academic Press, 1973, Chapter 14.
26. Wheatley, G.A., J.A. Hardman. Indications of the
~ presence of organochlorine insecticides in
rainwater in central England. Nature
"207:486 (1965).
27. Abbott, D.C., R.B. Harrison, J.O'G. Tatton and J.
Thomson. Organochlorine pesticides In the
atmospheric environment. Nature 208:1317
(1965).
28. Tarrant, K.R., and J.O'G. Tatton. Organochlorine
pesticides in rainwater in the British Isles.
Nature 219:725 (1968).
29. Peterle, T.J. DDT in Antarctic snow. Nature
224:620 (1969).
30. Sodergren, A. Chlorinated hydrocarbon residues in
airborne fallout. Nature 236:395 (1972).
31. Tessari, J.D., and D.L. Spencer. Air sampling for
pesticides in the human environment. J. Ass.
Off. Anal. Chem. .54:1376 (1971).
32. Akesson, N.B., and W.E. Yates. Problems relating
to application of agricultural chemicals and
resulting drift residues. Ann. Rev. Entomol.
9:285 (1964).
33. Hodgeson, J.A., W.A. McClenny, and P.L. Hanst.
Air pollution monitoring by advanced spectro-
spectroscopic techniques. Science 182:248
(1973).
34. Seiber, J.N., J.E. Woodrow, T.M. Shafik, and H.F.
Enos. Determination of pesticides and their
transformation products in air. In:
Environmental Dynamics of Pesticides. R.
Haque and V.H. Freed, editors, New York,
Plenum Press, 1975, pp. 17-43.
35. Yule, W.M., A.F.W. Cole, and I. Hoffman. A survey
for atmospheric contamination following
forest spraying with fenitrothion. Bull.
Environ. Contamin. Toxicol. 6.:289 (1971).
36. Miles, J.W., L.E. Fetzer, and G.W. Pearce.
Collection and determination of trace quanti-
ties of pesticides in air. Environ. Sci.
Technol. 4:420 (1970).
37. Enos, H.F., J.F. Thompson, J.B. Mann, and R.F.
Moseman. Determination of pesticide residues
in air. (022-pesticide Chem.), Division of
Pesticide Chemistry, 163rd Meeting, ACS,
Boston, Massachusetts. April, 1972.
38. Collins, R.D., and D.M. DeVries. Air concentra-
tions and food residues from use of Shell's
No-PestR insecticide strip. Bull. Environ.
Contamin. Toxicol. 9:277 (1973).
39. Mann, J.B., H.F. Enos, J. Gonzalez, and J.F.
Thompson. Development of sampling and
analytical procedure for determining hexa-
chlorobenzene and hexachloro-1,3-butadiene in
air. Environ. Sci. Technol. 8^:584 (1974).
40. Thomas, T.C., and J.N. Seiber. Chromosorb 102, an
efficient medium for trapping pesticides from
air. Bull. Environ. Contam. Toxicol. 12:17
(1974).
41. Que Hee, S.S., R.G. Sutherland, and M. Vetter.
GLC analysis of 2,4-D concentrations in air
samples from central Saskatchewan in 1972.
Environ. Sci, Technol. 9:62 (1975).
42. Aue, W.A., and P.M. Teli. Sampling of air pollu-
tants with support bonded chromatographic
phases. J. Chromatogr. 62^:15 (1971).
43. Herzel, F., and E. Lahmann. Polyethylene-coated
silica gel as a sorbent for organic pollu-
tants in air. Z. Anal. Chem. 264:304
(1973).
44. Beyermann, K., and W. Eckrich. Gas-chromatogra-
phlsche bestimmung von insecticid-spuren In
luft. Z. Anal. Chem. 265:4 (1973).
45. Thompson, J.F. (ed). Analysis of Pesticide Resi-
dues in Human and Environmental Samples.
Primate and Pesticides Effects Laboratory,
U.S. Environmental Protection Agency,
Perrine, Florida. 1972.
46. Seiber, J.N. and J.W. Woodrow. Unpublished
results.
47. Sherma, J., and T.M. Shafik. A multiclass, multi-
residue analytical method for determining
pesticide residues In air. Arch. Environ.
Contam. and Toxicol. 3^:55 (1975).
48. Zlatkis, A., H.A. Llchtenstein, and A. Tlshbee.
Concentration and analysis of trace volatile
organlcs in gases and biological tissues with
a new solid adsorbent. Chromatographia 6:67
(1973).
49. Dravnieks, A., B.K. Krotoszynski, J. Whitfield, A.
O'Donnell, and T. Burgword. High-speed
collection of organic vapors from the atmos-
phere. Environ. Sci. Technol. 5:1220
(1971).
50. Sutherland, G.L. Residue analytical limit of
detectability. Rea. Revs. 10:85 (1965).
51. Compton, B., and J. Bj.orkland, Design of a high-
volume sampler for airborne pesticide collec-
tion. (021-Pesticide Chem.), 163rd Meeting,
Division of Pesticide Chemistry, ACS, Boston,
Massachusetts. April, 1972.
52. Soderquist, C.T., D.G. Crosby, K.W. Moilanen, J.N.
Seiber, and J.E. Woodrow. Occurrence of
trifluralln and its photoproducts in air. J.
Agr. Food Chem. 23:304 (1975).
53. Bidleman, T.F., and C.E. Olney. High-volume
collection of atmospheric polychlorinated
biphenyls. Bull. Environ. Contam. Toxicol.
11:442 (1974).
6
7-2
-------
ANALYSIS OF PESTICIDE RESIDUES IN FIELD SOILS: OPTIMIZING SOIL
SAMPLING AND PESTICIDE EXTRACTION
Joseph H. Caro and Alan W. Taylor
Agricultural Chemicals Management Laboratory-
Agricultural Environmental Quality Institute
U.S. Department of Agriculture
Beltsville, Maryland 20705
Because pesticides incorporated into the soil are nonuniformly distributed both horizontally and vertically,
soil sampling to adequately represent the treated field is more difficult than is generally believed. This
paper compares alternative sampling methods, with emphasis on our experiences gained in field experiments over
a 9-year period. The most precise method consists of high-density core sampling at small sites and compositing
of sampleB from multiple random sites.
Pesticide residues in soil samples have been extracted by a variety of techniques and with numerous organic
solvents or combinations of solvents. The various options are reviewed with respect to techniques and solvent
systems and recommendations axe made for optimizing the extraction.
Modern analytical procedures for determination of
pesticide residues in substrates such as soil, plant
material, food, feed, water, animal tissue, or air
generally include a common basic sequence of oper-
ations: (i) sampling of the substrate; (ii) extraction
of the pesticide of Interest from the sample; (iii)
removal of unwanted substances from the extract in a
cleanup step; (iv) quantitation of the pesticide in
the clean extract by gas chromatography or other
appropriate technique} and (v) confirmation of the
identity of the sought-for residue. Sampling and
extraction differ with type of substrate, whereas
the other operations are essentially identical for
all substrates. Since our concern here is with soil
analysis, we will examine the sampling and extraction
steps, indicating what we feel is the best approach
to each, based on results gained in our research
program (sampling), and the experiences of others
(extraction).
Sampling of Soils
Substantial problems are inherent in obtaining
representative soil samples from a pesticide-treated
field. Consider the conditions: a chemical is
¦prayed or dropped onto the soil surface, in a less-
than-uniforra application, at a rate that often
corresponds to a concentration in the soil of 1 or
2 parts per million or less; frequently, the chemical
is then incorporated into the soil by discing or a
similar operation, which increases the heterogeneity
of distribution. After application, the pesticide
is subject to chemical decomposition, microbiological
degradation, volatilization, plant uptake, and loss
in runoff water and sediment, all of which may proceed
at different rates in different parts of a field.
Clearly, any individual analysis to determine a
pesticide residue level in the soil can reflect only
the condition at that particular sampling point and
at the particular time the sample was taken. Further-
more, usually less than a one-millionth part of the
whole substrate is analyzed in the laboratory and the
pesticide is highly and irregularly diluted in the
analytical sample. Results over a field area may
therefore be expected to vary considerably. Con-
sequently, the only meaningful expression of pesticide
level must be based on a number of samples taken
from various places in the field, and must take the
sampling variability into account. During a 9-year
program of field studies, we have tested a number
of techniques for collection and compositing of
samples and can now report a procedure that gives
relatively precise results without the necessity for
analyzing an inordinately large number of samples.
Table 1 describes 10 field experiments conducted
by our pesticide scientists since 1966. The objective
of all the experiments was to ascertain the fate of
the chemicals after application to the soil under
normal agricultural conditions. Details and results
of many of the experiments have been reported , and
Table X. Description of field experiments
>eriment
Location and
Rate of
Plot
No.
Year
Pesticide
Application
Area
kg/ha
ha
1.
Coshocton, 1966
Dleldrin
5.6
1.08
Ohio
2.
Coshocton, 1968
Dieldrin
5.6
0.63
3.
Coshocton, 1969
Dieldrln
5-6
0.68
It.
Coshocton, 1969
Heptachlor
5.6
0.68
5.
Coshocton, 1971
Carbofuran
5.1*
0.59
6.
Coshocton, 1971
Carbofuran
h.2
0.79
7.
Coshocton, 1972
Carbofuran
3.1
0.59
8.
Coshocton, 1973
Carbaryl
5.0
0.79
9.
Ithaca, 197b
Trifluralin
0.67
2.97
New York
10.
Ithaca, 197*»
Trifluralin
0.80
2.97
Conditions
of Application
Emulsion,
disced in
Granules,
disced in
Granules,
Emulsion,
disced in
Granules,
disced in
broadcast and
to 7.5-cm depth
do
do
do
broadcast and
to 7.5-cm depth
in seed furrow
do
do
broadcast and
to 7.5-ca depth
broadcast and
to 7.5-cm depth
7-3
-------
Table 2. Sampling schemes and analytical results of field experiments
ft J
No. of Depth Sampling Pesticide Content—
&
Samples
of
Technique
Uo.F
Std.
Coeff. of
Analyzed
Sampling.
Units
Ranse
Mean
Dev.
Variation
1
a
CM
17.5
1
ppm
2.11-9.83
¦J+,06
2.1+9
%
61.2
2
5
17.5
1
ppm
1.82-1+. 19
3.16
0.91
28.6
3
5
17.5
2
mg/nr
185-381
261
7U.1
28.1+
1»
5
17.5
2
mg/mi
196-378
265
68.6
25.9
5
2U
10.2
3
rag/m
215-726
1+04
126.2
31.3
6
16
10.2
3
mg /m*
365-1508
775
352.8
»»5.5
7
10
15.2
3
m/nz
571-1203
830
212.2
25.6
8
10
20.3
3
mg/m^
1085-2086
1579
303.0
19.1
9
5
15.2
k
mg/m|
35-!+5
38.4
3.9
10.0
10
5
15.2
It
mg/m
37-72
51.7
13.2
25.*
—^See Table 1 for description of experiments
For complete description of sampling techniques, see text. Numbers refer to the
described techniques;
1 * spade slice compositing
2* m random core compositing
2 a 900-cm2 area excavation
5" = compositing of closely spaced cores taken at random sampling points
£¦'All samples taken on day of pesticide application
the question of sample variability in determining
soil concentrations of the pesticides has been
examined5'6. The suitability of the various sampling
techniques used may be Judged by comparing results
obtained on soil samples taken in each case on the
day of application, before any of the dissipation
processes could exert a significant effect (Table 2).
Description of Soil Sampling Techniques
In the early years of the program (Table 1,
Experiments 1 and 2), we used a relatively crude means
to collect soil samples. We arbitrarily divided the
test fields, which were fan-shaped, sloping watersheds,
into 5 or 8 subsections, depending on field size, and
sampled each subsection individually • Twelve holes
were dug randomly in each subsection and a slice of
soil was taken from the side of each hole with a
spadet soil from below the 17.5-c® depth was discarded.
The 12 slices were combined and the bulk sample was
reduced by coning and quarteringi the final analytical
sample (15-25 g) was obtained with use of a small
laboratory sample splitter.
The high variability of results led to a search
for an improved sampling method. In 19°9 (Experiments
3 and U>, we took 75 21-mm-diameter cores in a random
pattern from each of 5 subsections in the test field,
composited the cores by subsection, and obtained the
5 analytical samples by reducing the composites in a
sample splitter. Precision of results was, however,
not substantially Improved. We therefore used a
different technique In subsequent work (Experiments 5
through 8)j Sampling points in the test field were
laid out in a regular pattern. At each point,
a 30- x 30-cm steel box, open at top and bottom, was
placed so that it straddled the crop row (because the
pesticide in some of the experiments had been banded
in the seed furrow) and driven into the soil deep
enough to encompass within the box ell the pesticide
in the 900-cm2 area. The soil was then excavated from
the box by shovel, weighed at the site, and divided
is a Baaple splitter to obtain the analytical sample.
In experiments 7 and 8, the soil was also rolled on
a tarpaulin after weighing, to improve uniformity of
pesticide distribution.
The size of the fiexd used in the latest experi-
ments (Experiments 9 and 10) precluded the use of the
box technique because too much soli would have to be
handled. A modified sampling procedure was insti-
tuted: 16 closely spaced Hi-mm-diameter cores were
2
taken from about a 900-cm area at each of 25 sampling
points arranged across the field in a grid pattern.
The So cores from 5 randomly selected sites were then
composited in a 0.28-jii^ (10-ft^) rotary drum mixer,
mixed 5 minutes, and a laboratory subsample was
Obtained fror th* T»ixert soil wit.h unn of a ssmule
splitter; 5 of the composite sample*, represented the
entire field.
Intensive Analysis Experiment
In all the past work, variability among field soil
samples was high irrespective of the sampling method
used (Table 2). In addition to the inherent varia-
bility due to unevenness in field coverage during
pesticide application and irregular mixing on
incorporation of the chemical in the soil, variations
can also arise during the handling of the samples
after they are taken from the field. To ascertain
which of these was the dominant source of variation,
we conducted an intensive analytical program on the
day-of-application samples taken in 197't 'Experiments
9 and 10), Eight 50- to 100-g subsamples taken from
each laboratory composite by dividing the soil in a
sample splitter were analyzed individually.
The results show that variability among subsamples
was very low where the pesticide had been applied as
an emulsion, but was considerably higher with the
granular application (Table 3). An analysis of
variance revealed the followingt (1) In the emulsion
application, difference between composites was the
dominant source of variation, indicating that the
subsamples were well mixed. (2) In the granular
2
7-3
-------
application, difference between subsamples was the
dominant source of variation. Since there is no
reason to suspect that the chemical analysis of the
two formulations vas in any way different, this result
reflects the difficulty in mixing and subsarapling soil
that contains granules, which axe concentrated point
sources of pesticides. (3) Although the overall
means for the emulsion and granular applications
differed by a factor of only 1.3 (Table 3), the
variance amonp; composites was tenfold greater for the
granular than for the emulsion treatment. Since all
field samples were collected in the same way, this
indicates that the distribution of granules in the
field was much less uniform than that of the emulsion.
Table 3. Analytical results in 197*+ trifluralin experiments£/
Mean
Std. Dev.
Trifluralin Applied
mg/m*
as Emulsion
mg/m^
mg/m^
1
36.6
36.6
36.6
35.0
bO.O
35.0
36.6
36.6
3 6.6
1.5'
2
38.9
38.9
37.2
37.2
1+0.5
1+0.5
38.9
I+0.5
39.1
1.1+
3
33.6
-37.1
33.6
33.6
35.3
35.3
37.1
36.lt
35.3
1.5
U
35.0
38.6
35.0
36.8
35.0
36.8
35.0
36.8
36.1
1.3
5
1*5.6
UU.O
U7.1
UU.O
UU.O
UU.O
1+5.6
1+1+.0
UU.8
1.2
Composites
38.U
3.9
Trifluralin Applied
as Granules
1
97.5
85.3
61.0
6b.0
58.0
51+.8
50.2
106.6
72.2
21.3
2
U5.6
75.5
20.lt
1+7.2
22.1
31+.6
17.3
36.1
37.1+
19.1
3
21.7
M+.8
82.1
69.6
1+6.1+
63.1+
52.7
61.9
55.3
18.U
U
39.3
36.8
U9.9
39.8
33.5
66.5
1+9.3
1+1+.3
1+1+.9
10.U
5
51.7
1+1+.7
56.7
1+9.9
Uo.l
39.3
U8.lv
60.1
1+8.9
7.1+
Composites
51.7
13.2
6»f
— Experiments 9 and 10, Table 1.
Comparison of Soil Sampling Techniques
We can now examine differences in sample varia-
bility encountered over the 9-year research program
in light of the information obtained in the intensive
analysis experiment. In experiments 1, 2, 3, and
9 (Table 2), the pesticide was applied as a spray
emulsion; in experiments 5, 6, 7, 8, and 10, granules
were used.
The coefficients of variation with emulsionB show
that sampling technique Ho. U, used with trifluralin
in 197U (Experiment 9), was clearly superior to the
methods used earlier. Considering inherent varia-
bility in distribution of the pesticide in the field,
the 10$ coefficient of variation obtained with tech-
nique Ho. U probably represents the 'best that can be
done without taking an impractically large number of
samples. Results of an earlier detailed sampling
study showed that coefficients of variation for cores
composited from a single location decreased very
slowly below about 20% at sampling densities greater
than 1 core per 3 m2. Using technique Ho. U, we were
able to report a single figure that was adequately
representative of the pesticide content in a 3-hectare
field by analysis of only 5 samples.
Coefficients of variation in the experiments with
granules show that the box technique (Technique Ho. 3)
gave results as precise as did technique Ho. U, but
only when the samples were thoroughly mixed before
size reduction (Experiments 7 and 8). The box tech-
nique is also disadvantageous in that it is more
cumbersome; relatively large amounts of soil must be
moved manually during the sampling. Consequently,
technique Ho. U is preferable. The intensive analy-
sis experiment showed that the 25$ variation observed
in field experiment 10 could be largely attributed
to the inherent nature of granules, not to short-
comings in the sampling. Only by analyzing very
large numbers of samples or by increasing the
analytical sample size impractically beyond 100 grams
could a result be obtained with granules that is
substantially more representative of the pesticide
content of a field soil.
3
7-3
-------
The Recommended Soil Sampling, Technique
The essential ingredients of the optimum technique
in sampling field soils to determine pesticide content
are: (1) high-density core sampling at small-area
locations scattered over the field; (2) compositing of
samples from several random locations; (3) intensive
mixing of the composited soil; (U) splitting the
composite 3ample down to analytical size by means that
do not introduce a bias, such as by use of a sample
splitter; and (5) chemically analyzing enough samples
that a statistically sound estimate of variability can
be obtained. The analytical sample should be as large
as possible within the limitations imposed by the size
of the laboratory equipment. Sampling depth should be
sufficient to collect all the pesticide under the area
of soil surface being sampled. When this is done,
the soil pesticide content should preferably be
2
expressed on a surface-area basis (mg/m ) rather than
on a concentration basis (ppm), because the latter
varies with depth. Samples taken to the same depth
at different times of the year are not strictly com-
parable with respect to pesticide concentration
because, the soil changes in volume with changes in
temperature and moisture level. When measuring pesti-
cide persistence in the field, sampling and analysis
should be conducted periodically over a long enough
time that the quantification of the dissipation rate
is.statistically valid. This means measurements must
be made over essentially the entire residence time
7
of the pesticide in the soil .
Extraction of Pesticides
After obtaining a representative sample, the next
step in soil pesticide residue analysis is the
extraction of the pesticide from the soil. This has
been done by a variety of techniques—shaking or
tumbling, blending, Soxhlet refluxing, ultrasonic
vibration—using any one of numerous solvents or
solvent combinations, with the particular choice
depending mainly on the type of pesticide being ex-
tracted, By far the largest amount of developmental
effort has been expended on the organochlorine insecti-
cides, but successful means for extracting numerous
other classes of pesticides have also been reported.
In the following sections, we review published infor-
mation on pesticide extraction and present guidelines
for optimizing the procedure.
Extraction Technique
The efficiency of pesticide extraction depends
largely on the soil moisture content. Extreme drying
of the soil should be avoided during preextraction
handling and storage because of possible decomposition
or irreversible adsorption of the pestieide ; on the
other hand, recovery is improved by adding some water
just before extraction. In a comprehensive study with
dieldrin, Saha et alf obtained superior results by
air-drying the soil and adding 20% water to most soil
types (80# for mucks) a few minutes before extraction.
Addition of water also Improved the extraction of
chlordane, but additions of more than 30% reduced the
recovery10. In the official AOAC method for determin-
ing organochlorine insecticides in soil, the sample is
premolstened with 0.2 M NH^Cl rather than water-, the
salt causes the soil colloids to expand for better
U
contact with the solvent .
Soxhlet refluxing usually fares beBt in compari-
sons with shaking or blending, in terms of efficiency
12
of pesticide recovery . Although the Soxhlet method
is efficient and requires little attention of the
analyst during extraction, it is not without its
difficulties. £ts overall time consumption is rela-
tively great, it consistently yields more coextractives
than other methods, and it sometimes gives poor
reproducibility because the reflux rate is difficult
to control1"^. It also appears to be less efficient
than tumbling when used on heavy clay or muck soils
that tend to aggregate when vet11*. On balance,
however, the high efficiency of Soxhlet extraction is
generally considered important enough to offset its
disadvantages. If the time required for Soxhlet
extraction cannot be tolerated, a much faster ultra-
sonic vibration technique can be used. The most
effective procedure involves use of a high specific
intensity ultrasonic generator equipped with a cutting
head to powder the soil. Treatment of soil with this
equipment for only 30 seconds extracted organochlorine
insecticides efficiently1"'. The cutting head is
advantageous because reduction in soil particle size
to less than 100 mesh has been shown to improve
q
extraction efficiency .
Extraction Solvents
Since no one solvent or solvent mixture will ex-
tract every possible pesticide or metabolite, the
solvent should be selected carefully for a particular
pesticide or class of pesticides. For organochlorine
insecticides, the preferred extractants are mixtures
of polar and nonpolar solvents (hexane-lsopropanol,
hexane-acetone, benzene-methanol), although polar
solvents can be used alone. Polar solvents In general
give less coextractives than nonpolar ones, but most
of the mixtures also yield acceptably clean extracts1**.
An exception is 1:1 (v:v) chloroform-methanol, which
Extracts both pesticides and interfering compounds very
g
efficiently . Of the mixtures, hexane-acetone In
either 1:1 or 1»1:59 (v:v) azeotropic proportions is
the most widely accepted. The low boiling point of
the azeotrope (50°) lessens the possibility that
labile pesticides will degrade1"^. The official AOAC
method for organochlorine insecticides in soils
specifies 1:1 (v:v) hexane:acetone as solvent11,
whereas 3:1 (v:v) hexane-isopropanol is used in the
National Soils Monitoring Program for both organo-
chlorine and organophosphorus insecticides1®.
I«1 general, the organophosphorus insecticides are
extracted efficiently from soil by the same solvents
that perform well with the organochlorine compounds,
A variety of satisfactory solvents has been reported
for use with other classes of pesticides. Generally,
polar solvents (acetonltrlle, acetone, methanol) are
effective where polar metabolites are involved, such
aB with the carbamate insecticides. Methanol, either
neat or diluted, is an efficient solvent for many
herbicides, including the ureas, s-triazlnes, and
nitroanillnes. The aromatic acid herbicides (plclorara,
dicamba, fenac) require acidified polar organic
solvents for proper extraction. Intensive treatment
with a strong acid such as 18h HgSO^ is necessary to
extract the strongly adsorbed quaternary herbicides
paraquat and dlquat. Whatever the solvent selected
for a specific pesticide, it must be systematically
tested before incorporation into an analytical
method, using soli samples fortified with the pesticide
under conditions that simulate those In the field-.
k
7-3
-------
Radiolabeled pesticides are widely used in these
fortification tests.
The Recommended Extraction Procedure
To optimize the extraction of pesticide residues
from soil samples, the following points Bhould be
observed:
1. Just before extraction, the soil should be
air-dried (except for volatile pesticides), then
moistened with 20-30% (75-85?! for mucks) of a dilute
solution (0.2M) of ammonium chloride.
2. The solvent-soil ratio should be at least 2:1
(v:w), but preferably higher, for the extraction step.
3. Except for mucks and heavy clay soils, ex-
traction should be conducted in a Soxhlet apparatus.
To assure complete extraction, Soxhlet refluxing should
be continued overnight unless preliminary tests show
that shorter tiroes are adequate. Mucks and heavy clays
should be tumbled in the solvent and the tumbling
should be repeated with at least one fresh volume of
solvent.
H. Where time is short, ultrasonic vibration
combined with soil pulverization may be substituted
for Soxhlet extraction.
5. Unless the solvent is widely reported as being
efficient for the pesticide of interest, it should be
rigorously tested for efficiency of extraction, using
fortified soils that simulate field samples.
Acknowledgement
The authors are indebted to H. P. Freeman for
conducting the soil analyses in the intensive analysis
experiment,
Literature Cited
1. Caro, J. H., W. M. Edwards, B. L. Glass, and
M. H. Frere. Dieldrin in runoff from treated
watersheds. Proc. Symp. Interdiscipl. Aspects
of Watershed Mgrat., Bozesnan, MT., 1970. Amer.
Soc. Civil Engrs,, H.Y., 1972, p. ll»l-l6o.
2. Caro, J. H., H. P. Freeman, D. E. Glotfelty, B. C.
Turner, and W, M» Edwards. Dissipation of soil-
incorporated carbofuran in the field. J. Agr.
Food Chem. 21, 1010-1015, Nov./Dec., 1973.
3. Caro, J. H,, H. P. Freeman, and B. C. Turner.
Persistence in soil and losses in runoff of
soil-incorporated carbaryl in a small watershed.
J. Agr. Food Chem. 22, 860-863, Sept./Oct.,197^.
k. Caro, J. H., and A, W. Taylor. Pathways of loss
of dieldrin from soils under field conditions.
J. Agr. Food Chem. 19. 379-381*, Mar./Apr.,1971 •
5. Taylor, A. W., and H. L. Barrows. Sample errors
in measurements of the persistence of dieldrin
in a field soil. In Methods in residue analysis.
Vol. IV, Proc. 2nd Internat. IUPAC Congr. Pestic.
Chem., Tel Aviv. Gordon and Breach, Inc., N.Y.
1971, p. 1+39-1+56.
6. Taylor, A. W., H. P. Freeman, and W. M. Edwards.
Sample variability and the measurement of
dieldrin content of a soil in the field. J.
Agr. Food Chem. 19, 832-036, Sept./Oct., 1971.
7. Freeman, H. P., A. W. Taylor, and W. M. Edwards.
Heptachlor and dieldrin disappearance from a
field soil measured by annual residue determi-
nations. J. Agr. Food Chem., accepted for
publication, 1975.
8. Monitoring Panel, FWGPM. Guidelines on analyti-
cal methodology for pesticide residue monitor-
ing. Federal Working Group on Pest Mgmt.,
Washington, D.C., 1975.
9. Saha, J. G., B. Bhavaraju, Y. W. Lee, and R. L.
Randell. Factors affecting extraction of
llj.
dieldrin- C from soil. J. Agr. Food Chem.
17, 877-882, July/Aug., 1969.
10. Saha, J. G. Comparison of several methods for
extracting chlordane residues from the soil.
J. Assoc. Offic. Anal. Chem. 51*, 170-171*,
Jan., I971.
11. Woolson, E. A. Extraction of chlorinated hydro-
carbon insecticides from soil: collaborative
study. J. Assoc. Offic. Anal. Chem. 57,
60H-609, May, 197!*.
12. Woolson, E. A., and P. C. Kearney. Survey of
chlorinated insecticide residue analyses in
soils. J. Assoc. Offic. Anal. Chem. 52,
1202-1206, Nov., 1969.
13. Chiba, M., and H. V. Morley. Factors influencing
extraction of aldrin and dieldrin residues from
different soil types. J. Agr. Food Chem. 16,
916-922, Hov./Dec., 1968.
. Caro, J. H. Analytical considerations in
determining residues of organochlorine insecti-
cides in environmental samples. Jn Methods in
residue analysis. Vol. IV, Proc. 2nd Internat.
IUPAC Congr. Pestic. Chem., Tel Aviv. Gordon
and Breach, Inc., H.Y., 1971, p. ^57-475,
15. Johnsen, R. E., and R. I. Starr. Ultrarapid
extraction of insecticides from soil using a
new ultrasonic technique. J. Agr. Food Chem.
20, U8-51, Jan./Feb., 1972.
16. Pionke, H. B., and G. Chesters. Determination
of organochlorine insecticides in soils and
waters. Soil Sci. Soc. Amer. Proc. 32, 7U9-
759, Hov./Dec., 1966.
17. Williams, I. H. Bote on the effect of water on
Soxhlet extraction of some organochlorine
insecticides from soil and comparison of this
method with three others. J. Assoc. Offic.
Anal. Chem. 51, 715-717, May, 1968.
18. Crockett, A. B., G. B. Wiersma, H. Tai,
W. G. Mitchell, P. F. Sand, and A. E. Carey.
Pesticide residue levels in soils and crops,
FY-70—National Soils Monitoring Program (II).
Pestic. Monit. J. 8, 69-97, Sept., 1971*.
5
7-3
-------
CONFIRMATION OF PESTICIDE RESIDUE IDENTITY BY CHEMICAL DERIVATIZATION
William P. Cochrane
Laboratory Services Section
Plant Products Building
Agriculture Canada
Ottawa, Ontario
K1A 0C5
Summary
Microscale chemical derivatization provides a
quick, simple and inexpensive approach to the confirma-
tion of pesticide identity at the residue level. A
wide selection of confirmatory tests are available for
all classes of pesticides. General and specific tests
that our laboratories are currently using on a routine
basis within our normal pesticide residue analyses
procedures are discussed. These areas include the
organochlorine, organophosphate, carbamate, urea,
triazlne and chlorophenoxy acid pesticides.
The analysis of pesticides in agricultural or
environmental samples poses difficult problems for the
residue chemist. The extent of these difficulties can
be obtained from a survey of the many methods that have
been utilized for the initial identification of a
suspect "pesticide or metabolite (Table 1).
Table 1
Methods Used in the Identification of Pesticides
1. Instrumental Methods:
- Spectral techniques - Infrared, UV, visible,
mass, NMR.
- Paper and thin-layer chromatography.
- Gas-liquid chromatography - Multiple column
and detector systems.
- Neutron activation analysis.
2. Microchemlcal Methods:
- Chemical and photochemical conversion of
pesticides into derivatives.
3. Biological Assay Methods:
- Enzymatic
- Immunological
- Phytotoxiclty etc.
4. Partition Methods:
- Liquid/liquid - p-values.
- Liquid/solid - column chromatography
Also residue analyses are made at the parts per
million (ppm) or parts per billion (ppb) on such
substrates as animal tissues, foods, plants, soils,
water and air. Misidentifications can. result from con-
fusion with co-extracted pesticides, naturally occurr-
ing products as well as from extraneous contamination
and therefore many samples require rigorous clean-up
to prevent interference in the final anafysis. It is
well established that a peak on a gas-liquid chromato-
gram is not sufficient evidence to establish the
presence of an insecticide in an extract. Therefore,
it is necessary to obtain convincing confirmatory
evidence. A number of the more commonly used techniques
employed to provide confirmation of the identity of
pesticide components detected by gas-liquid chromato-
graphy (GC) are shown in Table 2.
The most frequently used method is by determination
of the retention times on 2 or 3 columns of different
polarity. At best this method is only adequate. The
use of element "specific" (or selective) detectors, such
as the alkali flame ionization detector (AFID), the
flame photometric detector (FPD) or the Coulson elec-
trolytic conductivity detector (CCD) will show the pres-
ence of certain hetero atoms but they are not specific
for any particular pesticide. The specificity of these
detectors is only relative.Moye^ described the, on-
column transesterification of N-methyl carbamate pesti-
cides to methyl-N-methylcarbamate and alsoJ the trans-
esterif ication of organophosphate pesticides to their
corresponding methyl esters. However, these techniques
are only characteristic of the N-methyl carbamate or
P moieties and are better suited as screening Yðods
than as confirmatory tests.
Mass spectrometry, especially in combination with
GC, is by far the best method for confirming the iden-
tity of pesticide residues. With an interactive data
system it is possible to obtain a complete mass spectrum
on 1 ng of material. Alternatively, specific ion moni-
toring, (also called mass fragmentography-') can be used
where up to 4 ions can be monitored rather than scann-
ing the entire spectrum. For example, it has been
possible to detect 10 pg methyl parathlon using the 263,
125, 109 and 79 peaks. To achieve this sensitivity a
high quality noise free power supply is required.
Therefore the use of mass spectrometry is often limited
by availability, cost of instrumentation and in some
instances, insufficient sample.
The third category shown on Table 2 - microscale
chemical reactions - can however, provide a simple,
quick and inexpensive alternative. Here the pesticides
or metabolites concerned are converted before GC injec-
tion to derivatives with different retention times from
the parent compounds and hopefully from other common
pesticides that may be present. Chemical derivatization
techniques have been extensively used for the confirma-
tion of organochlorine insecticides, organophosphates,
carbamates and herbicides>'»°• The remainder of this
talk will deal with general and specific techniques used
for each of these groups of pesticides and attempts are
made to indicate the most useful tests that are current-
ly employed on a routine basis within our normal pesti-
cide residue analyses procedures.
Organochlorine (OC) Insecticides
A wide selection oi confirmatory tests (Table 2)
suitable for OC pesticide residues is available.*>>9
Saponification, often used as a clean-up step for cer-
tain pesticides, has been by "far the most commonly used
method for both the Identification and, in many instan-
ces, quantitation of DDT and related compounds. Hexane
solutions of pesticides such as p,p'-DDT, p,p-DDD, per-
thane and p.p^methoxychlor (and in many instances their
respective o,p-isomers) treated with anhydrous sodium
methylate will dehydrochloririate to their respective
olefins.-*-" However, in the case of dicofol, p ,p-dichl-
orobenzophenone is produced. The effect of alcoholic
KOH on 46 miscellaneous pesticides was reported by
Krause.^-1 New products were formed from a large number
of the OC compounds investigated. This has resulted in
a very useful procedure since the retention times of
both parent and any derivative peaks were recorded on 3
different GC columns. Other reagents that can be used
L 7-4
-------
are potassium tert-butoxlde (t-BuOK), sodium ethylate,
alcoholic KOH and NaOH. The use o£ the stronger
nucleophilic reagents, e.g., t-BuOK can simultaneously
identify DDT, its analogues and some cyclodiene insecti-
cides of the chlordane series.^ Normally p,p-DDT
occurs together with its metabolite p,p-DDE which will
interfere with the confirmation of the parent compound.
Previous removal of DDE by thin-layer chromatography
(TLC) is advisable. This TLC step can be circumvented
by the use of chromous chloride (CrCl2) to dechlorinate
p,p^-DDT to p,p^-DDD, Although CrCl2 reacts preferenti-
ally with the p,p-isomer of DDT it will also dechlorin-
ate o,p-DDT but does not react with p,p-methoxychlor.
By far the best use of CrC^ is in the dechlorination
of the various members of the cyclodiene series. CrCl2
reacts with the gem-dlchloro group of dieldrin, aldrin,
chlordane etc. , to yield syn and antl mono-chloro
isomers.
Table 2
Commonly used Methods of
Confirmation of Pesticide Residues
1. Gas Chromatography:
- Retention times on different columns.
- On-column reactions.
- Element selective detectors.
2. "Mass Spectrometry (with or without GC link-up):
- Low resolution: complete spectrum on 1 ng.
- Specific ion monitoring: up to A peaks,
10 pg.
3. Microscale Chemical Reactions:
a) Organochlorine Insecticides
- Dehydrochlorination - basic reagents e.g.,
KOH, NaOMe, t-BuOK.
- Dechlorination - Chromous chloride, Zn/HCl,
- UV irradiation.
- oxidation/epoxidation - CrO^, m-chloroper-
benzoic acid etc.
- Halogenation - Cl^, , aqueous HBr etc.
- Others - rearrangements, sulphite reduction,
acetylation etc.
b) Organophosphorus Insecticides
- Alkaline hydrolysis and derlvatization of
the P moiety.
- Hydrolysis and derlvatization of the alkyl
or aryl moiety,
- Derlvatization of the intact insecticide
c) Herbicides
- Triazines - Alkylation, silylation, methoxy-
lation, NH derlvatization.
- Carbamates and Ureas - Alkylation and tri-
fluoroacetylation.
- Chlorophenoxyacids - transesterification.
The reduction of 23 cyclodiene insecticides and
metabolites or derivatives has been studied and all
yielded electron—capture (EC) responsive derivatives.-^
However, the major drawback, as with the UV irradiation
confirmatory procedures, is that each compound gives
rise to a definite derivative pattern consisting, in
most cases of 2 or more GC peaks. Therefore, in prac-
tice, no more than 2 pesticide residues should be con-
firmed at the one time. CrCl2 can also be used to de-
oxygenate unhindered epoxides such as dieldrin. De-
oxygenation occurs readily if a hydroxyl or keto group
is adjacent to the epoxide and at a slower rate if
hydrogen is present..
Oxidation/epoxidation reactions also have their
specific uses with respect to 0C compounds. The chromic
acid (CrO^) oxidation of p,p-DDE to p,p-dichloro-benzo-
2
phenone has been used both for quantitation and confir-
mation at the residue scale. However, this reaction is
also applicable to other DDT-related compounds. Epoxi-
dation of aldrin and heptachlor with Cr03 to their res-
pective epoxides constitutes another approach. The
most convenient method of epoxidation of aldrin to
dieldrin is via the use of commercially available m-
chloroperbenzoic acid.
For most of the OC compounds an adequate selection
of chemical confirmatory tests is available. However,
one limitation in the application of these tests has
been the lack of information on the relative sensitivi-
ty of the tests and the chemical nature of the deriva-
tives produced. Recently an evaluation study of the
commonly recommended confirmatory tests for dieldrin
in food and agricultural products was carried out.-^
The reactions were evaluated, using pure dieldrin and
spiked sample extracts of grass, animal feeds and
butterfat. Two reagents, namely aqueons HBr solution
and BCl„/2-chloroethanol were especially useful,
because of their application and sensitivity. A lower
level of 0.003 ppm for a 10 g dry sample and 0.001 ppm
for a 25 g wet sample was obtained. Similarly, the
use of base/alcohol reagents versus an alkaline GC pre-
column technique were compared for the confirmation of
BHC isomers.-^ It was found that the GC pre-column
technique saved time in the analysis of cereal, animal
feed and cheese extracts. However, the GC pre-column
did not eliminate interferences in some meat and fat
samples. Dehydrochlorination using NaOMe solution did
eliminate these interferences. A lower level of 0.01
ppm lindane (K-BHC) in a 10 g sample could be confirmed
by either confirmatory procedure. The alkaline GC pre-
column can also be used for the confirmation of DDT,
DDD etc., els and trans-chlordane and other pesticides
which can eliminate HC1.
Organophosphorus (OP) Insecticides
Three general procedures have been used for the
confirmation of OP insecticides (Table 2) and a review
of these procedures has appeared recently.9 Briefly
summarizing, no general chemical confirmatory test
exists for the OP's or for the carbamates or herbicides
either. By far the best approach is the generation of
specific tests for use in category 3. One such method
is the reduction of compounds containing arylnitro or
aryl cyano groups to their respective amines.The
use of chromous chloride, which has been used extensiv-
ely in the confirmation of OC insecticides, was found
to reduce parathion (Table 3), fenitrothion and EPN.
While other metal chlorides could be used e.g., PdCl2
it was found that Zn/HCl had the wider application
since it also reduced surecide, paraoxon and fenitro-
oxon. Reaction solvent was found to be very important.
For parathion, benzene gave only a single product while
acetone gave 2 products. This second product resulted
from the condensation of aminoparathion with acetone.
Using CrCl2, as little as 50 ng/1 parathion in water or
50 ug/kg in fish or sediment have been confirmed.
OP insecticides which possess a P=S group can be
converted to their corresponding oxons (P*0) by oxida-
tion (Table 3) with reagents such as neutralized sodium
hypochlorite solution.^- The main disadvantage of this
reaction is the lower response of the P=0 compared to
the P=S compound. However, diazinon, ronnel, malathion,
parathion, methyl parathion and fenitrothion were easily
confirmed in peaches, blueberries, apples and lettuce
at the 0.25-0.5 ppm level using a FPD in the P mode.
Recently a base-catalyzed alkylation procedure
(Table 3) has been investigated as a general reaction
for the confirmation of not only the OP's but also
carbamates and herbicides which possess an NH or NH2
moiety.-®-® A NaH/CH3l/DMS0 reaction mixture is preferr-
7-4
-------
ed for practical purposes using standard conditions of
50°C for 10 Minutes. Crufornate, dimethoate, Monitor
and BAY 93820 have been confirmed at the residue level
using AFID detection.
Table 3
Chemical Confirmatory Tests for
Organophorus Insecticides
1. (Et0)2-P-0- no2
Parathion
CrCl
Me
2. (Me0)2*P-0- no2
Fenitrothion
(Et0)2.p.0.<5)-NH2
Aminoparathion
0 Me
[0] (MeO)2-P-0- no2
Fenitro-oy.on
0
3.
'2 2
Dimethoate
0 CI
4. MeNH-P-O- t-Bu
OMe
(MeO)2- P-S- CH2* C-K-Me2
Crufomate
hv> MeNH*P»0* ^0 t^-Bu
OMe
des-chloro-crufornate
of hydroxyatrazine in soil using AFID detection. Meth-
ylation and alkylation were the reactions of choice.
Alkylation gave fewer extraneous GC peaks while
silylation loweired the detector sensitivity due to SiO,
build-up on the electrodes. The lower limit of sensi-
tivity for both atrazine and hydroxy atrazine using
alkylation was in the region of 0.5 pptn for a 5 g sample.
Lawrence^0 also compared 3 confirmatory techniques
(alkylation, methoxylation and NH derivatization) for
triazines herbicides in plants. Methoxylation involves
the replacement of the labile chlorine atom with OCH3.
Using the CCD in the N mode it was concluded the meth-
oxylation reaction (NaOMe for 5 min ) was the easiest
but since this test is only applicable to the chloro-
triazines the alkylation reaction was the most gener-
ally useful.
Cl
Cl-< 0 )-NH-C*N'(CH3).OCH3
0
Llnuron
Alkylation
50 /10 win
Cl CH.
0CH,
EtNH N NH iPr EtN N iPr
'3
•C-(CH,)«0CH,
« J J
0
OCH
EtNH
NH iPr
CH,
CH,
5. (Eto>2-P.o.(o>s.cH3
s
(EtO)2- P-0- (q). SCH2- TFA
Hydroxy-Atrazine
Atratone
Fensulfothion
An alternative confirmatory test for crufomate does
exist involving its UV irradiation in hexane to give
the deschloro derivative (Table 3). This alkylation
procedure has been compared with silylation as a means
of confirming two diazinon metabolites in dog urine at
the ppsi level using CCD detection. Silylation was
found to produce less GC background than alkylation.
Detector sensitivity of the silyl derivatives was also
greater. However, both reactions were considered
simple, rapid and made the analysis of one of the meta-
bolites possible where it was otherwise undetectable.
Herbicides
a) Triazines - Silylation and alkylation have
been used to confirm both the parent s-trlazine herbi-
cides and their hydroxy metabolites. Silylation of
hydroxy atrazine with BSFTA in a closed vial at 150°C
for 15 minutes yields a mixture of the mono-, di-and
tri-silyl derivatives
19
Atrazine itself forms a
single derivative in 90% yield when the^reaction is
carried out 190°C, silylation having occurred at the
NHEt position. At the 2 ng/ml level prometone, atra-
zine, atratone and prometryne are 95-98% alkylated by
NaH/CHjI/DMSO (Fig. 1). Alkylation occurs at both NH
groups. Hydroxyatrazine can be alkylated to a trimethyl
derivative identical to that obtained from atratone.
Hydroxy atrazine can also be converted to atratone iti
50-75% yield using a large excess of diazomethane.
These three methods, silylation, alkylation and methyl-
ation were examined and compared for the confirmation
I Methylation with diazomethane |
50-802 yield
Figure 1. Alkylation of Linuron, Hydroxy Atrazine
and
Atratone with Sodium Hydride/Methyl Iodide
h) Ureas and Carbamates - The most generally useful
confirmatory tests that are applicable to both the
intact compounds and hydrolysis products are the tri-
fluoroacetylation and alkylation techniques. Alkyla-
tion-1-8 of carbamates, such as methiocarb, propham,
terbutol and solan, require very mild reaction condi-
tions (5 min. at room temperature) to give single
derivatives in 83-97% yields. A lower level of 0.44
ug/ml was indicated since propham impurities with re-
tention times less than 4 min. were observed which may
interfere in the AFID detection of other carbamates.
Ureas can be alkylated under standard conditions to
give GC stable methyl derivatives. For linuron (Fig. 1)
feriuron, monuron and diuron, derivatization can be
achieved at the 0.5 ug/ml level. Monomethylation occurs
at the NH group in all cases. For alkylation the order
of reactivity appears to be
carbamates^ phosphorasiidates-aaides >
phosphoroamidothioates ^ triazines > ureas.
Trifluoroacetylation21, with trifluoroacetic
anhydride (TFAA) aa reagent, constitutes a very useful
confirmatory technique for various pesticides contain-
ing active NH groups. It has also been found^ that
sulfoxides are reduced and a trifluoroacetoxy (TFA)
group formed. This is shown In Table 4. While methio-
carb itself forms a mono-TFA derivative, methiocarb
sulfoxide (and its phenol analogue) formed a di-TFA
derivative in which reaction had occurred on the sulph-
7-4
-------
oxide as well as Che amide moiety. The reaction has
been applied to other pesticides such as fensulfothlon
(Table 3) and Its oxon,phorate sulphoxlde, terbufos
sulfoxide, oxydemeton-methyl and carboxin sulphoxlde.
Therefore the reaction of TFAA with the sulfoxide
moiety to give a TFA derivative appears to be general.
Table 4
Trifluoroacetylation Reactions
Me
.s.£o).o.
TFAA
1. MeS.^.O-CNH.Me looVs „ln >
Me
Me 0
Methlocarb MeS • \ ' Q'C'N'Me
Me TFA
0 Me 0
" ' B
2. CH,
i310-o
Me
•C-NH-Me IOC/SO mln
Me
Me
Methlocarb Sulfoxide TFA-CH^S0 ) • O-C-NMe
TFA
S 0
n ii
3. (EtO),* P*S*CH2" S-Et R.T./15 min
0
ii
Phorate Sulfoxide (EtO^- P-S-CHjS'Et
Phorate Oxon
0 0
ii n
4. (MeO)2* P-S'CH2CH2" S-Et 100*715 mln >
0
il
Oxydemeton-Methyl (MeO)^' P* S-CH»CH*S-Et
cis/trans
5.
a Me
...
R.T./15 min.,
Me
CTFA
NH >
Carboxin Sulfoxide
^S^\*NH 0
0-TFA
Depending on the substitution of this group, TFAA
will be eliminated with relative degrees of ease. The
reaction can be used at the nanogram level as a chemi-
cal confirmatory test for insecticides, herbicides and
fungicides which contain a NH or S-*0 group or both.
c) Chlorophenoxyaclda®>9 - Herbicides such as, 2,4-D,
2,4,5-T, mecoprop, MCPA etc., are routinely quantltated
by GC after conversion to their respective methyl
esters. Normally confirmation is achieved by transes-
terlflcation of these methyl derivatives to n-propyl or
n-butyl esters with n-Pr0H/H,,S0, or n-BuOH/BF, reagents.
An alternative approach is broaination with BiL/I, in
glacial acetic acid at room temperature for 10 min.
The methyl esters of MCPA, mecoprop and MCPB yield the
corresponding mono-2-bromo derivatives.
Conclusion
Microscale chemical tests provide a quick, simple
means of confirming the identity of a suspect pesticide
residue. Such techniques as alkylation and trifluor-
acetylation can be generally applied to OP, carbamates,
ureas and various herbicides which possess active hydro-
gens. Also in this work it was found important to sel-
ect chemical confirmatory tests compatible with the
type of equipment available. For example, the relative
sensitivity of parathion and aminoparathion using an
electron capture detector is approximately 400:1. Hence
this test is best applied using a N or P selective det-
ector. Finally, chemical confirmatory tests are easily
incorporated into routine residue analysis where access
to sophisticated equipment is limited.
1. Natusch, D.F.S. and Thorpe, T.M. Anal. Chem. 45,
1184A (1973)
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
Moye, H.A. J. Agr. Food Chem., 1^., 452 (1971)
Moye, H.A. J. Agr. Food Chem., 21, 621 (1973)
Biros, F.J. Advan. Chem. Ser., 104, 132 (1971)
Rosen, J.D. and Pareles, S.R. "Mass Spectrometry
and NMR Spectroscopy in Pesticide Chemistry"
R. Haque and F.J. Biros, eds., Plenum Press, New
York, 1974, pp 91-98
Cochrane W.P. and Chau A.S.Y., Advan. Chem. Ser.
104, 11 (1971)
Elgar, K.E. Advan. Chem. Ser., 104, 151 (1971)
Cochrane, W.P. and Purkayastha R., Toxicol. Environ.
Chem. Rev. 1, 137 (1973)
Cochrane, W.P. J. Chromatogr. Sci., 13^, 246 (1975)
Mendoza, C.E., Wales, P.J., McLeod, H.A. and
McKinley, W.P., J. Assoc. Offic. Anal. Chem.,
51, 1095 (1968)
Krause, R.T., J. Assoc. Offic. Anal. Chem. 55,
1042 (1972)
Chau, A.S.Y. and Cochrane, W.P., J. Assoc. Offic.
Anal. Chem., 52, 1092 (1969)
Cochrane, W.P. and Forbes, M.A. "Methods in Residue
Analysis" Vol XV of Pesticide Chemistry.
A.S. Tahori, ed. Gordon and Breach, London, 1971,
pp 385-402
Maybury, R.B., and Cochrane, W.P., J. Assoc. Offic.
Anal. Chem., 56, 36 (1973)
Cochrane, W.P. and Maybury, R.B., J. Assoc. Offic.
Anal. Chem., 56, 1324 (1973)
Forbes, M.A., Wilson, B.P. , Greenhalgh, R., and
Cochrane, W.P., Bull. Environ. Contam. Toxicol
13, 141 (1975)
Singh, J. and Lapointe, M.R., J. Assoc. Offic. Anal.
Chem., 57, 1285 (1974)
Greenhalgh, R. and Kovacicova, J., J. Agr. Food
Chem., 23, 325 (1975)
Khan, S.U., Greenhalgh, R. and Cochrane, W.P.,
J. Agr. Food Chem., 23, (1975)
Lawrence, J.F., J. Agr. Food Chem., 22^, 936 (1974)
Greenhalgh, R., Marshall, W.D. and King, R.R.
J. Agr. Food Chem. (in press)
22. Greenhalgh, R. (unpublished results, 1975)
4
7-4
-------
APPLICATION OF ANALYTICAL METHODS RESEARCH
TO MONITORING ORGANIC RESIDUES IN FISH
David L. Stalling
Fish-Pesticide Research Laboratory
fish and Wildlife Service '
U.S. Departipent of Interior
Columbia, Missouri 65201
Summary
Numerous problems were encountered analyzing residues
In fish collected yearly from 100 sampling stations es-
tablished In conjunction with the National Pesticide
Monitoring Program. Conventional cleanup of extracts
by Florisil chromatography or solvent partitioning did
not permit GC-MS confirmation due to incomplete lipid
removal. Development of a more efficient cleanup meth-
od was achieved using a styrene-dlvlnylbenzene copolymer
gel permeation chromatography (GPC) column which result-
ed In >98% lipid removal without altering the residues.
Additional separations using either silicic acid or sil-
ica gel columns were required to resolve residues of
toxaphene and «hlordane in the presence of PCBs. Chem-
ical Ionization GC-MS using selective ion monitoring
permitted more selective toxaphene analysis. Addition-
al research is being directed toward development of au-
tomated cleanup Methods for phenolic moieties derived
from noii-persistent pesticides using LH-20 to separate
phenols from free fatty acids, lipids, PCBs and pesti-
cides. Also, combination of sequential charcoal adsorp-
tion columns with GPC 1b being explored to achieve se-
lective separation of residues including chlorinated
dibenzo-p-dioxins and other planar aromatics.
~ Moaccumulation ot persistent erivifotttaeHEftl eoniam-
inants by fish is well established. Residues consist-
ing of numerous pesticides, polychlorinated-biphenyls
(PCBs) other halogenated aromatic hydrocarbons and phe-
nols, and phthalate esters are compounds typically
found in fish from freshwater environments. Problems
encountered in analyzing these complex residues include
inadequate removal of coextracted lipids, formation of
stable emulsions, variability in procedures used for
fractionation of the residues into subgroups, contamina-
tion of samples with laboratory chemicals and masking
interference by multicomponent compounds.
Since 1968 part of our laboratory's research pro-
gram involved performing crosscheck analysis of fish
samples collected by the National Pesticide Monitoring
Program • In order to ensure that the analyses
we conducted were reasonably comprehensive and accurate,
we undertook methods research to develop improved proce-
dures for cleanup of fish extracts and separation of
residues into subgroups.
Techniques which could be automated were the sub-
ject of our initial Investigations since monitoring
Programs typically generate large numbers of samples
and manpower is limited. Another benefit we sought was
improved precision. We selected low pressure gel per-
meation chromatography (GPC) as our basic llpid-pesti-
cide separation tool3. This technique appeared wall
suited for automation and was a likely candidate to re-
place liquid/liquid partitioning for lipid removal.
GPC Cleanup <
GPC offers a major advantage over adsorption chromatog-
raphy in that the column material 1) does not require
activation; 2) can be used repetitively, and 3) does
not require solvent regeneration or coagslex gradients.
Freeman summarized the mechanism responsible for GPC
separations and Mudler and Buytenhuya discussed aspecta
of preparative separation ot small molecules by GPC and
listed the gels commonly employed and their manufactur-
er.
BioBeads Separations.
Tindle and Stalling3 described a chromatographic clean-
up system that provided for unattended sample introduc-
tion and GPC fractionation of up to 23 crude sample ex-
tracts. .Eolystyreae-divinyl benzene copolymer resins
(BioBeadsfij and cyclohexane permitted efficient separa-
tion of lipids and many environmental contaminants.
GPC cleanup removed 98 percent of coextracted fish
and oil lipids and quantitatively recovered PCBs and or-
ganochlorine insecticides. Two problems were encounter-
ed in the cyclohexane based GPC system: 1) in green
plant materials,lipids ware not readily eluted and 2)
animal feed extracts were not cleaned up satisfactorily.
An automated GPC cleanup system for pesticides became
commercially available in 1974®.
. Research on gel resins and solvent systems for
cleanup of pesticides and environmental contaminants has
resulted in improved GPC cleanups. Satisfactory pesti-
cide cleanup using the automated system was made on
plant and animal oil extracts including cleanup of soap
stocks®. Cleanup was effected for pesticides, PCBs,
chlorinated-dibenzo-p-dioxina (Cl-DBDs),-dibenzofurans
(CI-DBFs),-naphthalenes (Cl-NAPs),-dlphenylesters (Cl-
DPE) and -benzenes. A 2.5 cm id column of BioBeads S-
X3 was effective in removing up to 1500 mg of extracted
lipid while a 1.0 cm id column was found to have a ca-r
paclty of 500 mg of lipid. Elutlon volumes of planar
aromatic compounds increase with the degree of chlorlna-
tion and may be due to solubility and/or electronic in-
teraction with the polymer lattice network. Multicom-
ponent chemical mixtures such as the Aroelora) series
1242, 1254, and 1260 were quantitatively recovered with
98% or more of the lipid removed. Individual components
100'
90-
80-
70—
SO—
h-
z
Ul
40—
30—
20——J
exuT-i'ON volume tnu
Figure 1. Station profiles ot selected compound# demon-
strating separation capabilities of 25% tol-
uene-ethyl acetate aa a solvent. GPC column
was 1.0 cm Id x 56 cm containing BioBeads
S-X3.
S-X3 UITH 25* TOLIINE ET0RC
oDEHP
• DOE
~BtT0-BHC
aUPID
-------
In these mixtures tended to elute at different rates,
however, the total elution volume remained narrow. Cl-
DBFs had elution patterns similar to the PCBs. Cl-DPE
eluted between the PCBs and the later eluting Cl-NAP
components and hexachlorobenzene (CgClg or HCB) was a
late eluting compound.
Carbamates and ionizable compounds such as phenol,
aniline, and 2,4-dlchlorophenoxy acetic acid (2,4-D)
were only partially separated from lipid using BioBeads.
Phenol and aniline had identical elution profiles and
methylating these polar functional groups did not ap-
pear to Influence the elution. Methylating 2,4-D in-
creased the elution volume by several ml so that the
methyl ester was well resolved from lipid in contrast
to the free acid. An improved solvent system, 25 per-
cent toluene-ethyl acetate, greatly increased separation
of di-2-ethylhexyl phthalate (DEHP), phenol, carbamates,
and benzene hexachloride (BHC) isomers from lipid. This
mixed solvent separated 98 percent of the lipid compo-
nents from DEHP. Ethyl acetate separated only 50 per-
cent of the lipid from DEHP. The resolution of beta-
BHC was also improved (Fig. 1). Elution parameters of
various chemicals were also studied in different systems
with the 1.0 cm id column (Table 1).
GPC Separation with l.H-20.
LH-20 differs from polystyrene-divinylbenzene copolymers
in that LH-20 has no aromatic structural elements. Al-
though LH-20 has been employed for pesticide cleanup,
BioBeads have provided us with superior separations for
a majority of the neutral compounds studied. LH-20,
however, appears to be well suited for cleanup of ex-
tracts containing phenols, particularly in the presence
of long chain fatty acids.
We have encountered problems with emulsions when
attempting to separate phenolic materials from visceral
extracts of fish by partitioning with base. The pres-
ence of large amounts of free, long chain fatty acids in
viscera was confirmed by GC-MS following methylation of
the extracts with H2S04:CH30H. We have also found large
concentrations of methyl esters of fatty acids in bird
extracts which are recovered along with DDT using Flor-
isil cleanup. Thus, a procedure for separating phenols
from fatty acids would avoid the formation of soaps and
eliminate emulsion formation. Separation of various
phenols from triglycerides, fatty acids and their methyl
esters was achieved with LH-20 (Table 2). Recovery of
phenols and phenylacetate was quantitative. Increasing
the content of methanol in the solvent system decreased
the elution volume of triglycerides or fatty acids.
Base partition of the GPC phenol fraction before deriva-
tization provides additional separation of small neutral
contaminants from phenols.
Table 1. Solvent effects on GPC elution volumes!/
Elution Volume Range (ml)
Table 2.
Solvent
volumes.
COT1P0UND
ETHYL
2S?*0lu£N£
ACETATE
ACETATE.
-ETCAC
scu_noN OiL
20-3F,
4 2-30
SOYBEAN OIL
1G-35
-
TCDD
OtOD
NCR
ALPHA- 8HQ
PE.Tfl~BHC
connp-pHc
0ELTA-BHC
ALDRIN
HCPTflCHLOff
CJS-CWLOftOftNE.
P.P-00C
? ,? ODD
F .P -DOT
HALOWAX
34-50
40-50
32-42
-12-5-1
44-&J?
-
42-51
32-42
2€-K
32-42
2C- 3C
36-46
34-44
32-42
44-«>6
3^-45
34-42
3G-4S
J*-46
3C-4?
33-44
32-44
36-46
26-38
30-3C
3G-4C
32-40
32-40
30-40
2C-3^
34-45
3
-------
After completion of the GPC cleanup cycle, elution of
the adsorption column with desired solvent la accom-
plished by directing the solvent selection valve into
the adsorption column. This eluate is collected in one
of the second flasks (Eluate 2). Addition of an inde-
pendent rotary valve controller to the second set of ef-
fluent distributors would permit unattended fraction-
ation of each sample into a maximum of 23 fractions us-
ing different solvents. Fractionation on the basis of
constant flow is under software program control using a
programmable clock. Valve control is effected by a dig-
ital input/output control module (up to 256 lines) of
the microprocessor.
PULSE DAMPER
MICROPROCESSOR
SOlVENT| |
RESERUOIR
(W) I 1
SAMPLE STORAGE
•aOOPS(5ml)
24 POSITIONS
SAMPLE
INTRODUCTION
"'AIVE(SIV)
-BYPASS
GPC
COLUMN
-vRQIWW VAIVE
I .CONTROLLER
DRAIN
OLLECT/DUMP
EFFLUENT
DISTRIBUTOR 1 K
RESERVOIR
ADSORPTION
COLUMN
2
S REQUIRED)
SOLVENT -
|electkical -
ELUATE 1
ELUATE 2
Figure 2. GPC Sequential adsorption system.
Development of a continuous flow effluent concentra-
tor. would permit automation of much of the present meth-
odology based on Florisil cleanup. We have employed
minicolumns containing Darco charcoal combined sequen-
tially with GPC eluate to retain some of the above groups
of chemicals. Earlier, McLoedet a1.9 compared several
carbon adsorbents for the separation of 42 pesticides.
Substitution of other charcoal or adsorption materials
for Darco charcoal will also permit modification of the
fractionation capabilities of the present system.
Fractionation of Environmental Contaminants.
Following GPC cleanup of fish extracts, adsorption chro-
matography is employed to reduce the complexity of the
residues in order to permit more reliable Interpretation
of GLC chromatograms. Florisil columns, are vised to sep-
arate dieldrin, endrin and phthalate esters from FCBs
and other less strongly adsorbed pesticides which are
removed by elution with 6-10% diethyl ether in petrole-
um ether. Further fractionation of PCBs and organochlo-
rine pesticides is currently achieved utilizing silicic
acid or a miniature silica gel column (Underwood, J.C.,
FDA, Kansas City, Missouri, Personal communication 1975).
The method employs 4 g of Woelm silica gel (Act I, heat-
ed overnight at 130 °C) in small 1 cm i.d. glass columns.
Solvents for the elution of PCBs and pesticides are 43
ml of 0.5% benzene in hexane (v/v) and 15 ml of 42 ethyl
acetate in benzene (v/v), respectively. PCB-pesticide
separations for previously cleaned up fish extracts (<5
g tissue) appear similar to separations by purified
silici;acid chromatography. In addition, no significant
contaminants interierring, with PCB-pesticide separations
.have been detected in batches of silicic acid. When an-
alyzing fish samples containing PCBs and pesticides with
silicic acid, we find that most PCB components contain-
ing 2 and 3 dhlorine atoms are removed in the PCB frac-
tion. Thus the majority of Aroclor 1242 components, all
components of PCB mixtures more chlorinated than Aroclor
1242, and most of the ]>.,£'-DDE elute In the PCB fraction.
Other chemicals found in the PCB fraction include mirex,
HCB, aldrin, heptachlor (a component of chlordane) and
several of the earliest eluting GC components of tech-
nical chlordane.
Our experience with analytical cleanup methods suit-
able for preparing extracts of fish samples from the
National Pesticide Monitoring Program indicates that
silicic acid chromatography provides adequate PCB-pesti-
cide separations. However, consistant PCB-pesticide
separations with silicic acid can. only be achieved by
characterization of each prepared batch of adsorbent
and by elimination of interfering contaminants. The
additional time required to purify contaminated batches
of silicic acid limits the utility of the method to
situations requiring a large capacity separation system.
The miniature columns of silica gel offer an acceptable
alternative to silicic acid chromatography for many PCB-
pesticide separations and require less solvents and
analytical time.
Micro Cel-E.
Velth et aT!-0 described use of a relatively new adsorb-
ent, Micro Cel-E, for precleanup of large amounts of
lipids (5-30 g prior to GPC cleanup). We are currently
examining Micro Cel-E for use as a desicant-grinding
aid and adsorbent for separation of environmental resi-
dues. Phthalic acid, it's monoesters and 2,4-D are
selectively retained and separated from lipids, free
fatty acids, phenols, PCBs and non-polar pesticides.
Tissue is ground with 10 g Micro Cel-E and transferred
to a 2.2 cm id glass column. The lipids, fatty acids
and neutral contaminants are eluted with 200 ml of 5%
diethyl ether-petroleum ether. Then 2,4-D (95% recov-
ery) and most of the phthalate monoesters (u80%) are
eluted with 225 ml of 10% acetone:acetonitrlie. Phthal-
ic acid and remaining monoesters are recovered with 200
ml of 2% H3P04:2% toluene in methanol. Reduction of
lipid on a weight basis was approximately 80%. Thus,
Micro Cel-E with column extraction performs drying and
effective cleanup for these polar aromatic acids.
Residues in fish analyzed from the National Pesticide
Monitoring Program as crosscheck samples (1970-1973).
Sampling location, species selection and general as-
pects of the National Pesticide Monitoring Program
(NPMP) were descrived by Henderson et al. (1). Fish
are collected annually from each of the fifty states
for analysis of pesticide and metal residues. Currently,
there are 100 sampling stations. Fish collection and
administration of the contract for commercial analyses
were the responsibility of the Division of Technical
Services (formerly Fishery Services), U.S. Fish and
Wildlife Service, U.S. Department of the Interior, Wash-
ington, D.C. Analytical methodology employed in the
program in the early years was not adequate to remove
interferences caused by PCBs. Detection of chlordane
in the complex of environmental contaminants found in
fish from the environment requires special attention
and use of efficient separation methods to remove PCBs
and related interferences. We selected and analysed
for organochlorine residues approximately 10 percent of
samples collected each year. Residues in fish collect-
ed since 1968 have not been published and data are be-
yond the scope of this discussion. Therefore, emphasis
will be placed on selected residues at specific stations.
There are areas in which excessively high environ-
mental residues are found. We have summarized some of
the problem areas in which exceptionally high residue
levels of chlordane, dieldrin, toxaphene, DDT and PCBs •
occur (Table 3). Whole body PCB residues of 50-100 Aig/g
7-5
-------
occur In the Hudson and Ohio Rivers. Decreases in the
levels since 1970 at other stations are suggested, par-
ticularly in the Merriraac River and at Marietta, Ohio.
Toxaphene residues in the southeast, especially in
the Yazoo River, Redwood, Mississippi are excessively
high indicating a rather serious pollution problem. DDT
residues also are high at this station. Additional sam-
pling here could indicate if agriculture practices and/
or effluents from formulation operations are responsible
for the residue input. Chlordane and dieldrin residues
are also a problem in Manoa Stream, Honolulu, Hawaii.
Dashed entries in the tables indicate that the residue
was below the level of detection or was obscured by
masking interferences.
GC-MS confirmation of chlordane and toxaphena residues.
In order to verify proper identification of pesticides
in samples analyzed by GC we employed a Perkin Elmer
Model 270 GC-MS interfaced to a PDP-12 LDP computer (Dig-
ital Equipment Corporation). The computer system per-
mits the intensity of an ion fragment or molecular ion
to be displayed in a manner similar to a GC trace. This
capability permits ratios of coeltlting GC components to
be determined. Chlordane residues were confirmed by GC-
MS by locating spectra having an ion cluster at M/e-371,
the ions corresponding to the molecular weight of cls-
chlordane minus 35 CI. Contaminated fish from the fol-
lowing rivers or lakes were confirmed to contain cis-
chlordane: Lake Michigan; Raritan River, Bound Brook,
New"Jersey; Manoa Stream, Honolulu, Hawaii; Mississippi
River, Luling, Mississippi; Arkansas River, Pine Bluff,
Arkansas and Ohio River, Marietta, Ohio.
Gas chroraatography-chemical ionization-mass spectro-
metry (CI-MS) was essential in the characterization of
toxaphene in fish. Only a few of the more than 150 com-
ponents which make up toxaphene are adequately separated
by GC. CI-MS permits more specific detection and char-
acterization of these residues than electron impact-MS.
Selective ion monitor curves from the mass spectra of
toxaphene residues found in fish from the southeastern
U.S. and from laboratory-exposed fish were similar to
each other and to toxaphene standards.
Ratios of cis-chlordane to trans-nonachlor in fish.
Recently we undertook measurement of ratios of cis-chlor-
dane/trans-nonachlor in a limited number of samples us-
ing GC-MS. Ratios of these two coeluting GC components
were determined from the ratio of the maximum ion in-
tensity of M/e 373 and 405. The ratio of these two mass
fragments measured in technical chlordane was approxi-
mately 4.5 to 1. This ratio was 5.1 in the tilapia
from Hawaii 6.1 in white perch from the Raritan River,
New Jersey; 3.9 in striped bass eggs from South Carolina;
and 2.0 in yellow perch from Lake Michigan. From these
data we can conclude that both cis-chlordane and trans-
nonachlor persist in fish.
Table 3. Environmental contaminants in fish.
Residue pug/g whole body)—^
1970
1971
1972
1973
PCB's
Hudson R., Poughke-
epsi, NY
93
61
104
213
76
118
69
Ohio R.
Marietta, OH
122
77
16
7.5
38
-
6.0
Cincinnati, OH
133
156
20
45
66
-
43
Metropolis, IL
8.6
Lake Michigan
Sheboygan, W1
7.8
2.6
11
53
4.9
4.4
Merrimac R.
Lowell, MA
98
45
8.8
8.5
Table 3. Cont.
1970
1971
1972
1973
Toxaphene
Yazoo R.
Redwood, MS
34
-
24
300
. 31
48
40
20
Tombigbee R.
Mcintosh, AL
-
0.50
0.50
0.50
Mississippi R.
Luling, LA
-
4.0
1.0
2.0
2.5
Memphis, TN
-
2.0
0.50
Arkansas R.
Pine Bluff, AR
-
4.0
5.0
3.3
Lake Michigan
Sheboygan, WI
-
-
-
-
DDT
Yazoo R.
Redwood, MS
20
16
37
31
26
22
11
Tombigbee R.
Mcintosh, AL
7.9
6.1
26
12
Mississippi R.
Luling, LA
0.80
1.3
0.90
0.45
1.0
Memphis, TN
6.8
-
0.70
0.26
Arkansas R.
Pine Bluff, AR
10
22
3.5
3.0
Lake Michigan
0 A
Sheboygan, WI
5.2
z, o
1.7
23
3.2
8.4
Dieldrin
Manoa Stream
Honolulu, HI
3.8
0.94
18
Raritan R.
Bound Brook, tU
0.50
0.31
0.07
Lake Michigan
Sheboygan, WI
0.30
-
0.45
0.60
0.12
0.01
Chlordane
Manoa Stream
Honolulu, HI
0.30
2.8
24
Raritan R.
Bound Brook, NJ
0.32
-
3.3
Lake Michigan
Sheboygan, WI
0.25
-
-
-
0.18
-
1/Each value represents analysis of single fish.
Significant advances in cleanup of environmental
contaminants are afforded by GPC. Development of LH-20
should prcJvide a method for phenolic moieties derived
from non-persistent pesticides.
References
1. Henderson, C. et al., Pesticide Monitoring Journal
3,145 (1969).
2. Stalling, B.L., et al., J.A.O.A.C. 55(1):32 (1972).
3. Freeman, D.H., J. Chromatogr. Sci. 11,175 (1973).
4. Mudler, J.L., and Buytetlhuys, F.A., J. Chromatogr.
51,459 (1970).
5. Tindle, R.C. and Stalling, D.L., Anal. Chem. 44,
1768 (1972).
6. GPC 1001-Autoprep, Analytical Biochemistry Labora-
tories Inc., Columbia, M0 65201.
7. Stalling, D.L., al., Proceeding of 3rd Interna-
tional Congress on Pesticide Chemistry, Helsinki,
Finland, July 1974.
8. Johnson, L.D., et al,, J.A.O.A.C. (in press).
9. McLoed, H.A., at al., J.A.O.A.C. 50,1216 (1967).
10. Veith, G.D., et al., J.A.O.A.C. 58,1 (1975).
4
7-5
-------
DETERMINATION OF PCBs AND MIREX IN L, ONTARIO
HERRING GULLS
D.J. Hallett, R.J. Norstrom, CWS, Ottawa, Ontario,
F.I. Onuska, M. Comba, R. Sampson, CCIW, Burlington,
Environment Canada.
Abstract
Mirex (dodecachloroctahydro-1,2,3-metheno-2
H-cyclobut-(c,d)-pentalene), a monodechlorinated
metabolite, and polychlorinated biphenyls (PCB's),
were determined in a whole body lipid of Herring
Gulls from Pigeon Island, L. Ontario, Canada. PCB's
coelute with mirex in residue cleanup procedures and
peaks are superimposed under standard gas chromato-
graphic conditions. Samples were perchlorinated to
remove this interference. Mirex a metabolite, and
PCB's were identified by GC/MS and quantitated by GC.
Herring Gulls have shown severe reproductive failures
on L. Ontario and appear to be a useful indicator of
persistent environmental pollutants in the Great
Lakes.
The adult, breeding Herring Gull is a relatively
non-migratory resident of Lake Ontario^- and feeds
largely on small fish. Severe reproductive failures
have occurred with Lake Ontario colonies of Herring
GullsT Eggs from these populations showed severe
contamination with DDE and PCB's which was correlated
with eggshell thinning2. The species provided an
excellent indicator organism of trace organochlorine
pollutants in Lake Ontario. Direct analysis of lake
water and lower biota for trace contaminants is
difficult due to the low levels present. The whole
body lipid of 16 mature birds from Pigeon Island,
L. Ontario in which the residues are preconcentrated
was therefore analysed.
Kaiser recently reported the unexpected occur-
rence of Mirex in fish found in Lake Ontario. High
levels of PCB's and DDE interfere with routine
monitoring of trace amounts of other organochlorine
residues and metabolites, particularly Mirex, using
the present methods^>5. Therefore samples were
extracted, cleaned up, and fractionated using normal
routine procedures, and then perchlorinated to
remove any PCB interference®. Residues were then
quantitated by Gas Chromatography using a standard EC
detector. Samples were also analysed with a Hall
detector in the chloride mode prior to perchlorina-
tion. The Hall detector has a low specificity to PCB
interference and a high specificity to perchlorinated
compounds such as Mirex?. The identity of the
molecules detected in the perchlorinated sample was
confirmed by GC/MS.
The adult Herring Gull lipid was extracted by the
method of Porter and Burke®. Residues were separated
on a deactivated Florisil column. The column was
eluted first with hexane to yield a fraction contain-
ing PCBs, HCB, DDE, and Mirex (recovery 99+0.5%)
followed by 30% methylene chloride in hexane to yield
more polar organochlorine pesticide residues such as
dieldrin.
An EC chromatogram obtained on a 1850 x 2 mm ID
glass column containing 3% 00-1 (80-100 mesh) on
Chromsorb W of the hexane fraction showed major peaks
attributable to PCB isomers and DDE. Evidence of
other organochlorine residues was totally masked by
PCB interference. A large peak with a mass spectrum
and retention time corresponding to heptachlorobiphenyl
(found in Arochlor 1260) was superimposed on Mirex.
The same hexanefraction was then analysed under
identical gas chromatographic conditions using the
Hall conductivity detector in the chloride mode. This
yielded 3 distinct major peaks attributable to DDE,
Mirex, and an unknown compound (Relative Retention
time to Mirex was 0.70). Due to its long retention
time and specific response the unknown was suspected
to be a derivative of Mirex. The relative response
of Mirex was found to be 375 times greater on the
Hall detector than the response of the heptachloro-
biphenyl peak in Arochlor 1260 which totally masked
Mirex with the EC chromatogram.
A portion of the same hexane eluant obtained
from the Herring Gull lipid was perchlorinated. An
FID chromatogram of the perchlorinated fraction
showed 4 major peaks. The structures were identified
and confirmed by GC/MS to be hexachlorobenzene, a
monodechlorinated degradation product of Mirex with a
molecular composition of CioHCln, Mirex Ciq Cll2>
decachlorobiphenyl, These residues were quantitated
by EC and gave 3530 ppm PCBs as Arochlor 1260, 220 ppm
Mirex, and 84 ppm of monodechlorinated Mirex.
Quantitation of the nonperchlorinated fraction with
the Hall detector gave 330 ppm DDE, 210 ppm Mirex,
and 70 ppm of the Mirex derivative.
Analysis of extracts with the Hall detector
provided a useful method for routinely monitoring
residues of Mirex and DDE in the Herring Gull lipid,
and led to the detection of monodechlorinated Mirex
derivative. Removal of high PCB interference by
column chromatography was not necessary. Perchlorina-
tion of extracts was an excellent means of preparing
samples for GC/MS. In our hands the Hall detector in
the reductive chloride mode was capable of detecting
residues of organochlorine pesticides in the 100
program range. It was extremely specific for these
chlorine compounds in the presence of PCB interference
and would prove useful as a confirmatory procedure
when identifying trace contaminants.
The whole body lipid obtained from adult Herring
Gulls indicated that relatively high concentrations of
Mirex and a monodechlorinated degradation product of
Mirex along with PCBs and DDE are bioaccumulated in
the food web. These residues were most likely
absorbed from a diet of lake fish and are ultimately
the result of biomagnification of the organochlorine
compounds from the Lake Ontario water through the
food web.
References
1. Kaldec, J.A., W.H. Drury, Ecology 49, 644 (1968).
2. Gilbertson, M., Can. Field-Naturalist 88, 273
(1974). ~
3. Kaiser, K.L.E., Science JL85, 5?3 (1974).
4. Reynolds, L., Residue Reviews _34, 27 (1971).
5. Berg, D.W., P.L. Dios»ay, G.A.V. Rees, Bull.
Environ. Contam. Toxicol., 2, 338 (1972).
6. Armour, J.A., J. Ass. Offic. Anal. Chem. 56, 987
(1973). ~~
7. Hall, R.C., J. Chromatogr. Sci. 12, 152 (1974).
8. Porter, M.L., J.A. Burke, J. Ass. Offic. Anal.
Chem. 56, 733 (1973).
1
7-6
-------
PESTICtSSES IH DEVELOPING COUNTRIES--
SIGN(FICA.NCE OF CHLORINATED HYDROCARBON RESIDUES IK
HUMAN MILK FROM CENTRAL AMERICA
M, TAGHI FARVAR, Ph.D.
Director
Centre for Endogenous Development Studies
Bo* 938, Tehran, Iran
MARGARET WINTER, Ph.D.
Research Associate In Chemistry
Centre for the Biology of Natural Systems
Washington University, St. Louis Missouri, U.S.A.
and
MARGARET L. THOMAS
Oepartment of Anthropology
Washington University
St. Louis Missouri, U.S.A.
In 1970 a preliminary study of chlorinated-hydrocarbon insecticide
residues In human milk from developing countries, forty-seven samples of
human milk were collected from three rural areas in Southern (Pacific
Coastal) Guatemala: La Bombs, Cerro Colorado ant El ftosari0. The average
concentrations (and range) of total DDT in whole milk in the three commu-
nities were as follows; la Bomtia: 2,93 ppm total DDT to.HI-ll.50 ppm;
10 samples); Cerro Colorado: 4.07 ppm {I.57-I2.2I ppm; 9 samples); El
Rosarioi 1.84 ppm 10.342-4.97 ppm; 27 sample,*).1 These were the highest
residues of DDT reported so far In mothers' milk from any country.2 Com-
pared with the WDO/FAO's Acceptable Dally Intaks (ADI) of 0.005 mg DOT/kg
body weight for humane, and considering that a new-born Infant consumes
about 0.6 lit milk/day and usually weighs no mors than about 3 kg, the
Guatemalan peasant neonate Ingests some 14-488 times (an average of 94
times) more ODT per day than the Acceptable Dally Intake established by
WHO/FAO,
- Both El Roeario and Cerro Colorado have had heavy agricultural
spraying on near-by cotton plantations while La Bomba has had very little;
meanwhile the Malaria Eradication Programs had been constantly spraying
Inside the houses in ail three communities for some 13 years, Yst at the
P <0.01 confidence level there was no significant difference among the
DDT levels In nether'e milk from the three communities. Hence, we hypoth-
esized that the main cause of the unprecedentedly high accumulation of
DOT in human milk in this area was indoor spraying of DOT for malaria
eradication.
To test this hypothesis further, a more extensive study was carried
out In, late 1973/eariy 1974, Involving S60 samples from nins areas includ
ing the three previous communities. (Thirty more samples were collected-
from women In a tenth region, but we ehall not Include these in the
pressnt report because we do not yet have sufficient data on the commu-
nity type and the exact patterns of insecticide use). The present study
lends strong support to our previous hypothesis that the main cauee of
the unusually high residues of DOT in human milk Is Indoor spraying by
malaria eradication teams.
Methode
Analysis of residual chlorinated hydrocarbon peetlcides was carried
out on a Varlan 244510 Gas Chromatograph equipped with a high temperature
Sc^H electron capture detector. The method was that recommended by the
Environmental Protection Agency.' Blase 6'nl/ti" (l.d.) columns packed
with I,sf> OV-17 plus 1.95# Ov-210 on Gai Chrom 1) (80/100 mesh) were used
for quantitation. Confirmation of Identity of unknown substances was
achieved by comparison of their retention times with those of known pestj,
cldss on this column and on a column Z&'*l/l6" U.d.i/ of 5?" Qf-I °n
Varaport 30 (100/120 m«»h5. Standard solutions of peetlcidee were pre-
pared and used for both qualitative and quantitative purposes.
Results
Seventy-eight of the milk samples are from three groupa of women
In Guatemala City, and the remaining 182 are from six rural areas.
The reeults of the analysis are presented In Tablet I-7. We ehall
focue the discussion primarily on the emplee from the rural areas.
p,p»-0DE and p,p'-OOQ are expressed as equivalents of DDT.
Table I. Organochiorine Insecticide residues in human milk from
Livingston, Izabal, Guatemala (population: approx. 3,000)
Taole I. (Continued)
Sample
p,p'-ODE
o,p'-ODT
p,p<-000
p,p'-DOT
Total DOT
(7>4KC
L. 1
O.47O
0.025
0.031
0.254
0.780
0.005
L- 2
O.Oja
0.010
0.017
0.088
0.173
-
L- 3
0.177
0.013
0.025
0.236
0.451
-
L- 4
1.147
0.158
0.114
1.258
2.677
•
L- 5
0.617
0.024
0.093
0.435
1.169
•
L- 6
0.059
0.006
0.016
0.059
0.140
•
L- 7
0.142
0.022
0.034
O.334
0.53?
-
L- 6
O.I50
0.010
0.028
0.128
0.316
•
1-9
0.298
0.016
0.025
0.259
0.588
Samp It
PjP'-DOE
o.p'-ODT
p.p'-DOD
p,p<3DT
Total-DDT
fl-BHC
L-IO
0.699
0.023
0.120
O.443
1.285
-
L-11
2.964
0.134
0.220
2.368
5.686
0.014
L-12
0.522
0.020
O.O65
0.3J4
0.941
-
L-13
0.133
0.004
O.OIJ
0.090
0.240
0.006
L-14
0.133
0.016
0.048
0.204
0.401
-
L-15
0.668
0.038
0.133
0.457
1.296
-
L-16
0.196
0.015
0.032
0.169
0.412
-
L-17
0.486
0.016
0.027
0.210
0.739
L-18
0.853
O.O57
0.062
0.845
1.817
0.001
L-19
O.172
0.011
0.012
0.15?
0.354
0.001
L-20
O.O85
0.012
0.022
0.126
0.245
•
L-21
0.794
0.094
0.060
0.553
1.501
-
L-22
0.154
O.OI3
0.019
0.186
0.372
-
L-23
0.212
0.01;
0.041
0.201
0.467
-
L-24
0.143
0.014
0.019
0.159
0.335
-
L-25
0.284
0.016
0.024
0.224
0.548
-
L-26
0.145
0.009
0.024
0.100
0.278
-
L-27
0. 110
0.011
0.022
0.080
0.223
-
1-28
O.582
0.012
0.049
0.271
0.914
-
L-29
0.096
0.015
0.017
0.118
0.246
-
L-30
„V6i-
0.034
0.012
Of4§l .
0.792
-
Table 2. Organochiorine ineecticide reeiduee in human oil Ik from La Bomba
5
0.122
0.432
0.060
0.395
0,292
O.I99
0.272
o.oai
0.105
0.090
0.140
0,09a
0.221
0,174
O.432
0.319
0.J20
0.2=7
0.17'
0. 166
O.J 12
-0*522-
0.477
0.182
0.444
O.365
0.295
0.089
1.764
0. 136
0.420
0.296
0.662
0.203
I.OO7
0.784
0.461
0.628
0.194
0.4JI
0.294
O.33O
0.280
0.701
0.494
1.290
-------
Table 3. (Continued)
Tab It s, (Continued)
Sample p,p'-DD£ o.p'-ODT p^'-DOO p,p'-PPT Total DDT P-BHC Sample p,p'-DDE o,p'-DDT p,p'-DDD p,p'-DDT Total DDT (J-BHC
AM- 8
0.321
AM- 9
0.191
AM-10
0.378
AM-II
0.261
AM-12
0.101
AM-13
0.624
AM-14
0.339
AM-15
0.424
AM-16
O.573
AM-17
0.826
AM-18
0.405
AM-19
0.372
AM-20 .
0.212
AM-21
0.128
AM-22
0.06a
AM-23
1.872
AM-24
O.I57
AM-25
0.333
AM-26
0.194
AM-27
°.l»
AM-28
0.060
AM-29
0.741
AH-30
0.603
0.005
0.030
0.002
0.006
0.006
0.0 IE
0.012
0.00}
0.005
0.011
0.00J
0.007
O.OOJ
0.002
O.OOJ
Q.OIB
O.OO5
0.008
0.004
0.011
0.004
0.010
0.013
0.008
0.128
0.462
0.004
0.018
0.177
0.406
-
0.016
0.083
0.479
0.004
0.009
0.093
0.369
0.001
0.007
0.071
0.185
0.001
0.016
0.203
0.855
O.OO5
0.019
0.142
0.512
0.001
0.021
O.O77
0.527
0.004
0.014
0.171
0.763
0.001
0.018
0.259
I.I 14
0.008
0.019
0.262
0.691
0.003
0.025
0.177
0.581
0.005
0.008
0.072
0.295
0.002
0.010
0.04J
0.183
-
0.002
0.041
0.114
-
0.122
O.494
2.506
0.018
0.010
0.072
0.244
0.006
0,010
0.106
0.457
0.002
0.011
O.094
O.3O3
TR
0.011
0.117
0.292-
TR
0.004
0.031
0.099
0.001
0.022
0.140
O.913
0.003
AW
o.isq
0.830
O.OOS
TR ! Present but leee than 0.001 ppm.
Table 4. Oroanochlorlne insecticide reelduee In human milk from Carro
Colorado, U Gomera, Eacuintla, Guatemala (populations appro*. 4,000)
Sample p.p'-OOE o,p'-DPT p,p'-000 p,p'-OPT Total OPT (3-8HC
CC- I
CC- 2
CC- 3
CC- 4
CC- 5
CC- 6
CC- 7
CC- 8
CC- 9
CC-IO
CC-II
CC—12
CC-13
CC—14
CC-IJ
CC-16
CC-17
CC—18
CC-19
CC-20
CC-21
CC-22
CC-23
CC-24
CC-25
CC-26
CC-27
CC-28
CC-29
CC-30
CC-31
0.376
0.023
0.237
0.274
0.062
0.126
0.127
O.O55
0.071
0.253
0.022
0.127
0.148
O.O58
0.119
0.505
0.904
0.214
0.684
I.60J
0.286
0.198
0.186
0.327
0.284
O.698
O.617
O.O49
0.167
O.25O
0.007
0.003
0.007
0,003
0.006
O.OO3
0.004
0.003
O.OO5
0,005
0.008
0.014
0,005
0.013
O.OO5
O.OO5
0,005
0,004
0.003
0.008
0.006
O.OO3
O.OO7
Q.0'3
0.013
0.004
0.011
0.020
0.006
0.007
0.023
0,008
0.007
0,013
0.003
0.008
0.014
3
0.442
0,000
GC-3I
O.O67
-
0.004
0,027
0.098
0.009
Gc-je
0.029
-
0,002
0,013
0.044
O.OO3
ac-33
o,a65
0.003
0.017
0,101
0.386
0.007
Gc-34
0,231
0.002
0.006
O.OJJ
0.296
0.006
GC-M
0.210
-
0.009
0,042
0.261
0.017
GC-36
0,236
0,002
6.009
0.059
O.306
O.OIJ
0.247
0,006
0.013
O.O69
\8S
0,070
GC-38
0.201
0.003
0,013
O.O99
0.004
Ge-39
0.10)
0,004
O.O34
0.U3
O.OJ4
GC-40
0.874
0,004
0.018
0,205
UIOI
0.010
(Continued)
7-7
-------
M't, 2-
(Continued)
Samp Is
p,p*-DDE
o,p'-DDT
p,p'-DDO
p,p*-ODT
Total DOT
/J-BHC
6C-4I
0.091
0.002
0.003
0.022
0.124
0,014
GC-42
0.133
0.001
0.006
0.020
O.160
0.008
5C-43
0.l«
0.001
0.004
0.022
O.170
0.006
GC-44
0.06l
0.001
O.OO3
0.014
O.079
0.004
GC-45
GC-46
0.336
O.OO3
0.0 II
0.140
O.5IO
0.0|6
0.147
0.001
0.011
0.079
O.238
0.003
flC-47
0.029
-
0.002
0.010
0.041
0.0 II
GC-48
O.I34
0.001
0.009
0.037
0.181
0.003
GC-49
0.191
0.003
0.014
0.081
0.289
0.011
GC-50
0.164
0.003
0.006
0.047
0.220
0.011
GC-51
0.039
-
0.004
0.012
O.O55
0.004
GC-52
0.457
0.012
0.025
0.117
0.611
0.040
GC-53
0.113
TR
0.003
0.019
O.I35
0.009
sc-54
0.094
0.001
0.004
0.021
0.120
0.010
GC-55
0.105
O.OO3
0.009
0.039
O.I56
0.009
GC-56
0.074
-
0.003
0.012
0.089
O.OOJ
gc-57»
0.040
0.001
0.006
0.017
0.064
0.016
GC-58
0.212
0.002
0.004
0.023
0.241
0.007
GC-59
GC-60*
0.155
O.363
0.002
O.OO3
0.008
0.018
0.025
O.O96
0.190
0.480
0.004
0.005
GC-71
O.OJO
-
0.001
0.005
O.O36
0.003
8C-72
0. 161
-
0.004
0.025
0.180
0.006
8C-73
0.396
-
0.006
0.044
0.448
0.010
6C-74
0. Ill
0.001
0.004
0.023
O.I39
0.010
GC-75
O.O32
-
0.002
0.014
0.048
0.004
GC-76
0.400
0.003
0.0 II
O.O78
0.492
0.01a
GC-7?
0.194
0.002
0.007
0.006
O.O34
0.237
0.005
GC-78
0.179
0.001
0.026
0.212
0.014
GC-79
O.O65
-
0.002
0.012
O.O79
O.OO7
GC-80
0.395
0.003
0.011
0.084
0.493
0.020
GC-81
0.233
0.002
0.010
0.046
O.29I
O.OI5
GC-82
0.032
-
0.003
0.006
0.041
0.004
GC-83
0.192
-
0.009
0.042
0.243
0.006
GC-84
0.226
0.004
0.009
O.O56
0.295
0.264
0.006
GC-8J
0.197
0.004
0.010
O.O53
0.010
GC-®
O.36I
0.007
0.012
0.092
0.472
0.018
GC-87
O.IPI
0.002
0.008
0.046
0.247
O.OI5
GC-89
0.260
0.006
o.on
0.1 If
0.394
0.015
TO 1 Present but less than 0.001
ppm.
*l>s ahowad 0.001 ppm °^-BHC.
Discussion
Tin first observation la that Livingston (Tabls I) shows ths high-
sst rssiduss of total DOT In thla study (avsrsgsi 0.864 ppm DOT; rang*:
0.140-5.686 ppm total DDT), which approach ths concentration! in our prt-
vloua atudy, aummariztd about. Llvlngtton la tht only community In which
ssmiannual Indoor DOT spraying for malaria control had continued up to
and Including tha tima of our atudy In I974. Tha main occupation of tha
papulation In Livingston ts fishing, with taint maize and bsnsna cultiva-
tion at a suppltmtnt in about ont-thlrd of our sample population. Inaac-
tlcidaa art uaad raraly In tha flald (ont In tight) but quiteoftan In the
housa (two-thirda), but In most casta thaaa ara non-chlorinattd psstici-
dts, Bsvsn of our donors wars ladlna woman (numbere LI, 10-i3, and 29-30)
ana the rasalnlng 23 wars Black,
U Bomba (Tabla Z) haa tht next hlghsst concantrations, with an
average of O.587 (rangtl 0.089-I.764 ppm total DOT). Semiannual indoor
DDT apraylng hart continued through 1972, after which it wat rtplactd by
propoxur— a carbamate insacticide sprayed four tints par year. In 1970
whan wa flrat studisd La Bomba, It was mostly a cattle-pasture, maize and
banana producing araa. Tha population ara tha paona of a number of
cattle ranchtt, all being ethnically Ladino (a term applied to Spsnlsh-
tpsaklng descendants of Spaniards mixed with tht Indigtnoua population).
Of our sample population of 31, twanty-aaven plant maize, of whom 22 uaa
peat lei dee to control paata, mostly non-chlorinatad insecticides. Only
two utt pesticides In the houathold, neither of tham a chlorinated
compound.
Tht main Changs sines 1970 had bttn that two plantations to ths
southwMt of La Bomba had bttn turntd into cotton producing land in tht
I9??*74 ttaaon. Six of tht donort live vtry clott to tht cotton planta-
tions which art apraytd by airplane (Not. 1820-24 and 28), and ont used
to livt on a cotton plantation (No. 189). But proximity to cotton fltlds
is not a significant factor in DDT accumulation htre ainet tht avtragt
concentration of total 00T In milk from the former six womsn is 0.579 pp«i,
tnd if wt Include LB9 tht average drops to 0.556 ppm (compartd with ths
overall avsrsgs of O.587).
Aeuncldn Mlta (Table 3) Is ns«t, with an avtragt of O.49O and a
range of 0.051-2506 ppm total DDT. Tlx Malaria Eradication Service
apraytd OUT Indoors in this Mrs* rtgularly until Sspttmbtr 1970. At
tht time of our sampling thty had not sprayed for thret-and-t-half
years. Howavsr, not long aftsrwards, thty btgan to spray ths araa
again. Tht area products mostly tomstoss tnd otusr vsgstablss; thsrs
ars also msizs-fislds snd milk-cow patturte. Insecticides art ustd In
ths flslds and In ths houses rathsr llbtraHy, but our investigation
showed that In almost all inatancta they art organo-phosphoroua and
esrbamats inaacticldta. Our donors were all ethnically Ladinsa.
Tht naxt two communitlss, Csrro Colorado and CI Rossrlo (Tablet 4
and 5), have svtn Itss, although fairly considerable amounts (avtragt)
0.466 and 0.276; rangsi 0.041-2.193 and O.O50-O.912 ppm total DDT,
rsspsctlvsly). Indoor OOT-apraylng opsratlona in these two communities
stopped aftar 1970, tht ytar of our prtvioua study, and wars also replac-
ed by propoxur. Csrro Colorado is a narrow and long raatttltmtnt arta.
To tha south and aouthwat, cotton had bttn ralaed for tight ytars. Tht
largs plantation which lines tht tast aidt raised cotton until it awltch-
td to malzt and sugsr cans six yssrs before ths sampling. Ths plantation
to tha northwsst alao ralatd cotton until It awitchtd to cattlt four
ytara prior to tht study. Most residtnts ralaa maize, but a few have
cattlt and other miscellaneous crops. Htrt, too, thtrt it rartly any
chtorlnatsd paatlctdas ussd in ths fields. Ths only ons mentioned at all
was sldrin, slthough a good numbar of organo-phoaphatea ars ussd. Our
sntlrs samplt population conalata of Ladlna woman.
While Cerro Colorado consists of small plota of land (just a few
hsctsres psr household), CI flosarl0 la a resettlement scheme with medium-
slzs farms— some ssvsn timss larger than in Cerro Colorado. Thla resst-
tlement area is surroundtd on almoat all aldsa by large cotton glants-
tiona. In addition, atveral of the plot-holders themsslvts plant cotton
and join foreas to hirs spray-planes. We do not have rtcsnt data on what
Inssctlcldss ara sprayad on cotton in tithtr community; but thtrt is s
difference in DDT-conesntration in milk from woman who live ntar cotton
In Ctrro Colorado (O.851 ppm total ODT for tha aavan samples numbered
CC19-20 and 22-26 aa compared with 0.354 ppm avtragt for tht rtmalnlng 24
who do not live In clost proximity with cotton) whlIt thtrt is no such
difference In £1 Roaarlo (0.253 ppm total DDT for tht II aamplta numbered
ER6, 9-11 and 14-20, aa compared vith 0.288 for tha remsining 20 samples
from woman who do not live near cotton). Further data on detailed Insec-
ticide consumption of the cotton fields would bs nesdsd to know If ths
grsatsr residua values of DDT in tht 7 Cerro Colorsdo milk samplss pan bs
attributed to added Influence dus to drift or othsr factora. Our sampls
population in both of thsss communitlss Is Ladlna, sxcspt for Nos, EH2,
26 and 30 who ars Indiana.
Ons conclusion that can already bs reached baaed on tht above data
Is that both the 1970 and 1974 data strongly support the hypothesis that
themalnsourca of residues in human milk In tht rural artat of Guatemala
Is Intradomicl llarv spraylna of OUT hv the malaria eradication oroorammes.
Agricultural usss form a secondary sourcs. A second conclusion is that
when Indoor saravlno of DDT is interrupted, as in La Bomba, Cerro
Colorado and El Roaarlo, ths rssidus levels In human milk drop rather
rapidly, In a matter of a few yVara.
What yields further support to our flrat conclusion Is ths data
from Ntbaj, Ouicftd, whsrt tht avtragt concentration is only O.O35 ppm
total DDT, with a range of Q.005-0.I8J ppm for 2B samplss (Tablt 6).
Thlt it an iaolattd rural ssttltmtnt in a small vallty in tht CuchumstJn
mountain rangt In ths northwsstsrr highlands of Guatemala. Most psopls
ralaa malzt, but there are aoms coffst plantations not too far sway and
an orangs orchard, too. Only thras of ths donors rsportsd any insscti-
cidss being ussd inthtlrmaizs-fislds, but about two-thirda aald they
used household Ineectlcidas, including Gamexin, which la a mixture of BHC
and DOT. This probably explains the sxlsting, although relatively low,
levels of DOT and BHC rssiduss in Nsbsj mothsrs' milk. Ths rslstivsly
high domestic uss of various inssctlcids brands Is explained by ths fact
that s mobile pickup vehicle comss around on market daya to asll pssti-
cidts In this isolatsd community. Thsrs Is no DDT spraying for malaria
control aa Ntbaj Is not a malarious arts.
Finally, ths data from Guatsmala City ahow a Itvel of DOT rssiduss
which 1s lowsr than ths coomunitlss whtrt intradomiclMary spraying of
DDT for malaria control has bstn praeticsd. (No malarial apraylng of DDT
sxlsta In Guatsmala City). Nsvtrthtless, thsrs Is a rathsr eonsldtrabls
level ranging from 0.015 to 1.01 ppm total DOT, with an average of 0.233
ppm In 78 samplss. Tht most probable explanation hart it that residsnts
of ths capital city corns from a diverse background, and thsir dlst, too.
Included in thsir dlst ars itsms such aa swat and fiah from tht hsavlly
polluted south cosst (many cattla-paaturas ars formsr cotton plantations).
Ths normal diet in Guatemala city (aa we 11 aa ths fish) contsins a rathsr
heavy doss of ODT at rsportsd by us slsswhsrs. In addition, 38 of ths
78 womsn rsportsd using housshold Inssctlcldss, in soms casts DOT and
Gamexln.
In ordar to evaluate ths toxicologies! significance of thsss concejj
trationa, we can only compare ths rssiduss with ths WHO/FAO established"
Acceptable Dally Intaks (ADI) valuss for humans. Tablt 8 shows ths comM
risont
Table 6. Comparison of maximum, minimum and avsrsgs vsluss of ODT in
sach community with ths UHQ/fnO Acceptable Daily Intaks (0.005 mg/kg
body wsight)
Community
Livingston
La Bomba
Asunei6n Mita
Csrro Colorado
E) Roaarlo
Gustsmsla City
Ntbaj
NO. of
30
31
31
31
31
78
28
J^
71
100
86
36
44
70
-IiffllLML
minimum
5*6
3.6
2
1.6
2
0.6
0.2
. iv^at
24
20
19
II
9.3
1.4
3
7-7
-------
Th* tab la shows that th* *v*rag* valuta of th* dose actually takan
by infanta in all th* communltl** studied exceed th* *stabli*h*d maximum
tolaranee limit*, even In N*baj which It an laolated community, it also
(how* that th* infant* take an avtrag* of I.4-35 tin** nor* DOT In thair
diet from mot liar's milk a loot; that a II infant* In Livingston, u Bomba,
AaunciSn Klta, Ctrro Colorado and CI Rosarlo take quantltie* of DDT above
the tolerance level; that only In Guatemala City and Nebaj In thia *tudy
are thar* aome infanta who take lesa DDT than permitted Internationally
(In Nebaj they take a* little aa 0,2, I.e., five timoe ieaa than AOI);
but that th* maximum quint I tits taken in our etudy population* rang* frcei
7.3 to 227 times more than th* AOI.
Unfortunately the AOI Is th* only meaeure of harmful effects w*
have. Toxicity atudit* in insecticide* ere aa a rule ba**d on h*althy
adult papulation*, auch a* thoa* reporter by Hay**;4 unfortunately con-
clusions baaed on adult populationa In wll-f*d countrit* art of llttl*
value in evaluating acuta and, even nor* importantly, aub-acuta effects
of ouch poison* in und*rd*v*lop*d countries wh*r* there ia u*ually a
general nutritional deficiency coupled with a multitude of other threats
to health of th* infant population.
A brief evaluation of th* *xietlng information on DOT toxicity
ha* been made elsewhere,1 In th* Sam* work a detailed analysis of th*
raauft* of some 14 years of malaria eradication in Centra I America show*
that littl* or no gain hat b**n mad* againat thia diataa* following th*
conventional approach to malaria eradication by indoor spraying of DDT,
In aome Central American countrlee physiological and b*havloral resis-
tance in Anophei— alblmanua hat cau**d a atandatill tn th* antimalarial
war, and TheHncTdencTofth* disease he* aome tlmee reached or exceeded
preeradlcation lavel*. In moat Central American countrl*a dialerIn, SOT,
malatiion and propoxur hay* in turn r*placed each other at reaiatance in
anophellnea hat mad* them u**l**«. Th*r* i* now no jntecticid* u*abI*
in a malaria eradication programme that ha* r*main*d Immune from anophs-
IIne reel stance, it ie therefore time w* begin to revise our *trat*gi*«
againat thia tluaiva dittta* giv*n th* world-wid* **t-back*. Habitat
management, biological control, and above all, Improvement* in th* *ocio-
aconomic standard* of living of th* general population are th* only fac-
tor* that will ultimately h*lp u* conquer malaria In the developing
countries.
The normal concentration of DDT uatd In malarial apraylng ia
2 gm/m2 of wall aurfac*. If w* consldored th* *pray*d wall lurfac* to
b* equivalent to an agricultural surface art*, th* equivalent would be
40 kg/ha per year, which would b* an unpr*c*decit*d rat* of application
for DDT. It la no wondtr auch high concentrations are finding th*lr way
into the human body and ther*for* ultimately canta«lnating th* human mf Ik
on which chi ldr*n in developing countrl*e g*n*rally depend axclualvaly
from six montha to two year* or more. Our data from th* prasent atudy
•how, however, that then i* hop* for eliminating thia added burd*n on
th* growth and d*v*lopm*nt of childr*n in developing countrl** such as
Guatemala: that by etopplng th* Indoor *praying of Ineecticldea and
changing our atratagl** of malaria eradication to thoa* based on habitat
management and tru* economic d*v*lopm*nt, w* might at last achltv* both
an *limination of th* acourg* of malaria and of *xc***lv* DDT load* from
nature'e otherwis* perfect food—* mother's milk.
Beference*
I. Farvar, M. Taghi. ecological imp 11 est Ione of ln«*ct Control In
Central America] Aoriculture. Public H*alth and Otvalonmsnt. ~
hi gam u • —" ' - *
Ann Arbor, MIdhlgani Unlvereity Microti Ims, 1972.
2, Only on* prtviou* study by V.I. Oamatkini 0 *tl«pieni kumulatsli
ODT v organlzmlt chelovleka pri postupleniI piehchaviml produktaml
i yago tokeicheekom vpidleyetvll. 6ln. Ssnlt. JOi 109, 196;
(Degree of accumulation of ODT In the human body due to ita
presence in food*, and It* toxic effects) report* almlltr amount*
(1.22-4.88 ppm ODT] in human milk, but without mention of whether
these quantitlaa are In whole milk or only in tha lipid portion,
y*t, from a description of th* area, th* population Involved and
. tht dtgr** and kind of Ineecticide usage in th* ar*a, aa wall 1*
paraonai communication with hit Sovltt colleagues, wa have coneluded
that th*a* quantltlt* are mo*t likely In th*. lipid fraction.
Therefore, th* equivaUnt conctntratlona In who I* milk can be approx-
imately eat 1 mated aa 0.061-0.244 ppm total DPT (aaaumlng a lipid
contant of yji), which would plat* th*** valu** in th* mm rang* at
comparebl* populationa In tha U.S. and Europe.
,HnijMH.|.n .flwyrm
Fsrrln*, FUrUM frlmata and Pesticide*
3. Thompson, J,F, (*d)«
£ff*ct^Uboratory,%ErtvIromtnti I Protection Agency, 1972.
(revleed, 1974). Section 5* U) *"d (B), III, pp. 2-?.
4. Hayes Jr., W. J. U inoeuidad del DOT *n *1 hombrt dweostrada en *|
control d* l» malaria, youtfn d* la Otleint Sanitaria fanamarlcan..
?>48M99, 197".
-------
ENVIRONMENTAL MONITORING IN LATIN AMERICA AND THE CARIBBEAN
Ricardo Haddad and Walter Castagnino
Air Pollution Adviser Water Pollution Adviser
Pan American Center for Sanitary Engineering and Environmental Sciences (CEPIS)
Pan American Health Organization (PAHO)
Summary
An international air pollution monitoring network
(REDPANAIRE) was begun in 1967 by PAHO, which currently
sponsors and coordinates it. A summary is presented of
data collected, trends, methods used, rationale behind
their selection, and some of the1 difficulties encoun-
tered. Future development and operating alternatives
are discussed. As the only existing international net-
work of this type, it may serve as a model for other
developing regions, and could be used as a reference
for a global air pollution network.
Varying degrees of information on water quality
have been collected in Latin American countries. Short-
age of funds and skilled personnel are now limiting data
acquisition to the most pressing needs. Forecasting
of water quality conditions is being done in some areas,
and there are reports on basin-wide evaluation of base-
line water quality and pollution control programs. Out-
standing examples are cited and discussed. Stress is
placed on the need for proper balance between informa-
tion to evaluate and forecast water quality conditions
in L&tin America.
Introduction
Latin America and the Caribbean include at present
26 countries and several dependent territories, in very
different stages of development. Together with the
United States and Canada, they constitute one of the
six regions of the World Health Organization (WHO) - the
Region of the Americas - attended by the Pan American
Health Organization (PAHO).
The Latin American and Caribbean countries are, as
a whole, in an intermediate stage of development, some-
where between the less-developed countries of most of
Africa and Asia, and the industrialized world. Thus,
their health and environmental problems include those
typical of underdevelopment, as well as those recognized
as products of industrialization. Due to improved san-
itary conditions, their mortality rate is relatively
low and they have, as a consequence, the highest demo-
graphic growth in the world - an average of 2.9% annual-
ly. Although nearly 50% of the population lives under
rural conditions, scattered over a very wide territory,
normally with low standards of living, there are 17
urban conglomerates with over 1,000,000 inhabitants
each. Four of these have more than 6,000,000. Most of
them are very modern and are growing at a pace of 5 to
8% yearly. This rate of growth compounds the difficult
problems of underdevelopment and, at the same time,
modernization and industrialization have produced the
familiar ones related to the deterioration of environ-
mental quality. Air and water pollution are increasing
and have reached, in some cases, situations as severe
as those found in the industrialized world.
As one of its programs aimed at protecting health
in the Region of the Americas, PAHO has been providing
technical assistance in the environmental field for
many years. The first important activities were started
through its regional consultants and engineers assigned
to individual countries. These programs were intensi-
fied after the Directing Council of the Organization
resolved, in 1965, "to request the Director to give
appropriate attention to the expanding problems of air
and water pollution in the Organization programs of as-
sistance..." In 1969, to further support and expand
these programs, the Organization established a regional
environmental technology center - the Pan American Cen-
ter for Sanitary Engineering and Environmental Sciences
(CEPIS), located in Lima, Peru. This paper will de-
scribe and analyze some of these efforts as they are
related to air and water quality monitoring.
Air Quality Monitoring
Some Latin American countries became concerned with
air pollution during the 1950's. Sporadic measurements
were taken, and efforts were made to control certain
sources of pollution. These activities spread to other
cities as they became aware of the seriousness of the
problem and started to forecast their demographic, in-
dustrial, urban and transportation development. The
obvious and sometimes rapid deterioration of atmospheric
quality, combined with the knowledge of the situation
in the United States and Europe, lent greater urgency
to these activities.
Santiago (Chile) and Sao Paulo (Brazil) were the
cities where the situation was most publicized; conse-
quently, efforts for evaluation and control were greater
there. As could be expected, interest increased as the
problem worsened, spreading from health officials to
administrative authorities, to the press and the public
at large. In 1961, PAHO helped to organize and set up
the Institute of Occupational Health and Air Pollution
in Santiago to work in research and education in these
fields. Subsequently, the Organization was instrumental
in the creation of the Institute of Sanitary Engineering
of the State of Guanabara in Brazil, and collaborated
in a control program in Sao Paulo which began in 1963.*
At the same time, other cities such as Lima, Mexico City
and Rio de Janeiro were carrying out fairly extensive
studies. Buenos Aires was the site of a Latin American
Air Pollution Congress in 1962 and a World Congress on
the same subject in 1964. In 1966, PAHO decided to ap-
point a permanent Air Pollution Adviser and to sponsor
the establishment of the Pan American Air Pollution
Monitoring Network, currently supervised by CEPIS.
During the last 15 years, PAHO has assisted in
drawing up general programs and/or solving specific
problems at the request of almost every country within
the Region. The Organization has also sponsored inten-
sive short courses, graduate scholarships and travel
grants to help train local personnel and organized a
"Latin American Seminar on Air Pollution" in Rio de
Janeiro in 1968 and a Symposium on "Environment, Health
and Development in the Americas" in Mexico City in 1974.
1
8-1
-------
The Pan American Air Pollution Monitoring Network
(REDPANAIRE)
As a strategy to develop the interest of the Member
Countries and to gather information on the real magni-
tude of the problems, it was decided to start a network
of sampling stations which would gradually include the
main cities of the Region. Particulate matter and sul-
fur dioxide were selected as the parameters to be mea-
sured, since they are good indicators of existing con-
ditions and can be used to establish trends when a suf-
ficient number of measurements are available.
The standardized methods used in England and other
European countries, endorsed by the European Organisa-
tion for Economic Co-operation and Development,1 were
adopted. They have the advantage of being inexpensive
and easy to set up, and thus are particularly appro-
priate for early stages of new programs. Equipment for
a first station could be purchased in 1970 for US$ 1,050.
Additional stations in the same city coat only US$ 500
each, as there is no need to duplicate analytical
equipment. Ahnual maintenance cost is under US$ 100
per station. The support of a minimal laboratory is
sufficient for analyses. Only two technician-days per
month and 30 additional minutes per day to change fil-
ters apd solutions are needed.
The possibility of using the high-volume sampler
for particle collection and the colorimetric pararos-
aniline method for determining sulfur dioxide was also
studied in depth. However, it was felt that they would
require a trained chemist. Weighing of the high-volume
samples must be done under ambient conditions of less
than 55% relative humidity. Since humidity is normally
higher than that in most of the cities involved, expen-
sive chambers with controlled humidity would have to be
built.
Instrumental methods were not considered. Equip-
ment was too expensive for the available resources.
Lack of personnel for operating, calibrating and main-
taining instruments would have been an unsurmountable
difficulty. Besides, most instruments were not very
reliable, nor able to give much better results than
manual methods. Since this type of equipment was under-
going constant technological improvement, it would have
been difficult to select instruments that would not
become rapidly outdated.
Equipment for 10 stations was bought in 1966 and
a detailed Manual of Operations was prepared, both in
Spanish and English. Contacts were made in several
countries and agreements were drawn up outlining the
bases of cooperation between the respective Governments
and PAHO. The most suitable sites for stations were
sought in each city, and the participating countries
were encouraged to expand the program by purchasing and
operating additional stations.
Equipment began arriving in April 1967. The first
station started work in June of the same year. By No-
vember, eight of the originally planned stations were
operating. Various countries later used their own funds
to add more stations. In December 1974, REDPANAIRE had
93 stations spread among 30 cities in 15 countries.
Additionally, there were 126 smaller stations in 6 cit-
ies measuring only settled dust. At least 30 stations
are presently being installed or are in the planning
stage. Some Latin American cities such as Buenos Aires,
Caracas, Havana, Mexico City, Rio de Janeiro' and Sao
Paulo, have other evaluation programs besides the Net-
work,
Knowledge of the effects of air pollution on health,
welfare, and the economy is still not very precise and,
consequently, there is no agreement on just what should
be considered clean air. Laws passed by a few countries
show important differences in relation to permissible
levels, the ndmber of pollutants included in the regu-
lations, and ways of reporting concentrations. Shortly
after REDPANAIRE began operating, reference levels were
suggested to facilitate the interpretation of results.
The following levels were recommended after studying
existing legislation up to 1967, findings of other mon-
itoring programs, and data on the socio-economic and
health implications of air pollution:
Table 1. Reference Levels suggested by
PAHO/CEPIS for interpretation of REDPANAIRE results
Parameter
Monthly average
Settled dust
0.5 mg/cmz/30 days
Suspended dust
100 micrograras/m3
Sulfur dioxide
70 micrograms/m3
The figures given above do not attempt to set in-
ternational limits, a function beyond the scope of an
individual institution. Their sole purpose is to provide
a basis for comparison to help determine the quality of
ambient air in a city as a function of results obtained
by the Network stations. It must be clearly understood
that these reference levels are only intended to be used
for comparing the monthly averages of samples collected
and analyzed by the Network stations.
Five years later, in 1972, a WHO Expert Committee2
suggested long-term air quality goals which, for par-
ticulate matter and sulfur dioxide, were based on the
same methods used in REDPANAIRE. They were lower than
the reference levels suggested by PAHO/CEPIS, reaching
only 40 Mg/m3 for suspended dust and 60 pg/m3 for sulfur
dioxide.
REDPANAIRE results
Two consolidated reports have been prepared and
distributed by CEPIS. The first one3 included the data
collected from the beginning of the Network, in June
1967, until December 1970. The second one1' included
results of the first report plus those collected until
December 1973. The first report summarized approximate-
ly 40,000 results; the second included around 250,000.
Table 2 summarized the results obtained by
the monthly averages with the reference levels suggested
by PAHO/CEPIS:
Table 2. Distribution of monthly averages of
results obtained by REDPANAIRE stations, compared
with Reference Levels (RL) suggested by PAHO/CEPIS
1967 - 1973
Settled
dust
Suspended dust
SO;
J
Fraction of RL
So. of
stations
X
80. of
stations
*
MO. of
stations
t
Less than 0.5 RL
1
1.4
37
45.1
43
53.1
0.5 to 1.0 RL
12
17.1
29
35.4
13
16.0
1.0 to 1.5 RL
16
22.9
9
11.0
16
19.8
1.5 to 2.0 RL
9
12.9
6
7.3
7
8.6
Over 2.0 RL
32
45.7
1
1.2
2
2.5
Totals
70
100.0
82
100,0
81
100.0
2
8-1
-------
If the comparison is made with the long-term goals
proposed by WHO, the following distribution is obtained:
Validity of results
Table 3. Distribution of monthly averages of
results obtained by REDPANAIRE stations, compared
with Long-Term Goals (LTG) proposed by a WHO Air
Pollution Expert Committee. 1967 - 1973
Suspended dust SO2
Fraction of LTG
No. of
No. of
stations
%
stations
%
Leas than 0.5 LTG
14
17.1
40
49.4
0.5 to 1.0 LTG
17
20.7
14
17.3
1.0 to 1.5 LTG
12
14-6
13
16.1
1.5 to 2.0 LTG
10
12.2
7
8.6
Over 2.0 LTG
29
35.4
7
8.6
Totals
82
100.0
81
100.0
Since the Long-Term Goals proposed by WHO for sus-
pended particles are only 40% of the PAHO/CEPIS Refer-
ence Levels, important differences can be found between
Tables 2 and 3. For sulfur dioxide, figures suggested
by both organizations are similar and differences are
of a lesser magnitude.
Some measures of air pollution trends in the cit-
ies included in the REDPANAIRE can also be extracted
from the two previously-mentioned reports. In Table 4
a comparison is made between the values obtained during
the periods 1967-1970 and 1971-1973 for the 25 stations
which show results for both periods:
Table 4. Variation of results from 1967-1970
and 1971-1973 for 25 REDPANAIRE stations with
results for both periods
Settled dust
Suspended duet
S02
Type of
variation
No. of
sta-
tions
X of
total
Average
varia-
tion
*o. of ,
¦«- L°*\
tions
Average
varia-
tion
No. of
sta-
tions
X Of
total
Average
varia-
tion
Increment
18
72
36.5
15 60
29.3
17
68
42.4
Decrement
7
28
12.9
10 40
19.3
8
32
21.7
For the three parameters measured, between 60% and
72% of the stations show an increment, with only 28% to
40% of the stations showing diminishing values. The
average increment is also much higher than the average
decrement. This suggests that in most of the stations
there is a trend to higher pollutant concentrations.
To establish this trend, a statistical analysis
was made5 of 35 stations with data covering at least 24
months during the period 1967-1973. Results are pre-
sented in Table 5:
Table 5. Trend of monthly averages for three
parameters monitored by 35 REDPANAIRE stations
with 24 or more months of data. 1967-1973
Settled
dust
Suspended dust
S02
Trend
No. of
stations
Z
No. of
stations
X
No. of
stations
X
Increasing
Decreaeing
No statistically
significant trend
13
4
15
40.6
12.5
46.9
12
8
15
34.3
22.8
42.9
13
10
12
37.1
28.6
34.3
Totals.
32
100.0
35
100.0
35
100.0
The validity of results obtained by the REDPANAIRE
method has been a subject of debate. For suspended
particles it is difficult to decide which is the best
method, or even try to make a valid comparison between
two different methods. Particulate matter is not a
chemical or physical entity and is really defined by
the method of collection. Apparently minor changes in
the procedures or instruments used, characteristics of
pollutant production in a city, or seasonal changes can
greatly influence the results. No comparison can be
made between results obtained by two different methods.
Dust concentration should be compared only within the
same station, and used to determine trends and the ef-
fectiveness of control programs.
With regard to sulfur dioxide, the situation is
different. Sulfur dioxide constitutes a specific chem-
ical substance. Different methods can be used to es-
tablish its atmospheric concentration and those which
best reflect the real situation can be accepted. The
pararosaniline method is considered specific for SO2.
It has been officially approved by the U.S. Government
and has been suggested as a comparison method to estab-
lish the precision of other methods, both manual and
instrumental.
A paper prepared by the Institute of Occupational
Health of Peru6, with the assistance of CEPIS, showed
that results obtained with the hydrogen peroxide method
represented an average of 82.4% ± 3.4% of the content
of a special atmosphere prepared with the help of a
permeation tube instrument. Those results were not
affected by varying the flow between 2.6 and 5.2 liters
per minute, nor by different changes introduced in the
sampling train. Reproducibility can then be accepted
and precision can be established at around 82%.
Simultaneous 24-hour samples of Lima air are being
taken at the CEPIS laboratory to measure SO2 concentra-
tion by both the pararosaniline and the hydrogen perox-
ide method. According to the 35 samples collected from
May 15 to June 19, 1975, both methods yield similar
results. A mean value of 5,68 jjg/m3 was obtained with
the acidimetric method as compared with 6.04 yg/m3 by
the pararosaniline method - a 6% difference. Although
these results are considered preliminary, there does not
seem to be an important difference between the two meth-
ods. This coincides with the findings in the Mexico
City stations, which shifted from the hydrogen peroxide
to the pararosaniline method in September 1974. No
important difference has been found.
Outlook for the future
The PAHO/CEPIS experience with REDPANAIRE could be
a valuable guide for establishing an International Net-
work of Air Pollution Sampling Stations. Among the
factors which contributed to its initial success were
the selection of simple and inexpensive methods which
are easy to perform even by inexperienced personnel,
the low cost of the equipment required, and the use of
a detailed Manual of Procedures to assure similar per-
formance of all stations. One source of complaint from
the participating countries was the fact that results
were not published more frequently. There is an urgent
need to find a more expeditious way of collecting the
data produced in the countries, which should be pub-
lished and distributed at least twice a year.
Acceptance of PAH0 recommendations and the Manual
of Procedures has been very good. Most of the partic-
ipating countries and cities are pleased to have the
opportunity to do something concrete about their air
8-1
-------
pollution problems. The feeling of belonging to an in-
ternational network has given more strength to their
programs. Even countries which initially showed some
reluctance to join the Network, because they considered
the sampling and analyzing methods too unsophisticated,
decided to join it and later added several stations on
their own.
For the future, PAHO intends to keep expanding the
Network by encouraging the participation of new cities.
By 1980 it will probably include between 250 and 300
stations, located in every country and important city
within the Region. A choice of different methods of
analysis will be offered to the participating countries.
Of course this will complicate the periodic reporting
of data collected, but it seems to be a logical step
for a better evaluation of the magnitude of air pollu-
tion problems within the Region.
Evaluation is only part of a control program, and
measuring is only part of evaluation. The countries
are being encouraged to establish sound air pollution
programs with emphasis on prevention. This is being
accomplished in several cases, especially where the
problems are Host urgent. PAHO will continue to assist
its Member Countries with these activities and would be
pleased to have the opportunity to collaborate in a
global air quality monitoring effort.
Water Quality Monitoring
Although it may be an oversimplification, there
are basically two reasons for monitoring water quality:
1. To evaluate baseline water quality and water pol-
lution control programs.
2. To forecast future water quality conditions.
Water quality problems are currently affecting
water uses of socio-economic significance in many areas
in Latin America and the Caribbean. The increasing
menace to the availability of proper water quality for
different needs has been pointed out in a recent short
report.7 According to that report, in 1970 more than
45% of the population was living in cities with more
than one million people. Six of these cities, with a
combined population of 22,7 million, are coastal com-
plexes, while nine, with a total population of 29.1
million, are discharging sewage effluents in relatively
minor streama or inland water bodies. At that time,
population served by some kind of submarine outfalls or
wastewater treatment plants amounted to no more than 5%
of the total in the first group and less than 17% in
the second. A further indication of the problem is the
fact that 28% of the urban population and more than 70%
of the total industrial production of the Sub-Region are
concentrated in four metropolitan centers (Sao Paulo,
Buenos Aires, Rio de Janeiro and Mexico City).
Competition for water is already being felt in a
number of cases. As examples we might mention the hydro-
electric and the low water augmentation needs in the
Upper Cauca River, Colombia; the hydroelectric water
supply and recreation demands in Billiiigs Lake in Sao
Paulo, Brazil; the beach recreation and waste disposal
needs in Montevideo, Uruguay; and the navigation and
housing development requirements on Madden Lake in the
Panama Canal area, to cite only a few. '
Some early monitoring programs
Most countries in the Region have been collecting
water quality information for several decades with vary-
ing degrees of coverage and for different purposes. It
is, therefore, not easy to judge the overall usefulness
of the data obtained. Nevertheless, some common data
acquisition trends can be pointed out after being ex-
posed to large segments of the available information:
1. Most data were collected with a single purpose in
mind. Most frequently this purpose was to determine
the quality of raw water supply sources for communities
or industries.
2. Differences in population and economic levels, even
within a single country, are reflected in the amount
and quality of the data,
3. Scarcity of human and material resources, includ-
ing laboratory equipment and sampling devices, has de-
layed or even impaired monitoring programs.
4. The available hydrological data, in many instances,
either covers only a short time span or is not very
reliable. This has represented an additional roadblock
to proper monitoring studies.
Notwithstanding the above shortcomings, some stud-
ies have been done to evaluate basin-wide water quality
information. One outstanding example is the PAHO report
on "Water Quality in the Plata River Basin,"8 This
includes the compilation, collation, and evaluation of
data available in 1969 in this vast area (about 3.6
million square kilometers or 1.4 million square miles)
together with some analysis of impact on main water uses
and broad characterization of pollution sources. Some
illustrations are shown as Figures 1, 2, and 3. Other
efforts worth noting are the COPLANABH (ComisiSn del
Plan Nacional de Aprovechamiento de los Recursos Hidrau-
licos) study in Venezuela9 and several reports included
in the ECLA (Economic Commission for Latin America,
U.N.) country-wide studies on water resources,10
Figure 1
SOLID LOAD IN PARANA RIVER SYSTEM
ACCUMULATED
ANNUAL TOTAL
(In thousand tons)
17SH
Kw«»)r. UH7
toran*
R«MfW
I
OCT. WW
0!C.
WFUlKNCt OF BCTMEJO RIVF.H SOLID LOAD DISCHARGE WIO PARANA MVBt SYSTCM I» SHOW
DISCHARGE AT 1000 K« FK* BUENOS AIRES IS FH.T IN OIFERSOT IMNTHS IN TOE PORTS OF
ROSARIO (300 Km FRCW MOUTH) AND BUENOS AIRES
4
8-1
-------
Figure 2
WATER POLLUTION IN THE HATANZA RIVER - BUENOS AIRES AREA
PAHO also sponsored water quality information meet-
ings in the Plata River Basin11 and produced some re-
ports on uniform methods of laboratory analysis, data
storage and retrieval, and sampling programs. One ex-
ample of the latter is a comprehensive report on water
quality monitoring prepared for PAHO and its Member
Countries by the Battelle Memorial Institute.12 These
gatherings and reports should be carefully patterned
to the existing needs and conflicts regarding water use
in Latin America and the Caribbean, and take into ac-
count differences in resources and their availability.
Present monitoring needs
Member Countries of our Organization are increas-
ingly aware of the need for water quality management
programs and the costs involved. It has been estimated7
that water pollution control for the next 10 years will
cost in the neighborhood of six billion dollars, if the
goals set by the Ministers of Health of the Americas at
their III Special Meeting (Santiago, Chile, 1972)13 are
to be met. The Governments are, therefore, demanding
that their technical institutions use the limited re-
sources available for tackling these pressing problems
as efficiently as possible. Monitoring, as part of the
effort, will not be the exception. Decisions must be
made regarding what, how and how often to measure.
Figure 3
Torn* PT Moh V*«to
CIU0A0 OE PORTO AlEGRE
t Diluvio
Tom# PT M*nino D»u»
Dricirg* P do M«lo
flio Cuaiba
Torn* PT Trittr**
KEY
Cpl
-------
The answers to these questions are closely related
to the objectives set for water used in each case, and
to the predictions of water quality conditions in tine
and space. Forecasting water quality characteristics
is the subject of water quality models. After calibra-
tion and validation, a model that reproduces reasonably
well the present conditions in the relevant parameters
can be used confidently to predict future water quality
in time and space.
If the growing demand for forecasting future water
quality conditions in some Latin American areas is to
be met, there will have to be a shift in emphasis in
the collection of water quality information. In most
cases, monitoring cannot continue to be single-purposed.
Authorities must be consulted on objectives and advised
on the possible outcomes. The cost of studies and their
monitoring needs should be compared with the total costs
involved and, when practicable, with the benefits to be
accrued. Only in this way can the most efficient use
of resources be determined.
Current programs
_ A number of institutions specifically concerned
with water pollution studies have been established and/
or expanded in recent years in Latin America. Among
the best-equipped and staffed are:
CETESB - Companhia Estadual de Tecnologia de Sanea-
mento BSsico e de Controle de Poluijao das
Xguas, Sao Paulo, Brazil
CIECCA - Centro de Investigacion y Entrenamiento
para el Control de la Calidad del Agua,
Secretarla de Recursos Hidraulicos, Mexico
DISCA - Departamento de Investigaciones sobre Con-
taminaciSn Ambiental, Direccion deMalario-
logla y Saneamiento Ambiental, Ministerio
de Sanidad y Asistencia Social, Venezuela
FEEMA - Fundapao Estadual de Engenharia do Meio
Ambiente, Rio de Janeiro, Brazil
(ex-XES, Instituto de Engenharia Sanitaria)
INCYTH - Instituto Nacional de Ciencia y Tecnica
Hidricas, Secretarla de Recursos Naturales
y Ambiente Humano, Buenos Aires, Argentina.
PAHO/CEPIS, with its own resources and as an ex-
ecuting agency for the UNDP, collaborates in studies
designed to predict future water quality and related
pollution control works necessary to meet the socio-
economic objectives related to specific situations.
One such study is the Guanabara Bay modelling ef-
fort and its monitoring requirements. This large body
of water bordering Rio de Janeiro, Brazil, is already
suffering the impacts of navigational, industrial and
municipal wastes. Many beaches are polluted; aesthetic
and ecological values are also threatened. Calibration
of salinity, bacteria and dissolved oxygen models im-
poses very definite monitoring requirements in time
(seasonal) and in space. What ia at stake is whether
to build a long trunk tunnel under the mountains to con-
vey the northern wastewater to the south and the ocean,
or to discharge it into the Bay. Figure 4 shows some
of the monitoring stations, together with the segmenta-
tion for the finite difference models.
Studies of this type are leading the way in Latin
America. In some cases, they are even changing the
institutional and legal mechanisms for water pollution
control. Such is the case in the Cauca River Valley in
Figure 4
200,000
& tUANABAHA BAY
MOOCL SEOMtNTS
AND SAMPLING STATIONS
Colombia where the outcome of the study, done with PAHO/
CEPIS assistance, showed that it would be advisable for
the Corporacifin del Valle del Cauca (CVC) to be solely
responsible for executing works related to pollution
control in the watershed. The CVC is now beginning to
implement the recommended program, including the peri-
odic monitoring of water quality levels. Knowing the
predicted levels, it is now possible to design an effi-
cient and economic monitoring system. This is also the
case in many other' situations where basin authorities
now exist or are being set up.
Outlook for the future
Latin America and the Caribbean is a dynamic and
heterogeneous region. Water quality monitoring is faced
with problems of resource availability and conflicting
priorities. The preceding pages have pointed out the
trends and needs, the shortcomings, and some past and
present activities in this field. The following remarks
are offered regarding prospects for the future:
1. Water quality monitoring is closely related to in-
formation in other fields, specially hydrological and
oceanographic data. It could be, in some instances,
advisable to compare water quality information require-
ments with the acquisition of data in those related
fields before embarking on ambitious programs,
2. Water quality forecasting will play an increasing-
ly important role in future water management in 4i£fer-
ent areas in each country and is presumably going to
require a large chunk of available monitoring resources.
6
8-1
-------
3. Well-staffed and equipped centers have already been
established or are being formed in several Latin Ameri-
can countries. These could be the nucleus for expanding
the monitoring and modelling studies which are so badly
needed itv the Sub-Region to shape local environmental
policies. PAHO's CEPIS can serve a-a a technical assis-
tance and research unit for these activities.
4. Differences in resources, their availability and
needs should be kept in mind when planning future pro-
grams and studies in this field.
Final Remarks
Although there is still much room for improvement,
we can be pleased with the enormous progress made in
both air and water quality monitoring in Latin America
and the Caribbean as compared to the information avail-
able only 25 years ago. The Governments' plans to ex-
pand these activities should also make us optimistic
about the future.
References
1. ORGANISATION FOR ECONOMIC CO-OPERATION AND DEVEL-
OPMENT. Methods of measuring air'pollution. O.E.C.D.
publication No. 17913, Paris, 1965
2. WORLD HEALTH ORGANIZATION. Air quality criteria
and guides for urban air pollutants. Report of a WHO
Expert Committee. Technical Report Series No. 50 6,
Geneva, 1972
3. - PAN AMERICAN HEALTH ORGANIZATION. Van American air
pollution monitoring network; report of results', June
1967-December 1970. Lima, CEPIS, 1971. Technical Se-
ries of the Department of Engineering and Environmental
Sciences No. 10
4. PAN AMERICAN HEALTH ORGANIZATION. Pan American air
pollution monitoring network; results obtained, «Tune
1967-December 1973, Preliminary report. Lima, CEPIS,
1974
5. HADDAD, Ricardo. "ContaminaciSn del aire. Situa-
cion Actual en la America Latina y el Caribe". Paper
presented at the Simposio sobre Ambiente, Salud y Desa-
rrollo en las Americas (OMS/OPS/CEPIS), Mexico, D.F.,
August 1974
6. QUISPE, L., et al, Institute de Salud Ocupacional,
Peru. "Informs del estudio de la eficiencia del metodo
acidimetrico para la determinacion de dioxido de azufre
en aire, usando el instrumento 'Tracor' de tubo permea-
ble, modelo manual". Unpublished paper, 1972
7. CASTAGNINO, Walter A. "Estado Actual de la Polu-
ciSn de Agua en America Latina". Paper presented at the
Simposio sobre Ambiente, SaludyDesarrollo en las Ame-
ricas (OMS/OPS/CEPIS), Mexico, D.F., August 1974
8. OFICINA SANITARIA PANAMERICANA. Calidad de aguas
en la auenaa del Plata. Preliminary report. 1969
9. VENEZUELA. Plan naoional de aproVechamiento de los
recursos hidrdulicos. Caracas, 1970
10. NACIONES UNIDAS. Comiaion Economica para America
Latina. Dos recursos hidr&ulieos de AmSrica Latina.
6 t. 1968
11. RIO DE JANEIRO. Instituto de Eugenharia Sanitaria
& Oficina Sanitaria Panamericana. Seminario internacio-
nal sobre control de aalidad de aguas de la cuenoa del
rto de ta Plata. Final report. 1968
12. BATTELLE. Planning, design and operation of com-
prehensive water quality monitoring systems with empha-
sis upon the needs of Latin American countries. Draft
final report prepared by Battelle Memorial Institute
for PAHO/CEPIS, 1971
13. PAN AMERICAN HEALTH ORGANIZATION. Ten-year health
plan for the Americas. Final report of the III Special
Meeting of Ministers of Health of the Americas. PAHO,
Official Document No. 118. Washington, D.C., 1973
7
8-1
-------
LARGE SCALE AIR POLLUTION MONITORING IN THE NETHERLANDS
T.Schneider
National Institute of Public Health, Bilthoven, the Netherlands
Over the last few years environmental monitoring has
become an important issue in the Netherlands. Scientif-
ic, technological and political thinking has been fo-
cussed on it, intruments and techniques have been de-
veloped.
Apart from the beautiful fields of tulips and other
spring flowers the Netherlands possess one of the
highest densities of population in the world together
with major industrial centres. Meteorological condi-
tions and wind direction in this part of the world are
extremely variable and so are pollution conditions. Not
only will the emission in a given air space vary from
day to day and hour to hour so will its chemical compo-
sition and place of origin. Complicating the matter is
the fact that this country's neighbours, Belgium and
West Germany, are also heavily industrialised. So, in
addition to the pollution generated internally, the
Netherlands must often cope with that coming across
international boundaries, from the industrial concen-
trations in the Ruhr area and northern Belgium.
In the late sixties a decision was taken to create
a national air pollution monitoring network for the
Netherlands to achieve comprehensive surveillance
throughout the country. It was designed to provide ul-
timately all the following information:
1. a reliable estimate of the quantity, composition and
origin - international as well as national - of the
pollution.
2. data for establishing trends in the degree of pollu-
tion from year to year and the influence of zoning
decisions, growth of industry, traflc and pollution
on these trends.
3. determinations of the effect of oountermeasures cm
these trends.
4. information for short-term warning purposes, to al-
low forecasting of undesirably high degrees of pol-
lution under unfavourable meteorological circumstan-
ces and predicting their spread over the country.
Given these goals the following criteria had to be met.
a. Flexibility in the structure of the measuring system.
Designing the system in 1970 it became soon obvious
that the key element had to be flexability, At that
time it was impossible to determine once and for all
the total number of monitoring stations and the com-
ponents to be measured. With time the outlook
ohanges and stations will have to be added or re-
moved. Also changes in composition and geographical
distribution of the pollution caused by measures
taken, make it necessary that changes within the
system can be performed -without undue costs.
b. Low cost of maintenance and servicing.
To obtain long series of accurate measurements it is
imperative that disturbances in the system as a
whole are rare. Because of the relative high cost of
maintenance in the field the number of times ser-
vicing has to be carried out especially at the moni-
toring stations had to be reduced to a minimum.
c. Real time system.
Conventional ways of gathering air pollution data
consisted of samples taken in bottles, transport to
the laboratory, measurement with laboratory instru-
ments and manually recording of the results. Here
the problem is different. It involved in fact an air
pollution inventory of a whole country not only in
the dimension of space, but also of time. It in-
volves charting geographical drift patterns, in-
cluding short-term and long-term trends. Hie situa-
tion in each region must be followed very olosely.
Alert situations have to be determined, and mes-
sages sent to provincial authorities and Industries
so that measures oan be taken on short notice. All
this called for virtual real time data collection
and processing.
Structure of the Monitoring Network
The network consists of 217 monitoring stations, of
whloh 109 form a regular base line grid oovering the
entire country. The.remaining stations cover special
topographical features, densely inhabited areas, large
industrial centres and national borders. Each station
is equipped to monitor, on a regular basis, one or more
polluting canpanents in the atmosphere and to transmit
the data to its regional centre. Sane of the stations
also monitor meteorological factors such as wind veloc-
ity and wind direction.
The national network is divided into 9 regional sub-
networks, each with its own computerised regional
measuring centre. These centres in turn are linked to ft
central computer in the National Institute of Publio
Health in Bilthoven. This central communication unit
controls the total system and stores and processes all
information of nationwide relevance.
In the aystem structure three levels of information are
available.
1. monitoring stations consisting of a measuring pole
1 8-3
-------
(air intake at 3.80 m), a housing with one or more
monitors and telemetric equipment. In the first
phase of operation the network is equipped with mon-
itors for SOg only. The capabilities of approxi-
mately 80 stations will be expanded to measure in
addition CO, NO, N02 and 0^. Each measuring instru-
ment is calibrated twice daily at two levels, autom-
atically, by remote signal from the computer in re-
gional measuring centre.
2. each Regional Measuring Centre houses a process com-
puter. This computer has the following functions:
a) input, preprocessing and presentation of informa-
tion received from the monitoring stations.
b) transmission of concentrated information to the
central communication unit in Bilthoven.
c) control, including calibration, of all monitoring
stations in each region.
d) control of peripheral equipment in the regional
centre.
At each region information is collected every min-
ute, minute values and hourly averages are calcu-
lated. Every 12 hours a command signal for remote
calibration of each monitor is initiated. A dialog
typewriter in the regional centre records all error
messages and a logging typewriter records air pol-
lution component and meteorological information.
Data may also be continuously monitored on an analog
recorder.
3. The National Measuring Centre. The highest level of
information within the network houses the central
communication unit, a process steering computer with
the following main functions:
a) timing and control of the entire measuring net-
work.
b) organization and verification of the communicat-
ion with the regional data centres.
c) control of peripheral devices.
d) processing, storage and presentation of informa-
tion received from the regional data centres.
The central communication unit autanatically col-
lects information from the regions and stores it on
magnetic tapes and discs. It calculates and presents
hourly averages on a line printer. It checks pollu-
tion levels in the regions. Threshold values are set
and monitors exceeding these values are automati-
cally indicated by lights on an illuminated map. All
these and additional information can be presented
on a CRT display, a logging typewriter, an analog
reoorder or punched papertape. A facility for minute
to minute check and control of the whole network is
possible. The information stored on magnetic tape is
processed off-line for statistical and model studies.
The continuously recorded information is also pro-
cessed to national and regional surveys each month,
every six months and every year to study trends in
pollution levels.
Data analysis
Due to the large number of questions that have to
be answered within the framework of the national net-
work several different techniques are used. Both clas-
sification, transformation, presentation of data and
data reduction are needed.
1. Geographical description.
As a basis for yearly presentation, percentiles (50
and 98) are calculated based on daily averages. The
percentiles as a function of the geographical loca-
tion are used for the calculation of iso-lines. Per-
centiles based on averages of hourly measurements
per wind direction sector are also calculated and
presented in the form of wind roses,
2. Trend evaluation.
Changes in the general level of pollution with time
as calculated from year to year is called the air
pollution trend. Elimination as far as is possible
of the meteorological influences on the general
level is important. Therefore classification of all
the data available according to wind direction, wind
speed, air temperature and stability classes is per-
formed.
3. Warning function.
Forecasting undesirably high degrees of pollution at
an early stage will be performed. By transformation
of the original hourly averages using an P-index the
existing atmospheric dispersion condition can be
described. Multiple discriminant analyses are also
used on an experimental basis.
Additional measurements
The measurements made at fixed stations of the
national network are complemented using mobile units.
These mobile units (cars and a light airplane) give for
a smaller area a more detailed measurement pattern in
space or time. Especially during periods of severe air
pollution, they play an important role in the determi-
nation of local high levels. Measurements with remote
sensing methods are a necessary addition to emission
inventories based on inquiries, fuel consumption analy-
sis and direct stack measurements With this equipment a
direct relation between emission and ambient air pollu-
tion levels can be determined. Recent developments
2 8-3
-------
within the mobile program are the direct measurement
of NO^ and SOg fluxes over larger distances and the
measurements of specific pollutant components such as
HF and separate hydrocarbons.
Apart from this mobile addition to the network, the
effect of air pollution on plant diseases Is studied on
30 stations of the national network. Especially indica-
tor plants, specific for one or more pollutants are
used. Among the indicators used are spinach as indica-
tor for ozone and SOg, Petunia as indicator for
ethylene, Urtica urens as indicator for PAN, tobacco as
indicator for ozone and grass and lettuce as indicators
for PAN. During last year especially ozone damage on
tobacco plants was found all over the country. Separate
from the research on plants grown in the outside air, a
number of small glass houses is used to study the
growth-and development of plants under comparable con-
ditions with and without filtered air. The pollutant
or combination of pollutants to be studied is filtered
out of the normal outside air as drawn through the
glass houses.
Short evaluation of the first experiences with the
nationwide air pollution network
After a first year of operation it is now possible
to give an evaluation of the results obtained so far.
One of the purposes of the network is to produce a map
of the geographical distribution of the pollutants over
the country. This function is performed quite well pro-
viding that the other components (nitrogen oxides and
ozone) will be installed. For the trend research only
preliminary results are produced. Research on an ade-
quate elimination of the meteorological influences
still has to be carried out. The development of a
warning function valid for the country as a whole and
for the different regions separately looks promising.
The density of the network allows us to indicate sudden
rises in pollutant concentrations at an early stage
when general levels are still low. The. study of the
transport of the pollutants has shown that additional
measurements in the third dimension are necessary to
determine the flow pattern from the source areas to the
different receptor regions in the Netherlands. The com-
bination of continuous measurements at th,e ground
stations and intermittent remote sensing flux measure-
ments from a light airplane and especially developped
mobile measuring units (with measurements during the
runs) give very satisfactory results.
The experience to date has confirmed the wisdom of the
flexible approach to the design of the monitoring
system. The automatic measuring instruments are per-
fected and an increasing number of polluting substances
and environmental parameters will be added to the
existing stations. Monitors will also have to be in-
stalled on towers to gather information in three dimen-
sions. Apart from an improvement in the sensors itself
the data analyses of such a lar^e number of mea*
surements is already advanced to a stage where better
use of the data can be made both by scientists as well
as by governmental, provincial and municipal authori-
ties. Long uninterupted series of reliable measurements
are needed and these can only be obtained with a system
that consists of sub-systems and units that are well
adapted to each other, a system with its maintenance
carried out by well trained techniciens. In the coming
years the scientific research in the field of air pol-
lution will certainly benefit from the wealth of data
to come forth from this nationwide monitoring system.
3
-------
A DESCRIPTION OF A TECHNIQUE FOR DEVELOPING AN OPTIMUM AIR POLLUTION
AND METEOROLOGICAL SAMPLING NETWORK IN URBAN REGIONS
Fred M. Vukovich
Research Triangle Institute
Research Triangle Park, N. C. 27709
Abstract
A research program whose purpose is to develop a
technique for optimization of air pollution and meteo-
rological sampling network selection in urban regions,
is described. The technique employs dynamic and
statistical procedures to determine the optimum network.
The research program is composed of two phases:
Phase I: Theoretical Phase and Phase II: Applied
Phase.
In the theoretical phase, the necessary dynamic
and statistical models were developed by which the
optimum network was determined. The models were
applied, and an optimum network was established for
St. Louis, Missouri, which is the site of the Regional
Air Pollution Study (RAPS) sponsored by the
Environmental Protection Agency.
In the applied phase, the optimum network
developed in Phase I was set up in St. Louis, and a
field test was initiated. This field test was inte-
grated with the RAPS program.
Introduction
The establishment of the air quality standards
has brought about the need to develop air quality and
meteorological sampling networks in urban areas. Such
networks have two primary purposes. The first is to
provide data for analysis of air pollution over the
urban region to determine if and where a particular
pollutant concentration exceeds the standard. This
information is required to implement control measures
to reduce the concentration of the particular pollutant
to an acceptable level. The second purpose is to pro-
vide a data base for both the short and long-term
prediction of the concentration of a particular pollu-
tant over the urban region. This information provides
a base for the development of long-term control
strategies. Diagnosis and prediction of air pollution
is particularly necessary now that the energy crisis
has made it probable that industries and power plants
will burn high sulfur coal.
Sweeney (1969) attempted to establish objective
guidelines for determining the number and location of
sampling stations in an optimum sampling network (OSN)
for urban areas purely by statistical means. He con-
cluded that the present state of art of statistical
theory does not provide a basis for a general solution
to the problem. The basis for this conclusion lies in
the fact that the air pollution distribution and
meteorological conditions are dynamic phenomena that
vary in space and time in an urban region and vary from
one urban region to the next. However, Sweeney indi-
cated that a solution specific for a-particular urban
region can be obtained if a representative and statis-
tically significant data set is available for the urban
region. This data set should be made up of observa-
tions of n number of pollutants and m number of meteo-
rological parameters. For most urban regions, n is
greater than or equal to three and m is at least one
(the wind distribution) but can be larger depending on
the degree of sophistication desired.
In order to acquire this data set, Sweeney sug-
gested that a high resolution sampling network be set
up in a given urban region for the collection of data
on pollution concentrations and meteorological condi-
tions over a period of years. Though this is
attractive, it is economically unfeasible. The
Regional Air Pollution Study (RAPS), being sponsored
by the Environmental Protection Agency (EPA), will cost
millions of dollars and does not yield the high reso-
lution data required by Sweeney. It is unreasonable to
ask each urban region to support such programs.
In order to ease the economic pressure, Sweeney
suggested that the urban region should be divided into
sections. A high resolution network should be set up
in one section and data collected over a number of years.
When sufficient data are obtained in the first section
of the urban region, the sampling network would be moved
to another, and data collection would commence and con-
tinue until sufficient data were obtained. This
iterative process would continue until the data set was
complete.
There is, however, a dilemma that arises when one
determines an OSN, based solely, or in part, on air
pollution data. Most, if not all, urban regions are
constantly changing. New industries are moving in, old
industries are moving out; that is, new point sources
move in and old point sources move out. It is important
to note that the location of the old and new point
sources are very often different. Furthermore, over
the years, rural roads, which often have insignificant
traffic, become major line sources with the establish-
ment of housing projects and accompanying commercial
areas near or along the road. Other types of urban
expansion result in similar changes. Urban expansion
can produce a major change in the air pollution distri-
bution in a five to ten year period. Any OSN based on
today'8 representative, air pollution data sets may
soon become suboptimum. This effect would prevent
Sweeney's suggested technique for obtaining a represen-
tative data set from being practical. In the inter-
vening time between collecting data in the first segment
of the urban region and collection of the data in the
last sector, major changes in source configurations may
have occurred, making the segmented data sets totally
independent of each other.
It is obvious for economic reasons, that the basis
of the OSN must be very stable over a relatively long
period of time. A technique to establish an OSN, on a
more stable basis than is provided by the air pollution
distribution, is described in this paper. The models
developed have been completed and an OSN for St. Louis,
Missouri has been created. St. Louis, Missouri was
chosen as a test case because the RAPS Program will
provide sufficient corollary data to evaluate the OSN.
The development of the models for this research project
were sponsored by the National Science Foundation. The
second phase of this research project, in which we are
testing and evaluating the OSN, is being jointly spon-
sored by the National Science Foundation and the
Environmental Protection Agency.
Theoretical Development
The conservation equation for any pollutant, i/i,
yields the exact nature of the variables necessary to
completely specify the distribution of the pollutant,
i.e.,
|i » _ v • Vi|i + V • (KV1(0 + 11) £ k i/) * + S .
at r n n n
The first two terms on the right represent the meteoro-
logical influence on the distribution of ij>. The first
term is the advective term where V is the wind vector,
and the second term is the diffusion term where K is
1
8-4
-------
the exchange coefficient. The third term is the
chemical reaction term in which k is the reaction
constant,
-------
the wind field derived from the OSN, the observed
pollution concentrations from the OSN, and the source
emissions. This model is generalized so that these
parameters can be specified for different urban regions.
Figure 1 is a flow diagram for the entire
development concept. At the top of the figure is the
primitive equation model used to establish a repre-
sentative data set for the wind field in a given urban
region. In the case of St. Louis, Missouri, the data
set was composed of forty-eight (48) case studies.
The data are input for an algorithm that determines a
regression model that, in turn, adequately character-
izes the wind field for the given urban region. The
data set and the resultant regression model are then
combined with the site selection algorithm to yield
the OSN. The theoretical development phase of the
modeling ends here; that is, in the figure the
theoretical development phase is the region above the
lower dashed line.
£
Figure 1. Flow diagram for development concept.
The applied phase is the region below the upper
dashed line. For this phase, it is assumed that the
OSN for the wind field has been set up in the urban
region. The wind and air pollution data are obtained
at the stations in the OSN. The wind data are used to
develop the wind distribution through the same
regression model determined in the theoretical develop-
ment phase. The wind distribution, the observed air
pollution data, and the emissions inventory, are input
data for the OVAM, which subsequently yield the air
pollution distribution.
Optimum Sampling Network
Figure 2 is a plot of the stations determined for
the OSN in St. Louis, Missouri. The criterion that
established the OSN was the variance function, which
yields a measure of the predictability over the area,
as well as the correlation and sums of squares of
deviations between observed and predicted values.
Figure 2. Station plot for the OSN for St. Louis,
Missouri
The OSN contains nineteen (19) stations. However,
optimum subclasses were also determined. For Instance,
if an urban region could afford to set up only an
eighteen (18) station network immediately, with the
idea of adding the nineteenth station later, then the
subclass that would be created is that which includes
all the stations in Figure 2 except the one marked
"nineteen". If seventeen stations are required, delete
nineteen and eighteen; for sixteen stations, delete
nineteen, eighteen, and seventeen, etc. Thirteen is
the smallest subset which can be used because the
regression model used for St. Louis has thirteen terms.
A comparison of the mean correlations between
observed and predicted values for the nineteen station
OSN and its subclasses is given in Table 1. This table
suggests that the thirteen station subclass for the OSN
should yield poor results, and that considerable
improvement is had with the subclass with fourteen or
more stations in it.
Table 1. Correlation coefficient (R) between
predicted and observed wind speeds
versus station number where a thirteen
term model was used to describe the
wind field.
Number of Stations in
Optimum Network
13
14
15
16
17
18
19
It was economically unfeasible to set up a
nlneteen-station network in St. Louis; to reduce costs
of the field program required to validate these
results, many of the RAPS or city/county air pollution
stations were used as part of the OSN when locations
coincided or nearly coincided. This tractable version
of the OSN is designated as OSN* and the station plot
is given in Figure 3. Once again the numbers next to
the stations are used to designate the stations to be
deleted if optimum subclasses are desired.
R
0.50
0.63
0.65
0.65
0.66
0.67
0.67
3
8-4
-------
References
Sweeney, H.( "Development of Sampling Guidelines for
Regional an
-------
DESIGN AND REDESIGN OF AN AIR SAMPLING NETWORK
James H. Price
Gary K. Tannahill
Duane J. Johnson
Andy C. Wheatley
Roger R. Wallis
Texas Air Control Board
Austin, Texas 78758
Summary
The central consideration in designing or redesigning
any network should be whether it meets the needs that
justify its existence. This paper describes an approach
that keeps overall air pollution"control needs the cen-
tral criteria for sampling network design and redesign.
This approach is to define the needs to justify the
network's existence, translate them into questions that
can be answered by air quality data, design a network
that will collect data to answer the questions, and
analyze the data to get the answers.
Introduction
An adequate air quality evaluation system is an essent-
ial part of any reasonable and effective air pollution
control program. Information that can be provided
only by such a system is required for planning and
decisions within every phase of a control agency.
Such information must, of course, be comprehensive,
accurate, and timely. Care must be taken to insure
that a balanced control program is achieved and that
surveillance efforts are not carried beyond the level
needed to satisfy basic objectives. However, it must
always be kept in mind that control regulations based
on inaccurate or inadequate knowledge of pollutant
concentrations may be needlessly costly because of
unnecessary controls or because of failure to protect
public health and welfare.
The heart of an adequate air quality evaluation system
is its air sampling network. The central consideration
in designing or redesigning any network should be whether
it meets the needs that Justify its existence. This
paper describes an approach that keeps overall air
pollution control needs the central criteria for samp-
ling networks design and redesign. In outline, that
approach is to define the needs to Justify the network's
existence, translate them into questions that can be
answered by air quality data, design a network that
will collect data to answer the questions, and analyze
the data to get the answers to the questions. A net-
work redesign requires, in addition, sampling site
information that is adequate to determine what questions
the existing network can answer.
Definition of Needs
The first step is to define clearly the needs that
Justify the existence of the network. These needs
can include determining whether air quality meets
ambient air quality standards, evaluating the effect-
iveness of air pollution controls, providing inform-
ation necessary for designing control strategies,
detecting potentially dangerous increases in pollutant
levels under air stagnation conditions, providing in-
formation needed for land-use and transportation plan-
ning,discovering and defining new pollution problems,
supporting epidemiological surveys of health effects
associated with measured pollutant levels, satisfy-
ing public demand for measurement of air quality
even when other critera would not Justify air
sampling, and providing data needed for air pollution,
research.
Translation Into Questions
It is tempting to skip the next step, translating
the needs into specific questions that can be
answered by sampling data. However, it is very
important not to skip this step. Without a clear
set of questions for the data to answer, the normal
pattern in network design is to buy the number of
samplers the budget will allow, allocate these
samplers to areas in porportion to the population
and pollution associated with areas, and find
"good" sites for them. Unfortunately, the data
produced frequently fail to answer any questions
except what the air quality is at the sampling site.
There may be alternatives to a network for answering
the questions. Using an emissions inventory alone
or in combination with diffusion modeling or using
source sampling should also be considered. Conduct-
ing a short-term air quality survey, using either
fixed stations or a mobile van or even an aircraft
for sampling, may be preferable to establishing an .
air sampling network. A mobile van, equipped to
monitor carbon monoxide, can in a few days' driving
on city streets locate areas of high carbon monoxide
concentrations far more efficiently than a network
of stationary samplers. In its function, the van
is analogous to the use of sulfation plates for
locating areas with worse sulfur dioxide problems
than other areas have. A short-term survey using
a number of fixed sampling locations for a specified
time, followed by evaluation of the data to determine
whether any permanent sampling site should be estab-
lished in the area, is an example of network design
and redesign. The preestablished termination date in
a short-term survey gives a flexibility in design of
the subsequent network that needs to be sought de-
liberately in redesign of existing networks.
Network Design
Many factors are important in deciding on the instru-
ments, the sampling sites,and the sampling schedules
that will most efficiently answer the questions asked.
A useful principle is to choose the simplest and
most reliable sampling method that will provide
the data needed to answer the pertinent questions.
As an example, Table 1 lists the capabilities of
sulfation plates, gas bubblers, and continuous
monitors to answer questions about the sulfur dioxide
concentrations in air. The sulfation plate method,
1
8-5
-------
Table 1
Capabilities of Sulfur Dioxide (SO2) Measurement Methods to Answer Questions About Air Quality
V = Can Answer the Question, X => Cannot Answer the Question
SULFATION GAS CONTINUOUS
QUESTION PLATES BUBBLERS IKSTRUMENTS
Is measurable SO2 present? y s/ s/
Which location (within an area with
similar temperature and humidity) 1 i
has the maximum average SO2 concen- »
tration?
Is there .a problem with SOg levels! T
>/ \/
Is the average level of SO2 increas- / / /
ing or decreasing from year to year? V V V
Is a site in compliance with annual x \J \J
and twenty-four-hour SOg standards?
Is a site in compliance with short- x x \f
time period SOg standards?
Is the concentration of SOg rising x x \f
toward the emergency epidsode level?
which requires no electric power, only minimally
skilled field personnel, and very inexpensive field
equipment, is fully adequate to determine whether
or not there is measurable SOg. Sulfation plates,
because they are inexpensive, can be used in large
numbers in finding the areas of maximum average sulfur
dioxide concentrations. However, since measurements
made with sulfation plates are relative, not quanti-
tative, the plates are not useful for determining com-
pliance with standards. The averaging time of sulfur
dioxide standards determines whether gas bubblers or
continuous instruments are necessary to determine
compliance. Because of the wide difference in cost
among sulfation plates, bubblers, and continuous
sulfur dioxide monitors, it is obvious that it is
worthwhile to consider which of the methods will
provide the information actually needed to answer
the pertinent air pollution questions. In the choice
of sampling methods there are many additional consider-
ations: whether the sampling can be passive and
qualitative or needs to be active and quantitative,
whether temporal or spatial variation in pollutant
concentration needs to be known or both do, whether
the data are needed immediately or a historical record
is adequate, whether or not wind speed and direction
data are also needed for analyzing the data, what the
averaging time of the measurements should be, what the
interval between samples should be, what the alter-
native instruments cost to buy, operate, and support,
and whether or not the technical and financial resources
are available to operate the system.l
Site Selection
Actual selection of sampling sites is an Important step.
Lengthy discussions of site selection criteria already
exist,2 The height above ground level, wind-flow
obstructions, security from vandalism, permanence, and
alectric power availability all deserve consideration
In the selection process. One approach to evaluating
potential sites has been to ask whether valid, "repre-
sentative" data can be collected at the site. The
concept of "representative" meaning "typical of a
broad area" is very difficult to define and is a
potential source of misinterpretation and conflict.
Better criteria would be,first, that valid data be
collected reliably and, second, that the data could
answer one of the specific questions developed
directly from agency needs. It should be clear
that the questions about the validity of data from
a site are separate from the issue of whether or
not the data answer the pertinent questions that
are developed from the definition of agency needs.
Site Documentation and Classification
The documentation of air sampling sites is the logical
final stage in establishing a network. The documen-
tation. procedure at the Texas Air Control Board
starts with visiting each site and photographing
the sampler from the four main compass points to
show its immediate location and the broad area
surrounding it. While at the site, the person
doing the documentation prepares a diagram of
the roof or area on which the sampler sits. This
sketch includes distances to wind-flow obstructions
and their elevations. We also prepare a map of the
area within one mile of the sampler, showing the
land uses and significant sources and sinks of air
pollutants. Performing this documentation at the
time each sampler is deployed is the logical com-
pletion of the design and deployment phase. Class-
ification of the sites based on this documentation
serves as a check that all the sampling sites satisfy
the criteria used in selecting them. As time passes
and data accumulate, it is easy to lose sight of
the questions the data from each site were supposed
to answer. Site documentatioB and classification
are a way to avoid this pitfall.
Adequate site documentation and classification
are essential to analyzing collected data. Data
analysis serves two functions: First, it provides
the answers that can be given to the questions used
as a basis for the sites. Second, it identifies the
sites from which data fail to answer the questions
the sites were established to answer. The problem
may be that the data from the site cannot answer
the questions at all or simply that not enough
data have been collected.
2
6-5
-------
Network Redesign
The data analysis results provide an obvious feedback
to network design. If the questions asked as the
basis for the network remain unanswered, either they
were unanswerable or the network needs to be redesign-
ed to answer them. If they are answered, the redesign
needs to consider whether or not fewer samplers could
provide the answers or, if the answers are reasonably
certain to stay the same, whether or not the sampling
for that specific purpose should be discontinued.
In summary, these are three prerequisites to network
redesign: (1) to have the program objectives clearly
defined and translated into specific questions that air
quality data can help answer, (2) to have the sampling
sites documented and classified and (3) to have the
data analyzed to see whether or not the sites produce
the answers the site classifications indicate they
can.
Continuous Monitors
It may be inconsistent for this section to be in
this paper, but we are including it because of the
strong tendency, which we share, to think of continu-
ous monitors as distinct from other air sampling
devices. Continuous monitors present operational,
support, and quality assurance problems that require
specialized consideration. Because of the attention
devoted to these problems, it is easy to forget that
the principles discussed in this paper apply whether
the sampling device is passive, intermittent, or
continuous. One corollary to these principles is that
instruments that produce valid data are the heart of
a continuous monitoring system. Following this pre-
cept in our continuous monitoring work, we have put
the emphasis on instrumentation and quality assurance.3
Only now that the basic system has been developed,
deployed, and refined are we working on automation and
real-time data telemetry. We have been satisfying
the most pressing need for real-time data by having
station operators telephone raw data to headquarters
for immediate processing during air stagnation advisories.
Texas' Experience
We have no illusions that the idealized approach pre-
sented in this paper can always be followed rigorously,
but we are finding that it is a useful model even when
itemized budgets or Environmental Protection Agency
formulas for numbers of samplers dictate the number
and type of samplers that are available.
This approach to network design was developed piece by
piece. Early in Texas' network development, sampler
sites were frequently chosen to provide data go answer
specific, clearly defined questions. In 197^, it became
obvious that comprehensive redesign of the network
was necessary. Since the purpose of many of the
sites had never been adequately documented, we began
with a comprehensive documentation effort for all
97-State, 90-local, and 13-EPA sampling sites that
were active in Texas. Initially, classification
included only these questions: whether or not a
site could produce valid data, whether or not there
are sources near the sampler, and what the land use
or uses within a mile of the sampler are.1* Classification
is being extended to include the questions that each
site is expected to answer.
Most of the bubblers and hi-vols in the Texas network
are operated by volunteers or employees of local govern-
ments. In practice, there has often been reluctance
to have a site discontinued, so whenever a volunteer
operator drops out of the program, we take that oppor-
tunity to reevaluate the sites he has operated.
We are satisfied that the Texas network is adequate
to answer questions about compliance with standards.
Since our X-ray fluorescence apparatus began operation
in 1973, our coverage of toxic heavy elements has
been sufficient to identify public exposure to un-
usually high concentrations. With developing capa-
bilities in using wind data and in data analysis in
general, our ability to identify the sources of
pollution is improving. A major area of dissatis-
faction for us is our lack of objective criteria
for differentiating between the air pollution in
major urban areas and relatively-clean smaller cities.
Ozone and total suspended particulate data are in-
adequate criteria for this purpose. We are ex-
amining the potential for using light scattering
coefficient and respirable particulate, nitrate, and
certain heavy elements as measures of urban pollution.
Hopefully, these parameters will correlate better
with subjective public perception of pollution.
Conclusion
The central consideration in designing or redesign-
ing a network should be whether it meets the needs
that justify its existence. It is easy for an air
pollution control agency to begin by establishing
sampler placement guidelines and move rapidly to
the point that network growth and the percentage
of data retrieved are the criteria for success in
an air quality evaluation program. We have found
that it is possible to avoid this pitfall by adopt-
ing the approach described in this paper: defining
the needs that justify the network's existence,
translating these needs into specific questions
that can be answered by air quality data, designing
a network to answer these questions, and analyzing
the data to get the answers and evaluate the network
design. Site documentation and classification are
essential features of this approach, particularly
if the process begins with an existing network that
was established without adequate documentation of
the sites and the purposes each site was to serve.
References
1. G. B. Morgan, "Assessment of Ambient Air Quality
in the United States" in Design of Environ-
mental Information Systems, R.A. Deininger, ed.,
Ann Arbor Science Publishing, 1971*.
L. J. Brasser, "Questions Related to the
Selection of Air Pollution Measuring
Systems", in Design of Environmental Inform-
ation Systems, R. A. Deininger, ed., Ann
Arbor Science Publisher, 197^.
2. Draft, "Air Quality Monitoring Site Description
Guidance", OAOFS Ho. 1.2-019, Office of Air
Quality Planning and Standards, U.S. Environ-
mental Protection Agency, 197^.
Draft, "Guidance for Air Quality Monitoring
Network Design and Instrument Siting",
0A0PS No. 1.2-012, Office of Air Quality
Planning and Standards, U.S. Environmental
Protection Agency, January, 197^-
W. R. Ott, "Development of Criteria for Siting
Air Monitoring Stations", Sixty-Eigth Annual
Meeting of the Air Pollution Control Association,
June, 1975-
3. D. J. Johnson, "Texas ambient air quality continuous
3
8-5
-------
monitoring network", Proceedings of the First
Annual Symposium on Air Pollution Control in
the Southwest. November 5-7¦> 1973. Texas A & M
University, College Station, Texas, pp. 525-
5U9 -
R. W. Mann, "Continuous ambient air quality evalu-
ation determinations from digitally logged
monitoring data", ibid, pp. 550-568.
Project Connie Operations Manual, Texas Air Control
Board, Austin, Texas (February 15, 191b).
Air Quality Surveillance Procedures Manual. Texas
Air Control Board, June, 1975.
D. J. Johnson, R. L. Richardson, and R. R. Wallis,
"Assuring High Quality Data from Ambient Air
Monitoring networks", Sixty-eighth Annual
Meeting of the Air Pollution Control Association,
June, 1975-
Manual for Operation of Texas Air Sampling Network
(TASN), Texas Air Control Board, Austin, Texas.
August, 1975.
-------
THE REGIONAL AIR MONITORING SYSTEM
ST. LOUIS, MISSOURI U.S.A.
Dr. R. Lee Myers
Rockwell Air Monitoring Center
St. Louis, Missouri
James A. Reagan
U.S. Environmental Protection Agency
St. Louis, Missouri
The Regional Air Monitoring System (RAMS) is the
ground based air quality measurement network of RAPS,
the Regional Air Pollution Study. This 25 station
network is the most sophisticated air monitoring
system currently operational. The central purpose of
RAPS is to produce for a single urban area enough
information on all processes that determine the con-
centrations of air pollutants so that they can be
described in a system of mathematical models.l This
objective leads directly to the requirement that there
be generated an extensive base of air quality measure-
ments against which calculated values may be compared.
The RAMS stations also serve as experimental facili-
ties for other investigations. In the following
material, an extensive description of the network,
analyzers, data acquisition, quality control, and
operations is given.
Network Design
Although early planning included 40-50 stations,
economic considerations required reconsideration of a
"minimum network." In Figure 1 the twenty-five sta-
tibns are shown in relation to the design. Stations
are designated from 101 at the center increasing by
unit increments clockwise from north around each ring
going outward ending with 125 as the westerly station.
Originally designed to be in four rings of 6, 6, 8,
and 4 stations at distances from the center of 4, 10,
20, and 40 kilometers respectively as shown, the
FIGURE
,22
final rings have average radii of 5, 11, 20, and 44 km.
Elevations of the stations are fairly uniform averag-
ing 154+ 23 m MSL.
Site Criteria
Clustering the stations at the center of the net-
work provides an approximate concentration proportional
to gradient of the levels of air pollution.2 The four
outer stations were located as near 90° azimuths as
feasible and not close to any significant sources.
The remaining sites are located at least 1 km from any
significant sources, i.e., smokestack or major traffic
artery. No obstructions on the horizon generally
higher than ten percent were permitted. Topographic
features capable of local wind flow influence were
avoided.
Siting Experience
Acceptable areas for each site were designated.
A thorough examination of the area was made for poten-
tial sites and owners contacted to ascertain accepta-
bility. More problems were encountered than expected.
The 30 m towers installed at 17 stations violated
height restrictions in every community. Neighborhood
acceptance was important in all urban and suburban
locations. Power was unavailable in large parts of
the rural area. Many compromises were forced due to
trees on the horizon rising higher than ten percent.
Accommodation to farmers plowing plans resulted in
drawn out negotiations over access roads and exact
station placement. Zoning permits and variances nearly
proved impossible to obtain in a few cases even with
neighborhood acceptance. Several excellent sites could
not be obtained due to the owners development plans.
Park Departments and School Boards generally do not
want other facilities on their land. An average of
four candidate sites were reviewed before obtaining one
for each station. If site technical and economic
criteria are met, a small public relations program
with neighborhood associations, area council repre-
sentatives, and immediately adjacent property owners
will obtain community acceptance and patience will
then secure needed permits.
A network of stations, five in the City of
St. Louis, five in the County of St. Louis and one in
Illinois were in place before the installation of the
RAMS. No station was installed in close proximity to
one of this network's stations thereby avoiding dupli-
cation. Instead the data from these stations provide
greater resolution of the regional gradients.
Shelter and Support Assemblies
Each station consists of a shelter, tower, fence,
sensors and support equipment. The shelter is 4.88 m
(16') wide x 3.25 m (10'8") deep x 2.74 m (9') high, .
constructed of interlocking galvanized steel panels,
the exteriors of which are painted with a siliconized
l
8-6
-------
polyester paint. An end door provides norma 1 access
with a side door for alternate or emergency exit. The
flat roof is covered with expanded metal grating and
will support both instruments and personnel. Access
to the roof is fay a portable ladder stored in the sta-
tion. The shelter rests on a structural steel base,
equipped with lifting lugs to facilitate removal and
relocation. The shipping weight slightly in excess of
3000 kg is not much, but the width is a problem in
interstate shipment. Electrical service of 100
ampere, 120/240 volt, 3 wire single phase service is
provided. The heater/air conditioner has a cooling
capacity of 11,000 kg-cal and a heating capacity of
8,500 kg-cal. Four automatic fire extinguishers are
mounted on the overhead. Freon 1301 is used as the
extinguishing agent for its nontoxic properties.
Support equipment is detailed below.
Compressor - Dryer Assembly
A single stage oilless compressor supplies pres-
surized air tp the heatless dryer equipped with two
identical desiccant chambers packed with molecular
sieve and charcoal. Each chamber acts as a desiccant
and -hydrocarbon absorber. The chambers alternate with
one being regenerated while the other dries the air.
In add4t1on to the drying, ozone, sulfur dioxide,
hydrogen sulfide and oxides of nitrogen are removed.
A continuous stream of ultra-pure air is thus provided
the analyzers by this system.
Gas Sampling Manifold
Ambient air enters the station through an in-
verted glass funnel attached to a three inch inside
diameter glass tubing. The intake and tubing are
covered with a metal shield for rock protection. The
air passes into the station and is taken from the hori-
zontal manifold by Teflon tubing inserted into the
center of the manifold at a 45° angle to prevent con-
densate from being drawn into the analyzer. A Teflon
filter, 1/4" mesh, is placed over the bottom of the
intake funnel to prevent Insects from entering. Glass
traps are provided at 90° bends in the tubing. Each
analyzer intake has a 47 mm, lOOyTeflon filter inline.
Calibration System
A system for calibration of the gaseous analy-
zers providing a zero and upscale concentrations of
various gases has been incorporated. Mass flowmeters
and a mixing chamber allow dynamic dilution in ratios
as high as 1000:1 of calibration gases. Zero air
supplied by the compressor-dryer assembly is addi-
tionally purified by passage through a catalytic oxi-
dizer for removal of carbon monoxide and Is used as a
diluent and for determining instrumental zero.
Calibration Panel. This consists of flow con-
trollers, flow indicators, constant temperature
Haake Water Bath, SO* and H„S permeation tube assembly,
mass flowmeter, and Controls. It provides the regula-
tion capability for accurate mixing and dilutions of
test gases.
Sulfur Dioxide. An S02 permeation tube, certi-
fied by the National Bureau of Standards (NBS), 1s
maintained In the water bath at a constant temperature
of 30 + ,01°C. A monitored zero air flow 1s main-
tained~over 1t at all times. A thermistor sensor
monitors the temperature at all times. Periodic
checks against an NBS Standard Reference Material
(SRM) S02 permeation tube are made.
Hydrogen Sulfide. An HpS permeation tube Is
periodically placed in the water bath for multipoint
calibrations. A monitored flow of nitrogen is main-
tained over it( at all times when in situ.
Ozone. A 23 cm ultraviolet inside a triangular
slotted sleeve is used to generate ozone. An 8 mm
diameter quartz tube allows zero air to pass through
to produce a high concentration of ozone. The quanti-
ty produced is regulated by varying the position of
the slotted sleeve. Flow rate of the zero air is
monitored. See Oxides of Nitrogen for calibration
procedure.
Nitric Oxide. A cylinder of NO in N? 1s main-
tained in each station. The concentration is nomi-
nally 100 ppm, but is periodically checked by an NBS
SRM NO cylinder.
Oxides of Nitrogen. Nitric oxide from the cylin-
der 1s blended with ozone from the generator by gas
titration to form nitrogen dioxide. The depression in
NO is equal to the concentration of ozone. Conversion
efficiency of the N0? to NO 1s also determined for the
N0X measurements.
Carbon Monoxide. A cylinder of CO in nitrogen is
maintained in each station. The concentration is
nominally 5000 ppm, but 1s periodically checked against
an NBS cylinder.
Methane. A nominal mixture of 2500 ppm of CH- is
1n the cylinder containing the CO. The concentration
1s periodically checked against an NBS cylinder. Zero
air is prepared the same as CO zero air.
Operation. The calibration panel 1s used to
achieve an approximate dilution of 500:1. Solenoid
valves are activated to select the desired calibration
gas. Either manual selection or automated sequencing
under computer control can be made. A second selection
can switch an analyzer from ambient to calibration gas.
Zero or span gas 1s supplied to the sensor for a suffi-
ciently long time to achieve a stable response and then
switched again. Independent calibrations are periodi-
cally made of the mass flowmeters.
Status Panel
Certain conditions and limits are monitored on
other assemblies within the station. Some status
checks are connected to a display to alert personnel
in a station, such as flameout. Other conditions
such as door-open are set in the status bit array for
transmission with the data. The status panel func-
tions as the interface to the data acquisition system
for the many additional checks 1n the system.
Exhaust Manifolds
The air sample exhaust manifold provides the exit
from the gas sample manifold as well as the exhaust
provisions for the outputs from the hydrogen genera-
tors and chromatographs. The calibration exhaust
manifold provides the exit from the calibration mani-
fold as well as the exhaust provisions for the outputs
from the H„S and S0? permeation tubes and the ozone
generator. An acid gas trap in line contains four
acid gas filters which absorb contaminants that could
bias the monitoring systems, before expelling the
exhaust gases into the ambient air. C?H. is removed
by a heater which oxidizes it to COg and H^O.
2
8-6
-------
Hydrogen Generator
For safety purposes two Milton Roy Mark IV hydro-
gen generators are in each station. These were se-
lected in lieu of hydrogen cylinders which present a
tremendous safety and logistics problem. Each genera-
tor has a two-week water supply and is provisioned
with an automatic shut off in case of flameout. The
longevity of the cells has not been good, repair time
has been drawn out and the cost of repairs has been
expensive.
Particle Manifold
A special intake manifold was designed for parti-
cle sampling. The intake is of special design to
insure the entry without loss of particles from an air
stream for delivery to the analyzer. The conical
cover on the intake is designed to simulate the parti-
cle size censoring qualities of the standard hi-vol
shelter. This cutoff is nominally 200 p, but work by
other investigators indicates l6o p may be more accu-
rate. The manifold itself is of aluminum and by its
conductive property does not build an electrostatic
charge. A laminar flow of the air stream is main-
tained within the manifold.
Gas Closet
Compressed gas cylinders to support system opera-
tion and calibration are housed in a separate enclo-
sure within the station. This enclosure is continu-
ously purged with excess air from the heater/air
conditioner, eliminating buildup of potentially ex-
plosive ethylene and toxic carbon monoxide gases
should leaks occur. The air enters the enclosure at
the top, flows over the cylinder and is vented outside
the shelter at the bottom.
Pump Box
All pumps were removed from the gas analyzers
prior to installation 1n the system thereby eliminat-
ing the noise and failure problem characteristic of
them. Instead a vacuum pump and a compressor-dryer
assembly were installed in each station. These were
placed in an Insulated wooden box at one end of the
shelter. By Isolating the mechanically noisy equip-
ment, the station meets the design criteria of 75 dBA.
In addition the top of the box serves as an excellent
work or test bench.
Telephone
Two voice grade circuits are provided at each
station. One is a common business telephone for
communication between station personnel and the
central computer or repair facility. The second
is a shared data line described under Telecommunica-
tions below.
Tower
A tower is provided at each statfon. Seventeen
stations have 30 m towers while stations 108, 110, 114,
115, 116, 117, 118, and 121 have 10 m towers. Each
tower serves as an Instrument mounting stand. The
10 m towers are mounted on a concrete pad and brack-
eted to the side of the shelter at the roof. Each of
the outer stations has a guyed, 30 m tower mounted
separate from the shelter on a concrete pad. The
other thirteen 30 m towers are self-standing. They
are three sided with a base of approximately 3 m width
and a top of 25 cm width. Each 1s mounted on a block
3
containing about 15 m of concrete. A lightning rod
is installed on every tower. Each tower is in the
northern sector of the station plot.
Fence .
Every station is surrounded by a chain link fence
two meters high topped by three strands of barbwire.
The gate is positioned in line with the main shelter
entrance to facilitate installation and removal of
large equipment. Redwood strips or green plastic
covered links have been provided where needed to con-
form to neighborhood standards.
Instrumentation
The mix of analyzers, sensors, and sample collec-
tors varies between stations. During periods of
intensive investigations, other special instruments
are operated at the stations. The measurement tech-
nique for each parameter is described below.
Wind Speed
Manufacturer: Meteorology Research, Inc.
Model: 1022S
Measurement method: Three stainless steel cups
rotate a shaft attached to a chopper disk. A light
beam from a light emitting diode passes through the
chopper disk as it rotates and impinges on a photocell
transistor. The output signal is a sine wave whose
frequency is directly proportional to the speed of
rotation.
Effective measurement height: 10 m or 30 m
depending on tower height.
Range: 0-22 m/s (0-50 m.p.h.)
Starting threshold: 0.22 m/s (0.5 m.p.h.)
Response distance: 1.5 m (63% recovery)
Accuracy: + 0.17 m/s or 1% whichever is greater.
Experience: Winds in excess of 25 m/s occur in
St. Louis. These sensors are sensitive and have a low
threshold. There is a tradeoff between ruggedness and
sensitivity and these have required extensive main-
tenance.
Installations: All stations.
Wind Direction
Manufacturer: Meteorology Research, Inc.
Model: 1022D
Measurement method: A gauged two-section poten-
tiometer, of 20k ohms per section, is mounted with a
solid connector to the shaft inside the sensor main
housing, A non-friction keeper pin on the pot-mounted
blade arm maintains a sensor-to-pot orientation. The
coils are aligned to be in phase with one another
while the wipers are positioned 180° apart.
Effective measurement height: 10 m or 30 m
depending on tower height.
Range: 0 - 540°
Starting threshold: 0.3 m/s (0.7 m.p.h.)
Delay distance: 1.1 m (50% recovery)
Damping ratio: 0.4 at 10° angle of attack
Linearity: + 0.5% of full scale
Resolution: 0.1%
Experience: Extensive maintenance of these in-
struments has been required. Failures have been
primarily due to one of three causes: Loose shaft to
pot connection, pot failure, or loss of Internal pot
orientation. Interstation data comparison usually
flags failures fairly quickly.
Installations: All stations.
3
8-6
-------
Turbulence
Dew Point
Manufacturer: R. M. Young Company
Model: 27002 Gill UVW Anemometer
Measurement method: On each of three orthogonal
mounting arms a hellcoid propeller is mounted with
sufficient separation between the propellers to have
no significant effect upon each other. The propeller
drives a miniature d.c. tachometer generator providing
an analog voltage directly proportional to wind speed.
Both forward and reverse air flow are measured. The
propeller responds only to that component of the wind
which is parallel with its axis of rotation.
Effective measurement height: 30 m
Range: 0-22 m/s (0-50 m.p.h.)
Threshold: 0.22 m/s (0.5 m.p.h.)
Response: 1 revolution per 30 cm of wind
Experience: Data from these sensors is taken
every half second. Operationally the station data
tapes must be retrieved and processed to recover the
data. Due to the wind conditions in St. Louis, pro-
pellers and sensor arms have been separated and lost
from the sensSrs. Since the instrument is relatively
sensitive it is installed for operational periods,
then dismounted and stored during periods of nonopera-
tion. As a rule these sensors have been operated only
during^the intensive field investigations of RAPS.
Installations: Stations 105, 107, 109, 111, 113.
Temperature
Manufacturer: Meteorology Research, Inc.
Model: 840-2
Measurement method: A dual thermistor and resis-
tor network comprises the sensor. The circuit is
designed to provide a linear resistance change with
air temperature change. The sensor is inside a glass
reflective tube positioned downward such that it sees
a cone with a contained angle less than 90°, in order
to minimize long-wave radiation effects from the
ground or snow surface. The sensor 1s aspirated with
an air flow of some 5 m/s.
Effective measurement height: 5 m
Range: -20°C - +50°C
Experience: This sensor has consistently per-
formed well. No unusual or extraordinary maintenance
problems have been encountered. Linearity 1s not
known and the absolute accuracy is not known.
Installations: All stations.
Vertical Temperature Gradient
Manufacturer: Meteorology Research, Inc.
Model: 840-1/840-2
Measurement method: The sensor is the same as
that in the temperature instrument. A second sensor
1s installed on the tower and is connected by a sepa-
rate bridge to the ambient temperature sensor. The
sensors are electronically balanced to a zero differ-
ential In an ice bath. When mounted the signal condi-
tioning card matches the output from the two sensors
to produce an output proportional to the temperature
differential.
Effective measurement height: 5 m vs. 30 m
Range: -5°C to +5°C
Experiences: When installed the instruments
appear to function well. No determination of absolute
accuracy has been made. Electronic drift or board
failure appear to occur after extended operation.
There was indication the signal conditioning board may
have been temperature sensitive. Inversions with a
vertical temperature gradient in excess of +5°C occur
In the St. Louis area.
Installations: Stations 101, 102, 104, 105, 106,
107, 109, 111, 112, 113, 122, and 123.
4
Manufacturer: EG & G International, Inc.
Model: 8{30
Measurement method: The Instrument is an automa-
tic, optically sensed, thermoelectrically cooler, con-
densation dew point hygrometer. A highly polished
copper disk plated with gold to render an inert surface
has a precision thermistor embedded 1n it. The ther-
mistor is part of an electrical bridge circuit. The
thermoelectric element has two surfaces, one of which
can be made cool by rejecting heat through the other
surface upon the application of an electric current.
A neon lamp in the sensor directs light to the mirror
surface where it is reflected at an angle to a photo
conductive cell. The resistance of the cell, which
changes in relation to the amount of light striking it,
forms part of a bridge circuit where the electrical
output is indicative of the reflectance of the mirror
surface. The output of the bridge 1s amplified and
directed to the photoelectric element which is bonded
to the base of the mirror. The circuit is adjusted by
the automatic balance control which decreases the
cooling effect proportionately until equilibrium exists
in a constant amount of condensate.
Effective measurement height: 2 m
Range: -40°C to +50°C
Dew point depression: Maximum depression capa-
bility is 45°C at an ambient temperature of 27°C.
Cooling rate: 2°C/s, maximum
Resolution: 0.25°C, nominal
Accuracy: 1°C, nominal
Experience: These instruments have provided mean-
ingful data 30-40% of the time. The urban atmosphere
of the St. Louis area causes the mirrors to become con-
taminated with both soluble and Insoluble salts. More
cleaning of the mirror 1s required. An In-line filter
has reduced the rate at which a mirror malfunctions.
In the humid climate of that region the common failure
mode is the formation of a puddle of water on the mir-
ror surface.
Installations: All stations.
Barometric Pressure
Manufacturer: Meteorology Research, Inc.
Model: 751
Measurement method: A transducer, incorporating
several diaphragms of NISpan C alloy which are mechan-
ically linked with Invar fittings, 1s connected to a
precision potentiometer. Changes in resistance occur
proportionally to changes in pressure.
Effective measurement height: 2 cm
Range: 914 - 1067 mb (27.0 - 31.5 inches H„)
Linearity: + 0.3% 9
Sensitivity: Less than 0.2%
Hysteresis: less than 0.2S!
Experience: When matched against a properly
balanced signal conditioning card, these instruments
do not quite match the manufacturer's conversion equa-
tion. Therefore, a dynamic procedure employing an NBS
traceable barometer was used to establish response
curves for each Instrument. Local station differences
can be seen 1n the pressure patterns.
Installations: Stations 101, 109, 112, 122, 123,
124, and 125.
Ozone
Manufacturer: Monitor Labs, Inc.
Model: 841 OA
Measurement method: This sensor determines ozone
concentration by mixing a fixed air flow wfth a fixed
ethylene flow, and measuring any light output as a
result of this with a photomultlpller tube. If any
8-6
-------
ozone is present, a blue light, X max = 435 nm, Is
emitted. An optical modulator is placed between the
reaction cell and the detector so that the only out-
put signal is light emitted in the reaction cell.
Effective measurement height: 4 m
Range: Selectable, 0-.20 ppm is normal with
0.50 ppm used during the oxidant season.
Nominal span value: 0.10 ppm
Zero and span drift: 2%/day allowable
Minimum detection limit: 0.005 ppm
Experience: This analyzer has operated very well
since startup. The noise of the instrument is margin-
al acceptable, but other than that the instrument has
little or no drift problems.
Installations: All stations.
Oxides of Nitrogen
Manufacturer: Monitor Labs, Inc.
Model: 8440
Measurement method: This analyzer determines
nitric oxide concentration by mixing a controlled flow
of sample air with ozone, and measuring any light out-
put as a result of this with a photomultiplier tube.
If any nitric oxide 1s present, a deep red light near
the infrared region is produced. If the ozone concen-
tration is fixed, the light output is linearly related
to NO concentration. A dual channel system provides
simultaneous measurement based on the above reaction.
The incoming air stream is split with one half meas-
ured for NO concentration. The other half passes
across a low temperature converter consisting of moly-
bedenum metal which selectively converts nitrogen
dioxide to nitric oxide which is measured as above.
An optical modulator is placed between the reaction
cell and the detector so that the only output signal
is light emitted in the reaction cell.
Effective measurement height: 4 m
Range: Selectable, 0-.50ppm is used
Nominal span value: 0.20 ppm
Zero and span drift: 22/day allowable
Minimum detection limit: 0.005 ppm
Experience: Considerable difficulty was ini-
tially experienced at the beginning. -Numerous, fail-
ure occurred, but engineering modifications eliminated
these problems. Now the instrument operates for
extended periods of time without problem. By means
of the ozone-nitric oxide gas phase titration per-
formed, converter efficiency is monitored.
Installations: All stations.
Hydrocarbons and Carbon Monoxide
Manufacturer: Beckman Instruments, Inc.
Model: 6800
Measurement method: This sensor 1s based on the
detection of small ionization current flows between
two polarized electrodes, the burner jet and collector,
1n the vicinity of a hydrogen flame. The magnitude
of the current depends on the concentration of carbon
atoms In the carrier gas. Total hydrocarbons are
measured by injection of sample air to the flame ioni-
zation detector (FID) and the subsequent response.
Additionally carrier air sweeps sample gas into an
Initial stripper column which 1s used to separate the
fast-elutlng methane and carbon monoxide from other
components. These two undifferentiated compounds
are passed to another column where they are resolved
and then passed through a catalytic converter where
the carbon monoxide is converted to methane. The
separated compounds pass through the FID with the
methane giving rise to a signal proportional to
concentration and carbon monoxide converted to methane
generating a similar response. Peak detection and
hold circuits are Included In the Instrument
electronics.
Cycle time: 5 minutes
Effective measurement height: 4 m
Ranges: CO, 0-10 ppm or 0-50 ppm, auto ranged
CH«, 0-10 ppm or 0-50 ppm, auto ranged
THC, 0-10 ppm or 0-50 ppm, auto ranged
Span drift: 2%/day allowable
Experience: This instrument performs well when
it Is operative. Typical of most flame analyzer sys-
tems it is sensitive to many conditions. On a
continuing basis these sensors have required extensive
support. Valve failures, column replacements, timing
adjustments have all been required on one instrument
or another.
Installations: All stations.
Total Sulfur
Manufacturer: Meloy Laboratories, Inc.
Model: SA-185
Measurement method: This analyzer 1s based on the
chemlluminescence of sulfur compounds produced In a
hydrogen rich flame. The light emitted, X max = 394nm,
is proportional to approximately the square of the
sulfur concentration. The sensor employs a narrow
band pass filter which filters the light before it is
passed to a photomultiplier tube. This flame photo-
metric detector sensor provides continuous measure-
ment of gaseous sulfur species.
Effective measurement height: 4 m
.Range: 0-.2ppmor 0-1.0ppm, auto ranged
Zero and span drift: 2%/day allowable
Experience: Except for minor maintenance problems
associated with the flame used in the sensor, this
Instrument has operated with little or no attention for
extended periods. No silver scrubber for hydrogen
sulfide was installed in the RAMS.
Installations: Stations 102, 107, 109, 110, 111,
112, 117, 118, 119, 123, 124 and 125.
Sulfur Dioxide and Total Sulfur
Manufacturer: Tracor, Inc.
Model: 270 HA
Measurement method: This analyzer incorporates
the flame photometric detection method described for
the Meloy SA 185 above. Sample air Is passed directly
to the FPD for the determination of total sulfur. A
separate aliquot of air is injected Into a separation
column which resolves the sulfur dioxide and hydrogen
sulfide components. These are then passed sequentially
through the FPD for measurement. Peak detection and
hold capability is included in the electronics.
Effective measurement height: 4 m
Range: TS, 0-.2 ppm and 0-1.00 ppm, simultaneous
SO,, 0-.2 ppm and 0-1.0 ppm, simultaneous
H?5, 0-1 ppm and 0-1.0 ppm, simultaneous
Span drift: 2%/day allowable
Minimum detection limit: 0.005 ppm
Experience: This instrument was difficult to
bring into operation. Once the instrument stabilizes
it has operated for extended periods with minor adjust-
ments. It is very sensitive to power outages and re-
quires a long period to restabillze. Instrument drift
has quite often exceeded the operating limit, but not
by too much. For an instrument as complicated as it
1s, this one has performed quite well.
Installations: Stations 101, 103, 104, 105, 106,
108, 113, 114, 115, 116, 120, 121, and 122.
bscat
Manufacturer: Meteorology Research, Inc.
Model: 1561
Measurement method:. An air sample is drawn into
5
8-6
-------
the interior of the Instrument and illuminated by
light from an opal glass diffuser which is Illumi-
nated by a continuously operating quartz-halogen in-
candescent lamp. A reference phototube maintains a
constant light intensity. Both phototubes have opti-
cal filters to correct the spectral response of the
instrument to that of the human eye. The photomulti-
plier tube views a small solid angle parallel to the
diffuser surface. It measures the scattering portion
of the extinction coefficient of an air sample. The
scatter coefficient obtained is the sum of the aerosol
scatter and Rayleigh scatter from the atmospheric
gases.
Effective measurement height: 5 m
Range: 0.1 to 10 x 10"^ m"'
Accuracy: ,+ 10% of scale
Flow rate: .0024 mVs (5 cfm)
Experience: At the start of this program, the
1561 was a relatively new instrument. To date they
have required extensive maintenance. Common problems
have included lamp failure and electronic drift, other
nonspecific problems requiring factory readjustment
have been extensive. The instrument has a fast res-
ponse and appears to correlate with human observation.
An additional heating coll 1s available from the manu-
facturer for removal of the effects of humidity on the
observed, scattering coefficient. Although Installed
in the network, little use has been made of the
heaters.
Installations: All stations.
Solar Radiation
Manufacturer: The Eppley Laboratory, Inc.
Model: Precision Spectral Pyranometer
Normal Incidence Pyrheliometer
Precision Infrared Radiometer
Measurement method: A wire round thermopile,
nonwavelength selective and coated with a black lac-
quer 1s the basic sensing element. The pyranometer
Incorporates two hemispherical globes, an Inner one
of quartz and an external one of a wavelength selec-
tive filter, which measure global (direct plus diffuse)
radiation 1n the range 300-3000 nm. The pyrheliometer
looks only at direct radiation by the placement of a
sensor at the end of brass tube with an aperture to
length ratio of 1:10. The aperture is covered by a
quartz window and selective filters are placed between
the sun and aperture. An equatorial mount 1s the base
for the pyrheliometer and it rotates to track the sun
as it moves across the horizon. A nine position fil-
ter wheel rotates 1n front of the aperture presenting
a dark (covered) position, open position, and filters
selective at 395, 475, 530, 570, 630, 695, and 780 nm.
The Infrared radiometer (pyrgeometer) 1s like the
pyranometer except the single globe is a hemisphere of
KRS5. It has a vacuum deposited filter on the Inner
surface and a weather protective coating on the outer
surface which allows 1t to measure from 3.0-50 jan. A
thermistor resistance circuit measures the net outward
flux of the sensor itself.
Effective measurement height: 5 m
Range: 0-4 cal cm'/min. -
Sensitivity: 5 mv per cal cm /mln.
Temperature dependence: + 1%
Experience: The pyranometers perform quite well
when properly Installed. There are three different
types used 1n RAMS. The outer globes are one of three
types: quartz, 395 nm selective or 695 selective fil-
ters. The pyrheUometers track the sun and rotate to
a new filter each minute. The equatorial mounts drift
with time and require constant realignment and the
signal cable controlling the wheel positioning is
large and does not twist gracefully. The pyrgeometer
hemispheres have corroded badly 1n the St. Louis atmos-
phere. Dirt settles onto the globes and filters of all
6
these Instruments requiring constant cleaning.
Installations: Stations 103, 104, 108, 114, 118,
122.
Gas Bag Samplers
Manufacturer: Xonics, Inc.
Model: None
Measurement method: This sampler provides an
integrated sample of ambient air in a Teflon bag for
subsequent analysis. The bag is made of TEF Teflon.
It holds 100 liters and is approximately 76 cm wide by
91 cm long. It 1s hung from one end inside an airtight
metal box approximately 97 cm high by 81 cm deep by
10 cm wide. Air is drawn Into and out of the box until
sampling begins at which time a valve switches over and
draws air from the box allowing ambient air to enter
the bag. The flow can be controlled from 0-1200 cc/m1n
by a regulator.
Effective measurement height: 4 m
Range: 0-100 liters
Experience: The adjustable flow rate is used to
take 100 liter samples to correspond to the three hour
period, 6:00 a.m. - 9:00 a.m., of the United States
hydrocarbon standard. The second box in a station is
set to take 100 liters In two hours usually from
4:00 p.m. - 6:00 p.m. Bags are removed from a station
and sample air is analyzed, usually for hydrocarbon
compounds from C,-C1Q. Normal operation results in
60-80% of the bags containing sufficient sample air for
analysis. Bag leaks, failures of the boxes to sample,
kinks In sample lines, flow regulators going out of
adjustment have caused many samples to be lost or
Inadequate. This system Incorporates the first
Installation of this type as a routine sampler.
Installations: Two at each station.
Total Suspended Particulates
Manufacturer: Sierra Instruments
Model: 305
Measurement method: A standard high volume
sampler with a carbon brush motor draws ambient air
through a 20 x 25 cm glass fiber filter. The flow
rate is nominally 0.02 m3/s. and is continuously
measured by a thermal conductivity based anemometer.
Presample weight and postsample weight are made 1n a
humidity controlled atmosphere.
Effective measurement height: 4 m
Experience: These h1-vols have functioned as
expected. With the exception of missed sampling
commands, the samplers have operated very satisfacto-
rily. Component determinations of nitrate and sulfate
are routinely made on the TSP.
Installations: Two each at stations 103, 105,106,
108, 112, 115, 118, 120, 122, and 124.
Data Acquisition System
As shown 1n Figure 2, each remote station has a
rather extensive data acquisition system as one of its
main components. At the heart of the data acquisition
system 1s the Digital Equipment Corporation PDP-8m
mini-computer. The PDP-8 exists with the following
options: power fall and automatic restart, a real time
clock, the extended arithmetic element and two asyn-
chronous Interfaces. In addition to the normal 8k, 12
bit words of memory which came with the PDP-8, another
8k were supplied. In normal operation the data acquisi-
tion system communicates its data 1n real time to the
central computer facility. However, due to a variety
of problems which do happen such as failure of the
telephone lines or failure of the central computer it-
self, back-up storage was provided at each of the re-
mote stations. This storage 1s an Industry-compatible
8-6
-------
KSR - 33
Teletype
n
Oigi tat
Equr p mem
Corporalioi
PDP-8*
Xlncom
Mode) 35SO
32 Channel
)iff. Input
A/0
Converter
Pertec Mode) 7820-9
tnduiiry Campat i ble
flag Tape
and
Cont rol
j Te I ecommonic
Dec ima1
Oi sp lay
st rument<
and
Spec i a
Device1
FIGURE 2
Remote Station Data System Configuration - Block Diagram
magnetic tape, 9 channel, 800 bits per inch, Pertec
Magnetic Tape Unit, model 7830-9. To get the analog
data into the system it was necessary to install a
multiplexer, analog-digital converter system. The
unit chosen for this application was a Xincom Model
3150 with the following characteristics: input + 5
volts, through-put rate 20 kHz, 48 channels fully
differential, conversion resolution 12 bits. To pro-
vide other important information about the system's
performance, digital input bits are provided. Ninety-
six total digital inputs are available of which about
80 are used. This allows the system to measure and
record such information as flow out-of-limit, air
pressure out-of-1imits, and associated support charac-
teristics, such as the position of all the valves in
the system. To provide for the automatic unattended
operation, it was also necessary to add digital out-
put commands. Forty-eight commands are available for
use in controlling the sequencing of valves for the
automatic zero and span function-, computer controlled
range changing for several of the instruments; and
miscellaneous functions such as modem control. Also
included in the data acquisition system is a Teletype
for displaying data and communication with the system
for the operator. A six digit, digital display is
present for displaying time or the value of any analog
channel in the system.
The software provided with the system is very
flexible in nature and allows all instruments to
function independently of all others. For example a
variable averaging time, nominally one minute, is
selected for each of the parameters in that station's
startup. While the constraints of the RAMS network
are such that data are usually recorded at one-minute
intervals, it is possible for any instrument to record
at more or less frequent intervals than one minute.
The datum recorded for each interval is the average of
data points taken every half second during the record-
ing Interval. In addition to providing for the acqui-
sition of data, the program also has features built
Into it,which independently for each instrument,
sequence through the requisite valves for the automa-
tic zero and span. The software provides for various
operator functions, storage of data on magnetic tapes,
and communications with the central computer facility.
Telecommunications
In a real time computerized system such as RAMS,
it is important that the telecommunications network be
designed to work efficiently, economically, and reli-
ably. As was shown earlier in Figure 1, four leased
telephone lines run out from the Central Computer
Facility to different parts of the RAMS network.
These lines are a configuration of four wire half du-
plex with multi-drop access. The actual communication
is done with Novation modems running with Bell 202
type compatibility. Communication rates are 1200 baud
in both directions using ASCII character formats. In
addition to the parity function provided by ASCII, each
transmission by the remote or the central includes a
check sum for greater redundancy in error detection.
While the telephone lines are unconditioned the choice
of equipment in protocol has allowed us to achieve a
bit error rate on the order of 2 in 10'. This is a
typical value. However, as is well known, in regions
of intense electrical storms the bit error rate during
those periods rises much higher.
Central Computer Facility
The Central Computer Facility of RAMS is the real
time focal point for all data and operations. Data are
received from each station at one-minute intervals.
The central is the place to do all real time limit
checking, flagging of errors, and providing a central
dispatch for all the field operations. As will be seen
in the following section the hardware and software have
been integrated in a fully comprehensive system for
real time treatment of the RAMS data.
The central computer, Figure 3, for RAMS is actu-
ally two computers; hardware interleaved and running
in tandem. The first computer is a small PDP-11/05
with 16k, 16 bits words of memory, a Teletype, and four
asynchronous interfaces. The 11/05 is hooked through
a Unibus window to the larger PDP-11/40 with 40k words
of memory, and the majority of the system peripherals
including discs, tapes, printer-plotter, line printer,
card reader and cathode ray tube and operator's console.
As will be seen in later discussions the PDP-11/05 is
in charge of real time data handling and as such is the
master of the system. The PDP-11/40 is used primarily
to do background processing, but with the stipulation
that its peripherals are used by the PDP-11/05 in real
time on an as required basis. Basically the perfor-
mance characteristics of the peripherals are as
follows: each disc unit has 1.2 million-16 bit words
of storage. The tape drives are nine track, 45 ips,
800 bpi with the exception of one drive which is 7
track at multiple density. The line printer prints a
full 132 columns, 300 lines per minute minimum print
speed. The printer-plotter is an electrostatic version
which when operated as a printer will print in excess
of 4000 lines per minute and as a plotter can generate
a two dimensional or three dimensional plot in about
three seconds, providng the computations have been done
in advance. The card reader operates at about 300
cards per minute. The cathode ray tube display (ADDS
880) is a standard unit with keyboard.
Software for RAMS can be divided into two logical
parts; the real time software and the background data
processing. There are several functions which are per-
formed 1n real time under software direction. The main
function is the polling task in which the central se-
lectively retrieves the data from each of the stations
in the network once each minute and stores the data on
tape. As the data are being stored on tape they are
also trapped and displayed on the cathode ray tube as
required by the operator one station at a time 1n addi-
tion to the time each station was last successfully
contacted. The second major function of the real time
aspects of the system is to send out messages to the
remote stations for the purposes of altering the
7
8-6
-------
TELEPHONE LINES
TU10
MAG TAPE
TUIO
MAG TAPE
MAG TAPE
CONTROLLER
MAG TAPE
CONTROLLER
GOULD 4800
PLOTTER
AND
CONTROLLER
LA30
DECwrlter
LC
CONTR
11
OLLER
4 MODEMS
4 DL11
ASYNCHROUS
LINE
INTERFACE
PDP-11/05
AND
BUS WINDOW
KSR-33
TTY
UN I BUS
lpii
LINE PRINTER
AND
CONTROLLER
CRT 1
CARD READER
AND
CONTROLLER
PDP-11/40
COMPUTER
BOOTSTRAP
32K CORE
HARDWARE FLOATING POINT
MEMORY MANAGEMENT
REAL TIME CLOCK
DL11
ASYNCHRONOUS
INTERFACE
ADDS
CRT
TERMINAL
FIGURE 3
Central Computer System - Block Diagram
Instruments. This might be to initiate a calibration
change, a range change, the averaging time, or other
similar functions. The resident real time monitor
assigned to POP-11/05 also has the responsibility of
arbitrarily deciding which of the two processors get
any of the peripherals at what time and also works
to control the memory management option of the 11/40.
The background programs which run on the PDP-11/40
are more extensive 1n nature than those required for
just the polling process. Since the RAMS generates
in excess of one-million data points per day, the
programs required to process it must run in a rela-
tively small amount of time and also work efficiently
to summarize the data where 1t can be helpful to the
persons responsible for operating the network. By
far the most important background program for RAMS is
a program called TAPGEN. The purpose of TAPGEN 1s to
validate data. It takes the data from the remote
zero and span calibrations done each day and updates
the computer's calibration files. The raw data from
the stations are then read 1n through the program
and updated calibration constants are applied to
generate data 1n engineering units. At this time all
system status words are also checked to assure proper
validity. For instance if the vacuum pumps have
failed in the station then the data would be invali-
dated because there would be improper sample flow
through the instruments. If the valves were left In
a wrong position this would also be cause for Invali-
dation. The program, TAPGEN computes all validated
data points for one-m1nute values from each of the
sensors and writes a second level or a Level II
tape. On a separate pass It reads the validated
one-minute averages and computes hourly averages
for Inclusions Into a third file on the tape, the
first file being calibration values. Several other
utility programs are also routinely run to generate
frequency distributions of wind speed and wind
direction, to provide performance summaries for the
drift of the Instruments, the telecommunications
performance, and to generate hourly averages of the
parameters for immediate inspection by the field
personnel.
Operations
Since the RAMS is a completely automated system,
the operations and maintenance philosophies used have
been designed to take complete advantage of the system.
RAMS is run on a 7 day a week, 24 hour a day basis.
The maximum amount of activity occurs during the day
shift, but there are also personnel available to Work
and on duty for second and third shifts. The com-
puter room 1s the key part,of the 24 hour a day
operation. All dispatching is done from 1t in the
case of immediately detected errors. Field operations
are divided among preventive maintenance, corrective
maintenance, and multipoint calibrations. During
normal operation each of the remote stations 1s
visited once every three days for the preventive
maintenance functions, such as changing filters,
replenishing water supplies, cleaning the mirrors of
dew point sensor and other similar functions.
Corrective maintenance Is performed by dispatching
a qualified Individual to a remote station when ah
error is detected. The error can either be de-
tected 1n real time from the raw data coming in from
the station or can be a system status error. This
would occur In the case of a compressor failure or an
Instrument flameout or any other error which gener-
ates an immediate status error. The second way of
dispatching for corrective maintenance 1s through the
dally generated computer outputs which summarize the
zero and span drift of each of the Instruments. This
often indicates errors and problems which are hot
obvious from the 1 mediate status readings. Each sta-
tion Is visited every five weeks to perform a 5-p©int
calibration on the Instruments. This is an addition
to the normal zero and span function performed on a
8
-------
daily basis. Occasionally the 5 point calibration
will also indicate components which are failing or
perhaps which even have failed but do not reflect
in the normal day to day characteristics of the
systems operation. A fully equipped repair facility
is also maintained to support the field operations.
In this facility sophisticated test equipment and
calibration racks are maintained to permit the major
overhaul of any piece of equipment in the RAMS
network.
References
1. P. W. Allen, "Regional Air Pollution Study - An
Overview," Paper No. 73-21 presented at the 66th
Annual Meeting APCA, Chicao, Illinois,
June 24-28, 1973.
2. F. Pooler, Jr., "Network Requirements for the
St. Louis Regional Air Pollution Study," J. Air
Pollution Control Assocation 24:3 (1974).
9
8-6
-------
mraoHMBrau. cpiLirr arcmLLiya n nrcowau
KrB.SRI SOKWASTl 30K3ANT0
National Institute of Health, Research and Development
Ministry of HMlth
Jakarta. Indonaaia
WWWtt**
V* are all aware that nature has tat Halted capa-
city to neutralise the ever inoreaalng amount and types
of pollutanta emitted to the environment.
Indonaaia aa ona of tha developing countries suf-
fers from a rather different nature of pollution com-
pared to thoaa vail developed and industrialized coun-
trlaa. Similar to elsewhere In tha developing ooun -
tries tiie major oauae of pollution in Indonesia la
poor hygiene and aanltatloa due to ladk of sanitary
faellitlaa and community Ignorance. This oondltion be-
oomes worse everyday dna to the rapid growth of popu -
lation ( 2»3% par year at present ) whlah consequently
increases the overorowdedaeaa and tha demand on -water
supply and tha need of disposing vast*. Besides a high
rata of population growth, Indonesia la also facing tha
problem of unbalanoed distribution of population, since
alnost 6tf> of ita population lives on Jaya which cob-
prlaea only 7f> of tha country's total land area.
So acme extent development would help tha country
fight againat pollution through tha improvement of ao-
oio-economio condition of tha country aad tha people.
Both the government and tha ooamunity would be able to
have better and adequate aafa water aupply and could
afford to have sanitary and aafa waata disposal.
Tha majority (89$) of tha population at present
live in rural areaa. Thia shows that tha development
trend will for some daoadaa be eaphaaised on agrlcul -
ture and other rural developnent related to it. Bnvl -
ronnantal pollution whlah is associated to these ttg-
ricultural activities is pollution caused by peaticl-
daa reaidue aad part of tha fertiliser washed sway by
tha irrigation water.
On tha other hand, rapid urbanisation la also
threatening big cities such as Jakarta, Surabaya, Me dan
and Pjung Pan dang. Tha increasing uae of motoricad ve-
hicles aad the uncontrolled waste of the growing number
of induatriaa is the usual result from this urban!
aation. it present the water pollution in big oitiea ia
mostly biological is nature originating from domestic
waata. However wa may not underestimate the potential
sourcea of both water and air pollution mentioned
earlier.
QOTBHHMfflT POLICY AHD ACTIONS ,
Since 1970 there has been growing concern of
the government about thase pollution problems. Several
institutions carried out surreys and investigation* to
dataot tha natura and degree of environmental pollution
In 1972, a National Committee on tha Environment m
established to faoilltata tha coordination and coopena-
tion among agenolea in coping with tha environmental
problems. This interdepartmental oonaittea has aa ita
main taaka and fuaotions: to make an inventory of pro-
blems, tha formulation of policy aad tha maaagamant of
programs at tha national level in tha environmental
field.
A joint minis trial oomdasion on Pesticide has
been eatabliahad in Indonaaia in order to control the
uae of pestloidaa, either for agricultural, health,
domestic or othar purpoaea. Tha Center for Blomedloal
Studiaa of the Katlonal Institute of Health Hasearch aad
Development ia equipped with aophiatioated instruments
such aa the Oaa chromatography, Spectrophotometer eto,
to analyse tha typaa of pesticides which pollute tha
environment, Bvery provinoe haa baea facilitated with
simple laboratory for this paatioida control.
Tha Center for Koologioal Studiaa of the above
mentioned institution haa also a Division called tha
Division of Physical Environment which among others is
in charge of the Studiaa of Baviromantal Pollution .
Tha ministry of mining haa also a Research Xnsti -
tute which has been equipped with a complete and sophis-
ticated laboratory to dataot among others marina pollu-
tion due to oil spill or aeoidanta happened to tankers.
Tha National Research Institute for water resources de-
velopment has been carrying out monitoring on the phy -
aloal conditions of big rivers all over Indonaaia.
Tha Institute of Koology of tha University of Padjadja-
ran, Bandung haa bean oanyiag out the Environmental
Impact Asaeasmaat on tha development projects. It ia
alao Interested in tha urban and rural ecoaysterns.
Tha National Institute of Ooaanology has beea atudyiag
the biologioal quality of water along tha northern coas-
tal area of Java.
Quite a number of workshops, seminara and aympoaia
vara oarried out. Lam, regulations and standards in
relation to the Environmental Pollution are being pre-
pared,
Tha above mentioned aotivitles prove that tha Go-
vernment of Indonesia is quite oonoemed with the oon-
trol of environmental pollution.
However to carry out complete monitoring syatem
would be premature at present considering the limited
facilities aad qualified man-power to do thla. Monito-
ring is expensive aad it will only be dona when there
ia urgent needa of doing thla. Jakarta for Instance has
a program to do tha monitoring on the quality as well
aa tha quantity of ita water bodiea.
POPULATION GROWTH AJTD DENSITY.
Tha population of Indonesia Is currently growing
by more than 2.4 million per year. Alternative pro-
jections for tha year 2001 indicate that the population
will be between 200 aad 280 million depending on how
rapidly fertility is reduoed. The bast estimate Is
about 220 million, almost twloe what it ia today. The
doubling power results from tha fact that 53 f> of tha
population ia under 20 years old aad has a long repro-
ductive period ahead.
1
8-7
-------
table I. Population 1b Indonesia by region,
1930 - 2001
(In Billions)
Region
1930
1961
1971
2001
Java
41.7
a.9
76.1
133.8
Sumatera
8.3
15.7
20.8
47.6
Kalimantan
2.1
4.1
5.2
10.3
Sulawesi
4.2
7.2
8.5
15.0
Other islands
4.4
7.1
8.6
16.1
Indonesia
60.7
97.0
119.2
222.8
Source : Population sensuees of 1930, 1961 and 1971 and
population projection for the year 2001.
The expected density for Java will not be loss
than 1000 persons par sq. ka la the year 2000. In con-
trast, the density on Kalimantan ia estimated only
24 persona par sq. km. *)
"Java in tha year 2000 will be an "island city*1 which
will present a mixture of sesriurban-semirural linea-
ment, " said Dr. Sumitro our Minister of Kesearcin
tfa axe facing a aeriooa problem of housing and
lack of sanitary facilities which will make the condi-
tion wont than what we have bean faolng at present.
Therefore the government gives high priority for the
housing and water supply pro gran both for the urban and
rural areas.
In tha Second Fire Year National Development Plan the
following baaic policies on population and settlements
have been adopted,
a. To reduoe the rate of population growth through fa-
mily planning programs within an overallpopulation
policy.
b. To improve tha balanoe of population distribution
between Java and the outer islands, through transmi-
gration program and better distribution of develop-
ment activities throughout the country.
c. To reduce the rural-urban Migration and to avoid
concentration of population in big oitiea through
tha creation of new growth oentres and the intensi-
fication of rural development.
d. To organise people living in remote areas through
resettlement programs and to raise their sociocultu-
ral level.
MAJOR CAP3BS OP KmroWBHTAL POLLPTIOH
In Indonesia the major causes of environmental po-
llution are s
1. Poor hygiene and sanitation.
2. Community igoomnoe.
3. Marine pollution by oil spills, tanker accidents and
offshore oil drillings.
4. The use of pesticides and fertilisers for agricultur-
al purposes.
Besides these, there are also growing concerns on
pollution as a result of :
1. Uncontrolled Industrial waste disposal.
2. Increasing number of motoxised vehicles,
3. Erosion resulting Aram felling the forest.
The improvement of hygiene and sanitation should
start with public health education in order to avoid
any constraint in carrying out this expensive program.
Hygiene and sanitation are also very much influenced
by the sodo-eoonoalo condition of the people.
The Straits of Malaka and the Java sea have been
polluted by domestic and industrial wastes, oil spills,
tanker accidents and the off-shore oil drillings.These
may create problems to fishery in the region. While
fishes are one of the Important sources of protein in
Indonesian diet.
Pesticides are used extensively to protect the. agri-
cultural produots and to control disease vectors. In
Indonesia.DDT is still used because of its effectiveness
and comparatively low cost. In 1974 £ 1,400 tons ddt. 7.5%
was used for malaria control, while the total use of
DDT from 1959 up to 1974 for this purpose ~ 42,500 ton
DDT 75$. Other Insecticides which were used were Diel -
drin, Arootine, Pyrettarin, and Malathion.
There are 19 insecticides, 2 weed killer pestioides,
4 rodenticides, and 3 moluscicides which are currently
used by Pest Control Agencies in Indonesia.
A considerable part of the fertilizeriused for paddy
fields will be washed away by the irrigation water, it
flows into the rivers, estuaries or other bodies of wa-
ter. This fertiliser will cause entrophioation of the
above mentioned water bodies and supports the growth of
unwanted speoies euoh as the water plant Budhornia rem-
ains*. This unwanted plant will grow very fast oovering
major areas of water bodies,clogging the gates and dis-
rupting tiie beneficial uses of these water bodies.
Presently many industries am located in high den-
sity areas or mixed with residential areas, in the city.
Regulation controlling the industrial waste ia absent
or if there is any it could not be properly enforced.
Although the number of industries is relatively small
the industrial waste problem should be dealt with seri-
ously.
In big cities most of the existing streets can not
accomodate the ever increasing mo tori Bed vehicles re -
suiting in traffic jams which in turn increase the air
and noise pollution. Widening the streets will force ma-
ny people living along the streets to saorifioe and also
will invite more private care. Hon polluting public
transportation such as "becak" has been banned for many
streets and replaced with the highly polluting transpor-
tation such as "bemo", "helioak" or "minlcar". Ho conve-
nient side walks or slow transportation lines for pedes-
trians or for cyclists are available.
The growing population pressure in rural oommuni -
ties have resulted In greater land hunger, which In turn
has Increased the tendency for people to invade lands
which were previously not allocated for settlements and
agriculture. Consequently, forests have been seriously
affected. As a result, floods are beooming more frequent
and more disastrous. Discharge of springs located in pre-
viously forested mountainous areas have decreased,along
with the clearing of forests, which also results in the
aooeleration of the rate of erosion in oertain areas,
turning rivers into muddy streams during rainy seasons
and some into completely dry oreeks during the dry sea-
sons.
MAffljtlTODB OP THE PROBLEMS
Water pollution
Several surveys on the quality of water at rivers
and atom water canals in Jakarta showed that the BOD
was ranging between 100 - 200 ppm and the D.O, many
times dropped to 0 ppm or even less.
The bacteriological quality of raw water taken from
the storm water canal for Jakarta Water Treatment Plants
was very bad showing VPS 500,000 - 2,000,000/100 ml. The
2
8-7
-------
condition 1b often worse during dry season where dilu -
tion is lacking1. The Dumber of pathogenic organisms auoh
as Salmonella and viruses are expected to be very high
also since the main sources of pollution are human waste
and refuse. Only proper treatment of water supply and
extra precaution of boiling the water before drinking
would protect; those who are consuming water from such
sources of raw water.
For the purposes of flood oontrol during the rainy
season and additional supply of water during the
drought several water reservoirs were oonstruoted in Ja-
karta. One of the reservoirs is the Pluit-reservoir and
located at the extreme and close to the sea, it receives
also wastes from industries whioh are located in that
area.The BOS of the reservoir water sometimes went up
to 1000 ppn or more.
In June 1975 dead fishes ware found in Brantas River
East Java olose to the water intake of Surabaya Water
Treatment Plant. Investigation shown that these fishes
were killed by HOT from a certain vetsin factory.
Several oaies were also reported about fresh water
fishes which were killed by pesticides used carelessly
for agriculture.
Only 4 cities In Indonesia are sewered and none of
these have any sewage treatment. These 4 cities do not
include Jakarta, because Jakarta sewerage system Is yet
under study by the UNDP. A complete sewage treatment
plant is an expensive process whioh mostly
becomes a
burden t* other establishments such as water works. In
a way Indonesia is probably fortunate to have the op-
portunity to think of the best way of treating or uti-
lising the sewage. Ve oould benefit from the experien-
ces and findings of other countries in this respeot.
Some of the human waste and most of the industrial
wastes esoape into open canals whioh are meant for
a to to water drainage. During the dry season these oa -
rials were septic causing undesirable smell to the
surrounding areas and creating an unaesthetic view of
black stagnant water.
Most water from shallow wells in densely populated
cities hardly meets the requirements of drinking water
due to serious soil pollution by septio tank effluents.
In certain parts of the oountry fer instance in
Vest Kalimantan w* have a problem of oolor present in
water due to organio matters from the trees along the
rivers. The oolor is not readily removed by ordinary
treatment process and filtration.
Some coastal areas such as the northern coast of
Java and the southern coast of Kalimantan suffer from
the saline water intrusion. In Kalimantan the saline
water intrusion was reported to be further drawn to the
inland waters due to the building of Gigantic Hydro
Electric Power in Riam Kanan, The pure water of the
virgin island of Irian Jaya in Tembagapure was reported
to be unfit for human consumption due to the waste from
coppermining in the area. These two cases need to be
studied further as examples of the effects of develop-
ment projects on health and ecology. Along this line
several studies have been carried out by the National
Institute of Health Research and Development in collab-
oration with the US Navy Medical Research Unit concern-
ing the occurrance and spread of Oncomelania Bupensia,
a certain species of snail which is responsible for
Schistosomiasis in Lake Lindu, Central Sulawesi. In
this area a dam was built for irrigation purposes which
may contribute to the spread of the snails.
V&ste from the fertiliser plant in South Sumatera
and from the petzo-ohemioal industry In Bast Java do
oreate nuisances for the people living in those areas.
The seriousness of coastal and marine pollution by oil
spill and tanker accidents was obvious from the big
fire at Tanjung Priok harbor in Jakarta (1973) and the
Showa Maru accident (1975).
Mr PQJ-Wto-
Industrial development in Indonesia has not yet
reached the stage which causes serious problems of air
pollution. Air pollution in big cities is mostly
caused by motor vehicles, aircraft, train, water trans-
port and other transportation means. These are refleo-
ted from the high 00 concentration between 50-95 PP«
(l hour grab sample, 4-5 samples taken within 10
hours/day In roads with heavy traffic in Jakarta du-
ring peak hours of the day) while the concentration of
dust in such roads ranges between 1-9 mg/cum of air
(Samples taken 6 days in a week between 8.00 a.m. -
4.30 p.m. with 30 minutes break for the engine). An-
other source of duBt besides the transportation means
is construction activity, especially road mainte-
nance using sand to protect the asphalt from melting
under the sun. In rural areas natural dusts and fo-
rest fires including agricultural burning are the
main causes of air pollution, which at the moment are ne-
gligible. Probably unpaved toads contribute more to
dost concentration in the air for rural areas.
The Industrial development will continue at a
much faster rate In the Second Five Tear Development
Plan. Mining of oil, tin, nickel, oopper eto. will al-
so continue, especially in the Baatern parts of Indo-
nesia Where agriculture is unpromising if not
impossible.
Due to the fact that Indonesia consists of 13,000
islands out of whioh 3,000 were inhabits tad we have
thus quite a long ooastal line to maintain and to pro-
teot against pollution.
KBPRB33IVB IBP PRS7EHTIVE MBASPHKS.
The existing conditions should be improved and
preventive measures should be taken to preserve tha
environment.
For the above mentioned requirement the following
steps should be taken s
1. Improvement of water supply and waste disposal both
for urban and rural areas, supported by the public
health education,
2. Control of ooianuni cable diseases which are associated
with poor hygiene and sanitation such as cholera,
diarrheal diseases, skin and aye diseases, soil-
transmitted helminthiasis etc.
3. Slim clearance followed by transmigration or rese-
ttlement of Its inhabitants.
4. Setting up policies, rules and regulations oonoern-
lag :
a. The water and land use.
b. The water resources development. tke
c. The environmental pollution oontrol in'foms of s
- Stream or Effluent Standards for water.
- Ambient or Emission Standards for air
including the oontrolvmarine pollution by oil.
d. Selective policy on import of vehicles, machina-
ble«, pesticides etc.
a. The oonservation of forest
eto.
5. Careful city and regional planning with special
3
8-7
-------
reference to planning of settlesients Including the
infra-structures, and also planning of industrial
development.
6. Intensification of the existing control of pestici-
des regarding the import, distribution, preparation
and use.
7. Recycling of all kinds of waste, converting than inr-
to beneficial aaterlals.
- hunan wast* and animal dunga for fertiliser
or to produce gas for cooking and limiting.
- nut shells, rloe bosks, saw duat etc. pro-
cessed for building aaterlals.
- refuse oould be turned to composed for ferti-
liser or for sanitary landfill and other be-
neficial uses.
- savage used in sewage fanning
- etc.
To check the quality of the water or air in the
environment to neat certain standard, a monitoring system
la needed. Appropriate inatmaentatlon and qualified
personnel arc also required to carry out tills nonlto -
ring system. Thereby proper planning and oareful feaala-
bility study oonoarnlng this monitoring system should
be nade before one decides to start with continuous moni-
toring, especially in developing countries where
there aw lade of such Instruments and qualified perso-
nnel.
4. International Assistances In teras of expertise,
training and provision of equipments are very Much
welcome for developing countries like Indonesia.
5. Information systems between neighboring countries
and with more developed countries might be useful.,
6. Standardisation of nethods and prooedures of moni-
toring ef the environmental quality will be useful
to avoid any difficulty in the interpretation of
quality of different environments.
#). The National Preparatory Connittee of Habitat
Indonesia "Hunan Settlements in Indonesia*
National Interim Report, May 1975.pp 5,6.
Lsr.iL met.
Water Pollution and Air Pollution Control Aots
will be formulated and this night tales quite a long
time before thesa oould be passed as National Aots.
Meanwhile cities like Jakarta have passed their
own regulations due to urgent nee da for controlling the
fast increasing pollution, because the existing *ui-
ssanoe Abatement Aot (1919) will no longer suit the
present conditions.
Standard for the quality of raw water for drinking
water supply has bean fomulated. Other standards rele-
vant to the preservation of the environmental quality
will also be prepared.
It sight be worthwhile to learn from the US experi-
ence with its Glean Air Aot for instanoe that we have
to consider not only the health haeaxds of pollution
but rather public health and welfare as a whole. This
public health and welfare (for air pollution) includes
Injury to agricultural crops and livestock, daaage to
and the deterioration of property and ha sards to air
and ground transportation.
(30KCUJ3I0K3 ASP RZ00WKHMTI0H3.
In order to be able to carry out a monitoring system
of the environmental quality In Indonesia we need :
1. Training of personnel with suitable beak-ground
knowledge ( oheniata, physioo-chemists, cheaioal-
englneers ate.;
2. Provision of no ni to ring equipments and materials
needed ( including sampling and analysis of samples )
These equipment* should be suitable for local Indo-
nesian conditions. If possible we should develop
our own monitoring apparatus.
3. Control of the sources of pollution should be conti-
nued not only through engineering Masures but also
toxougi management and regulation.
4
8-7
-------
LEGAL REQUIREMENTS FOR MONITORING GROUNDWATER QUALITY
Harvey 0. Banks
Harvey 0. Banks, Consulting Engineer, Inc.
Belmont, California
and
Leslie G. McMillion
If.S". Environmental Protection Agency
Las Vegas, Nevada
Summary
The needs of the Environmental Protection Agency
(EPA) for those data pertaining to groundwater quality
which must be obtained through monitoring programs,
are dictated by the mandates and requirements of the
Federal Water Pollution Control Act, as amended by
PubHc Law 92-500 (33 USC 1151 et seq.), and the Safe
Drinking Water Act,Public Law 93-523 (42 USC 300f,
et seq.). relating to groundwater quality protection.
An adequate data base is essential to fulfill the
mandatory duties imposed on EPA by these statutes in
a cost-effective manner. As indicated in the EPA
publication, "Model State Water Monitoring Program"
(EPA-440/9-74-002, July 1975), primary responsibility
for monitoring in fulfillment of the provisions of
Public Law 92-500 will be placed with the States.
Certain interstate compacts provide for monitor-
ing authority and activities by the compact Commis-
sions. State agencies charged by statute with
authority and responsibility for water resources
management and water quality control carry out monitor-
ing programs; these vary widely in scope among the
states and among state agencies within a State. Local
agencies often conduct extensive programs.
The purposes to be served, of which there may be
several, by the data to be obtained through a monitor-
ing program for a groundwater system must be defined
and the data needs carefully evaluated. A specific
monitoring program must be designed and implemented
for each groundwater system. The extent of the moni-
toring program required will depend upon present
quality problems, the sources and causes of pollution
and the relative present and future significance, the
types and importance of the uee(s) of the groundwaters,
the geologic, hydrologic and hydraulic complexity of
the system, and the data already available. The
dynamic nature of groundwater systems must be clearly
recognized in designing and implementing monitoring
programs. The need to develop the capability to
predict potential quality problems in order to formu-
late preventative programs on a timely basis is
important in evaluating data requirements.
Federal Water Pollution Control Act, aa amended
The objective of the Act as amended is stated in
Sec. 101(a) as:
"Sec, 101(a) The objective of this Act is
to restore and maintain the chemical, physical,
and biological integrity of the Nation's
waters..."
The definition of "pollution" given in the Act
indicates clearly the wide spectrum of groundwater
quality problems and of the sources and causes of
those problems that may need to be considered in de-
signing a groundwater quality monitoring program:
"Sec. 502.(19) The term 'pollution' means
the man-made or man-induced alteration of
the chemical, physical, biological, and
radiological integrity of water,"
Man's activities that may result itv impairment of
groundwater quality cover a wide range including not
only the disposal of municipal, industrial and agri-
cultural wastes, but also other causes not directly
associated with waste disposal such as reduction in
recharge or overpumping from coastal segments of aqui-
fers resulting in saline water intrusion.
The Act directs that specific programs be de-
veloped to improve and maintain groundwater quality:
"Sec. 102. (a) The Administrator (of EPA)
shall ... , prepare or develop comprehensive
programs for preventing, reducing, or eliraini-
nating the pollution of the navigable waters
and groundwaters and improving the sanitary
condition of surface and underground waters.
In the development of such comprehensive pro-
grams due regard shall be given to the improve-
ments which are necessary to conserve such waters
for the protection and propagation of fish and
aquatic life and wildlife, recreational purposes,
and the withdrawal of such waters for public
water supply, agricultural, industrial, and
other purposes.,,"
Development of cost-effective programs in fulfill-
ment of this mandate will require an adequate data
base which can be obtained only through carefully de-
signed and implemented monitoring programs encompass-
ing a wide range of types of data.
Establishment and conduct of groundwater quality
monitoring programs by both EPA and the States is
mandated by the Act:
"Sec. 104.(a) The Administrator (of EPA)
shall establish national programs for the
prevention, reduction, and elimination of
pollution as part of such programs shall: •••
(5) in cooperation with the States, and
1
9-1
-------
their political subdivisions, and other
Federal agencies establish, equip, and
maintain a water quality surveillance
system for the purpose of monitoring
the quality of the navigable waters
and groundwaters and the contiguous
zone and the oceans... and shall report
on such quality..."
"Sec. 106.(e) Beginning in fiscal
year 1974, the Administrator shall
not make any grant under this section
(grants for pollution control programs)
to any State which has not provided or
is not carrying out as a part of its
program...
"(1) the establishment and operation
of appropriate devices, methods, sys-
tems, and procedures necessary to
monitor, and to compile and analyze
data on (including classification
according to eutrophic condition),
the quality of navigable waters and
to the extent practicable, ground-
waters including biological monitor-
ing; and provision for annually updating
such data and including it in the report
required under Section 305 of this Act;"
Safe Drinking Water Act.
This Act, which became effective 16 December 1974,
places additional responsibilities of great public
health and economic significance upon the Administra-
tor of EPA to protect the Nation's groundwater re-
sources. The following excerpts from the Act indicate
the nature and extent of those responsibilities:
"Sec. 1424. (e) If the Administrator (of
EPA) determines, on his own initiative
or upon petition, that an area has an
aquifer which is the sole or principal
drinking water source for the area and
which, if contaminated, would create a
significant hazard to public health, he
6hall publish notice of that determina-
tion in the Federal Register. After the
publication of any such notice, no commit-
ment for Federal financial assistance
(through a grant, contract, loan guarantee,
or otherwise) may be entered into for any
project which the Administrator determines
may contaminate such aquifer through a re-
charge zone so as to create a significant
hazard to public health, but a commitment
for Federal financial assistance may, if
authorized under another provision of law,
be entered into to plan or design the project
to assure that it will not, so contaminate
the aquifer."
"Sec. 1442(a)(1) The Administrator may
conduct research, studies, and demonstra-
tions relating...to the provision of a
dependably safe supply of drinking water,
including...
"(E) improved methods of protecting under-
ground water sources of public water systems
from contamination.. .
"(4) The Administrator shall conduct a
survey and study of. ..
"(A) disposal of waste (including resi-
dential waste) which may endanger under-
ground water which supplies, or can
reasonably be expected to supply, any
public water systems, and---
"(5) The Administrator shall carry out
a study of methods of underground injection
which do not result in the degradation of
underground drinking water sources.
"(6) The Administrator shall carry out a
study of methods of preventing, detecting,
and dealing with surface spills of con-
taminants which may degrade underground
water sources for public water systems...
"(8) The Administrator shall carry out
a study of the nature and extent of the
impact on underground water which supplies
or can. reasonably be expected to supply
public water systems of (A) abandoned
injection or extraction wells; (B) intensive
application of pesticides and fertilizers in
underground water recharge areas; and (C)
ponds, pools, lagoons, pits, or other surface
disposal of contaminants in underground water
recharge areas."
Decisions made responsive to this Act will have
great economic impacts. To carry out these responsi-
bilities effectively, detailed information for many
aquifers on groundwater quality, geology, hydrology
and the sources and causes of groundwater quality,
impairment, will be necessary. The recharge areas
must be carefully delineated. This will be especially
difficult for some aquifers, the recharge areas of
which are extensive. Groundwater quality monitoring
programs must be designed to provide such data where
and when, required.
Interstate Compacts
A few interstate compacts—the Federal-interstate
compacts for the Delaware River Basin and the Susque-
hana River Basin are examples— provide authority for
the compact commissions to monitor groundwater quality,
and to develop and implement plans for quality protec-
tion.
State Statutes and Regulations
State laws and implementing regulations vary
widely from state to state as regards both groundwater
quality monitoring and the development and implementa-
tion of plans for groundwater quality protection.
There are generally several state agencies which have
statutorily defined authorities and responsibilities
with respect to groundwater, and which conduct monitor-
ing activities in the State. Such state agencies may
include:
.Water resources planning and manage-
ment agencies;
2
9-1
-------
• Mater quality control agencies;
.Water rights administration;
•Department of public health;
.Universities;
•Soil conservation agencies;
The authorities, responsibilities and activities of the
several state agencies may overlap to some extent.
At the regional and local levels of government,
counties, municipalities and special political juris-
dictions such as water supply, sanitary, flood control
and water conservation districts, have interests and
responsibilities with respect to groundwater, and may
conduct extensive monitoring programs.
In California, the State Water Resources Control
Board and the nine regional water quality control
boards have broad authority and responsibilities with
regard to groundwater quality monitoring and quality
control under Division 7 of the Water Code (Porter-
Cologne. Water Quality Control. Act, as amended). Like-
wise, the Water Code provides with respect to the State
Department of Water Resources:
"229. The department of Water Resources...
shall investigate conditions of the quality
of all waters (including groundwaters) with-
in the state, including saline waters, coastal
and inland, as related to all sources of pollu-
tion of whatever nature and shall report thereon
to the Legislature, to the board (State Water
Resources Control Board), and to the appropriate
regional water quality control board annually,
and may recommend any steps which might be taken
to improve or protect the quality of such waters.
The department shall coordinate its investiga-
tions fully with the board.
The Water Code of the State of Texas provides
similar authorities and responsibilities to the Texas
Water Quality Board, and, to some extent, the Texas
Water Development Board,
Some States have not yet progressed as far in
monitoring and planning for the Improvement and pro-
tection of groundwater quality.
Monitoring Objectives
Simply stated, the objective is to collect, manage
and analyze the data on groundwater quality and the
sources and causes of groundwater pollution, and other
information, geologic, hydrologic and economic, neces-
sary to enable EPA and the state (s) Involved to fulfull
their statutory responsibilities as regards protection
of groundwater quality, that is "...to restore and
maintain the chemical, physical and biological integrity
of the Nation's waters..."
To fulfill that objective, a monitoring program
must be developed and implemented for each groundwater
system, which may consist of several interrelated
aquifers with different characteristics. The data to
be obtained may be needed for one or more of the
following purposes, depending upon the types, extent
and seriousness of present and potential quality pro-
blems and the present and future importance of the
system as a source of water supply for various uses:
.Provision of background information on
quality;
.Detection of quality trends;
•Identification and assessment of present
and potential sources and causes of pollu-
tion;
•Planning for groundwater quality control—
including formulation, calibration, verifi-
cation and use of planning models;
.Establishment of water quality standards and
effluent limitations;
•Formulation of other regulatory control and
management actions necessary to protect
quality;
.Compliance;
•Enforcement—including identification of re-
charge areas pursuant to the Safe Drinking
Water Act;
•Providing verification data for scientific
and research endeavors;
•Development and utilization of hydrologic
and water-quality models;
.Reporting.
For a groundwater system which is already extensively
developed, or in process of development, all these
purposes must generally be served, although the
relative emphasis will differ.
Data Needs
The needs for groundwater data to be supplied by
monitoring programs are both extensive and urgent if
the statutory requirements relating to groundwater
quality protection are to be properly fulfilled.
The types and extent of monitoring data needs for a
specific groundwater system generally will depend on
a number of factors, often interrelated, including
present quality problems, sources and causes of
pollution and their relative present and future sig-
nificance, types and importance of the use(s) of the
groundwaters, geology and hydrology of the particular
system, and its complexity and variability. A wide
range of types of data may be needed, including in
addition to quality data:
.Geologic characteristics, surface and sub-
surface, areally and vertically;
.Hydrology;
.Hydraulic characteristics, areally
and vertically;
.land use;
.Water resources development and use;
.Demographic and economic;
.Waste generation and disposal.
3
9-1
-------
Some of the needed data may already be available from
the monitoring programs of other Federal agencies,
State agencies and other sources.
An effective monitoring program must recognize
the dynamic nature and variability of groundwater
systems as affected by both natural phenomena and man-
induced changes. The program must, therefore, be con-
tinuing. Its scope and relative emphasis will change
over time. The data obtained must be adequate to en-
able prediction of potential quality problems and for
formulation of timely plans for prevention.
To obtain the required data, use of several
different monitoring techniques may be necessary as
will be discussed in another paper. Acquisition of
the needed information should be accomplished in the
most cost-effective manner.
4
9-1
-------
ECONOMIC FRAMEWORK FOR GROUNDWATER QUALITY MONITORING
Robert L. Crouch
Department of Economics, University of California, Santa Barbara
Consultant, General Electric-TEMPO
Ross D. Eckert
Department of Economics, University of Southern California
Consultant, General Electric-TEMPO
Donald D. Rugg
Neil Campbell Associates, Los Angeles
Consultant, General Electric-TEMPO
Summary
This paper discusses the economic considerations
which must be taken into account in selecting an opti-
mal groundwater quality compliance monitoring strategy.
The economic issues raised by other monitoring problems
(such as monitoring for information) cannot be treated
here but are discussed in Crouch, et al_., from which
this paper is extracted.
generates a chloride concentration at the farmer's well
equal to OR, thereby imposing marginal damages on the
farmer equal to RR'. Using GNP maximization as the ob-
jective function, the optimal level of groundwater pol-
lution is OS— which generates a net gain to society
equal to A minus the monitoring (and other enforcement)
costs incurred to implement that policy.
Introduction
If (1) all markets were competitive, (2) market
transactors had full information, (3) resource-owners
were utility maximisers who convey the resources they
own to their highest valued use, and (4) enforceable
property rights existed in all resources, then society's
resources would be efficiently allocated.^ Under such
an allocation of society's resources, groundwater re-
sources would be used in optimal amounts for irrigation,
cooling, drinking, etc, as well as for the disposal of
wastes. The reason that groundwaters become excessively
polluted is that the fourth condition necessary for ef-
ficient allocation of resources (enforceable property
rights) is not now met with respect to groundwaters
As a result, certain parties use other parties' ground-
water for disposing of their wastes without authoriza-
tion, thereby imposing uncompensated damages on the
groundwater owners and external costs are generated. In
the absence of any obligation to compensate those whose
groundwaters he is using as a waste disposal site, such
waste disposal services appear free to the polluter and
he will, therefore, utilize such a service until the
marginal private benefit obtained from using that ser-
vice is driven down to zero- even though the marginal
social cost may be positive and large. Consequently,
the aquifer's waste receptor capabilities are abused and
it becomes excessively polluted.
MARC.lNAt DAMAGE
f UNCI ION
Preventing Excessive Pollution of Groundwater
In principle, there are two ways of preventing ex-
cessive pollution of groundwater. First, by government
intervention in the effluent discharge process (through
either discharge licenses, an effluent tax, or an ef-
fluent reduction subsidy); second, the property rights
to groundwater can be respecified so that they approach
more closely the basic requirements of a well-specified
property right. (This latter alternative is discussed
fully in Section 5 of Crouch, et aK )2. Space limita-
tions dictate that, here, we can only discuss the pre-
vention of excessive groundwater pollution via discharge
licensing, or a permit scheme.
The Expected Cost of a Standards Violation
As explained above, the standard contained in the
oil company's permit should correspond to OS. We shall
assume that the permittee will comply with the terms of
its permit only if the expected costs of noncompliance
exceed the cost of compliance and will not comply if the
expected cost of noncompliance is less than the cost of
compliance. The expected cost to a permittee of violat-
ing a standard is equal to
Figure 1.
Figure 1 illustrates the situation where we assume
that the chlorides from an oil company's unlined brine
disposal pit (the polluter) are contaminating a farmer's
irrigation well (the polluttee). The oil company
E(0 = Pd • Pc
E(F)
o:
where Pj is the probability of the standard violation
being detected, Pc is the (conditional) probability of
conviction (given detection), and E(F) is the expected
fine (given conviction). Without loss of generality,
we always assume Pc = 1.
The Probability of Detection Function
The probability of being detected violating a stand-
ard is a function of the amount of compliance monitoring
undertaken- as measured, say, by expenditure on compli-
ance monitoring, M--and the severity of the violation,
V-as measured by the difference between the actual level
of pollution emitted and that permitted by the standard.
Thus,
= f(Mc,V)
(2)
with f!| > 0, f^' < 0, f'2 > 0, and f'L = 0 since, plausibly,
we may assume diminishing returns to monitoring and that
the probability of detection increases as the size of
the violation increases (but at neither an increasing
or diminishing rate). The ceteris paribus relationship
between Pj and Mc (for varying size violations, Vg > V-|)
1 9-2
-------
is illustrated in Figure 2 while the ceteris paribus
relationship between and V (for varying levels of
expenditure on compliance monitoring M£ > mJ) is illus-
trated in Figure 3.
Figure 2.
SI VIHI
[tXCI'A CCJNC.If,ltA!ir
Figure 3.
The Expected Fine Function
We postulate that the expected fine E(F) is a func-
tion of the size of the violation V (with an upper
bound equal to the statuory maximum). That is,
E(F) = j(V)
(3)
with j'> 0 and j"> 0, by plausible assumption. This is
illustrated in Figure 4.
V, SfVEKMY Of VIOUIION
(EXCm CONCINICAHON
Of ClUOIMDfS in ity/l'lo')
The _E)^_ec ted _Cost_of_ Vjk)l_a_tincj a Standard,
Cornpliance Monitoring Expenditures,
and Size of Violation
By substituting equations (2) and (3) into (1), we
find that
E(C) = f(Mc, V) • j(V) (4)
(since Pc = 1, by assumption). The ceteris paribus
relationship between E(C) and V (for varying levels of
expenditure on compliance monitoring, M£ > M^) is il-
lustrated in Figure 5.
Figure 5.
This relationship is merely a scalar transform of Equa-
tion (3) and (Figure 4). It will emanate from the
origin, be convex from below, and terminate at Fmax.
The Permittee's Optimal Violation
The permittee's total benefits function (corres-
ponding to his marginal benefits function illustrated
in Figure 1) with respect to the chloride concentrations
generated at the farmer's well is illustrated in
Figure 6 (for the range of chloride concentrations in
excess of the standard).
. SEVEtlTY
VIOlAJ
(CONClNlfeAl
ot ct
Figure 4.
Figure 6.
It increases monotonically between S and R when it reaches
a maximum—which is the chloride concentration the per-
mittee will generate if no attempt is made to regulate
2
9-2
-------
his activities. The permittee's expected cost function
(from Figure 5) is also included in Figure 6.
The permittee's total benefit and cost functions
can be used to establish the level of emissions that it
is optimal for the permittee to generate and, therefore,
the optimal severity of the standard violation for the
permittee to undertake. The assumed objective is to
choose V so as to maximize his net expected gain H—
where H is defined to be the difference total benefits
and expected cost. Diagrammitcally, this occurs at V-|
when the slopes of the total benefit and expected cost
curves are equal. Analytically, the maximum occurs
where the marginal benefit from increasing the viola-
tion is equal to the marginal expected cost from in-
creasing the violation. (If the permittee derives dis-
utility from engaging in illegal activities, the opti-
mal violation V-j is reduced-but not necessarily to
zero. See Crouch, et aJL , Section IV). Clearly, the
optimal violation V] will be reduced by increasing ex-
penditures on compliance monitoring since this increases
the slope of the expected cost function everywhere. As
explained in the next section, however, it is not neces-
sarily optimal from society's point of view to increase
compliance monitoring expenditures enough so as to re-
duce V] to zero. That depends on the marginal benefits
derived from monitoring relative to its marginal cost.
Optimal Compliance Monitoring Expenditure
The "output" from compliance monitoring is varying
levels of probability of detection. The marginal cost
of compliance monitoring as a function of the proba-
bility of detection will increase (due to diminishing
returns) as shown in Figure 7.
com, *
- MAKL.IMAI BfHll 11
COMNIANCI
MARi.iNAi cost cum »r«-
coMPUAwa MONUOKiNc
Figure 8.
AavOINII), s
Figure 9.
The relationship between various levels of proba-
bility of a violation being detected and the total
social loss avoided may be inferred from Figure 10.
IXPENOITIME ON—
Figure 7.
Since a permittee's optimal violation of a stand-
ard will decrease as the expenditure on compliance
monitoring increases (and vice versa) this allows es-
tablishment of the functional relationship between ex-
penditure on compliance monitoring and the social gain
from that expenditure. The social gain from monitoring
is defined to be the social loss avoided by that mon-
itoring. To illustrate this, consider Figure 8. If no
monitoring is undertaken, the expected cost to the
permittee is zero—since Pj = 0, see Equation (l)-and
he will generate OR pollutants at the farmer's well.
The loss to society is equal to A + T-j + T?. If suffi-
cient monitoring is undertaken to reduce trie permittee's
violation by one unit (to SR], the loss avoided is equal
to area T]. Analogously, if enough additional monitor-
ing is undertaken to reduce the violation by one more
unit (to SR2), the total loss avoided is T, + T2 and the
marginal loss avoided is Tg. Thus, the total loss
avoided increases as the violation decreases—and it
does so at a diminishing rate towards a maximum equal to
A (which occurs at zero violation (see Figure 9).
^2__IOTAl social
' LOSS AVOIDED
VIOLATION
tXPfCIED COST, 1(C)
Figure 10.
The upper-left quadrant shows the relationship be-
tween Pj and Mc (Figure 2). The lower-left quadrant
shows the expected cost of a violation to the violator
3
9-2
-------
as a function of the probability of detection- which,
from Equation (1) is a proportional relationship. The
lower-right quadrant shows the optimal violation as a
function of the expected cost. This relationship shows
that as the expected cost of a violation increases, the
optimal violation decreases (but at a diminishing rate-
as may be inferred from Figure 6). Finally, the upper-
right quadrant reproduces Figure 9. (There is no con-
nection between the upper-left and upper-right quad-
rant). To establish the relationship between the prob-
ability of detection and the total social loss avoided,
assume that the expenditure on compliance monitoring is
increased from to in order to increase,the proba-
bility of detection by one unit from P^ to P§. As a
result the total social loss avoided increases from l_i
to l_2- If expenditure on monitoring is increased to
Mjj, the probability of detection is increased by one
m6re unit and the total social loss avoided increases
from 1-2 to Lj. Obviously, the relationship between
marginal social loss avoided and the probability of
detection (ie, the marginal benefit from compliance
monitoring) is an inverse relationship- as, in fact,
illustrated in Figure 7. Referring to that figure, we
may infer that the optimal probability of detection to
aim for is PJ—where the marginal benefit from compli-
ance monitoring is equal to its marginal cost. Once
the optimal probability of detection to aim for is es-
tablished, the optimal expenditure on compliance mon-
itoring may be inferred (from Figure 2). This identi-
fies the optimal compliance monitoring budget. Note
that, in general, the optimal detection probability is
not that which would induce perfect compliance with the
standard. (This should not be surprising. To use a
rather precise analogy, the maximum highway speed limit
is a standard which is enforced by policing the high-
ways. The highways could be policed so intensively
that no one would exceed the speed standard. However,
they are not because the marginal social cost of such
an increase in policing intensity would exceed the
marginal social benefit from deterring the few persons
who now speed from speeding.)
Irnglementation Problems. of_Monj.toring_
Some practical difficulties attend the implementa-
tion of the optimal compliance monitoring strategy. It
is inconceivable that Pjj will remain unchanged over
time. The marginal cost curve for compliance monitor-
ing will change as technological innovations in the
monitoring industry occur and as the relative prices of
various alternative monitoring techniques change. Sim-
ilarly, the marginal benefit curve will change as the
marginal social gain from monitoring (ie., the loss
avoided) changes. And this will change as the pol-
luter's marginal benefits function from emitting
pollutants and the polluttee's marginal damage function
from the impact of pollutants change. Clearly, these
depend on a whole host of factors such as relative
product and factor prices, the state of technology, and
so on. It was these practical considerations which led
to the exploration of an alternative, market-oriented,
scheme for obtaining optimum use of the waste-
assimilative capacities of the nations groundwaters
analyzed in Section 5 of Crouch, et al/
PIkLioflrspllicaL References_
1. Corker, C. (ed), Groundwater Law, Management, and
Administration, National Water Commission,
NWL-L-72-06, Final Report, Oct. 1971.
2. Crouch, R.L., R.D. Eckert, and D.D. Rugg, Monitor-
ing Groundwater Quality: Economic Framework" and '
Principles, General Electric-TEMPO, Report
GE75TMP-52.
3. Quirk, J. and R. Saposnik, Irjtroduc.yon to General
Equilibrium Theory and Welfare Economics, McGraw-
Hill ,"1968:""
4
9-2
-------
DEVELOPMENT OF A METHODOLOGY FOR MONITORING GROUNDWATER QUALITY
David K. Todd
University of California
Berkeley, California
and
Lome G. Everett
General Electric TEMPO
Santa Barbara, California
Summary
Both the Federal Water Pollution Control Act
Amendments of 1972 {PL 92-500) and the Safe Drinking
Water Act (PL 93-523) stress the need for a national
groundwater quality monitoring program. This paper
outlines a methodology for developing such a program,
based on studies currently underway for the U.S. Envi-
ronmental Protection Agency. The methodology consists of
a series of procedural steps which define a sequence of
actions leading to a monitoring program. The approach is
general so that it can be applied in a local area under
a variety of land-use and hydrogeologic conditions.
Introduction
Recent actions by the Congress to protect the
quality of waters in the United States have included
provisions for a national groundwater quality monitor-
ing program. A major impetus came from the Federal
Water Pollution Control Act Amendments of 1972 (PL 92-
500), which specified that EPA is to protect under-
ground waters, to withhold grants to States not having
a groundwater quality monitoring program, and to require
disposers of wastes to monitor their activities. Sub-
sequently, the Safe Drinking Water Act (PL 93-523)
stated monitoring and reporting requirements associated
with all water supply and underground injection systems.
The purpose of this paper is to present a procedure
for development and implementation of a cost-effective
methodology for monitoring groundwater quality. The
methodology is expressed in a generalized form so that
it can be usefully employed by State and local water
pollution control agencies and is applicable to all
types of groundwater aquifers, areas, and basins.
For convenience the monitoring methodology is pre-
sented in the form of a series of procedural steps
arranged in chronological order. By so doing a
straightforward sequence of actions is outlined which
can lead to a groundwater quality monitoring program in
a given area. It should be apparent that factors such
as climate, hydrogeology, population, pollution sources,
and water use vary from place-to-place; therefore, the
design of an appropriate monitoring program will also
vary accordingly. No one set of guidelines can cover
all situations; however, with judgment the approach pre-
sented herein can be extended and interpreted to meet
most extenuating situations.
A national program to protect groundwater quality
must focus primarily on measurements relating to sources
and causes which contribute to pollution. Furthermore,
because it is infeasible to monitor all sources and
causes of pollution, the methodology concentrates on
identifying the most important sources and causes. In
essence then the methodology becomes a resource alloca-
tion problem with the goal of developing a cost-effec-
tive monitoring program which will contribute most to
the protection of the nation's groundwaters.
Concept
A monitoring methodology should serve two roles
simultaneously. One, it should assist a local desig-
nated monitoring agency (DMA) to design and to imple-
ment a monitoring program. To accomplish these goals,
the methodology must address itself to the specific
details of pollution sources and causes, of monitoring
techniques, and of monitoring costs. The second role
should be to guide governmental agencies at the State
and national levels in establishing realistic monitor-
ing priorities, not only in terms of what should be
monitored and where, but also in terms of timing and
funding goals.
The monitoring methodology described in the follow-
ing sections is predicated upon the technical effort
being organized and conducted by or under the direct
supervision of personnel with professional training in
water resources engineering or groundwater geology.
Implementation
The following series of steps describes a chrono-
logical procedure for implementing a groundwater qual-
ity monitoring program. This applies to a specified
local area under the jurisdiction of a DMA. Activities
of different steps will, in practice, overlap in order
to make efficient use of personnel and time.
Step 1 - Select Area or Basin for Monitoring
The selection of areas to be monitored will be
made within a State by the appropriate State water pol-
lution control agency that works cooperatively with EPA
in carrying out the mandates of PL 92-500. The basis
for selecting areas will be governed, in general, by a
combination of (a) administrative, (b) physiographic,
and (c) priority considerations. Each of these factors
will be reviewed in the following paragraphs.
(a) Administrative Considerations. The initiation
of a monitoring program requires that a local desig-
nated monitoring agency (DMA) be specified. In many
situations the requisite agency with the necessary
technical staff may be a county, district, or regional
water organization. Thus, the area to be monitored can
often be made to correspond to the jurisdictional area
of the DMA. The size of a particular area may vary
from a few square miles to thousands of square miles.
Size alone is less important than the ready accessibil-
ity of all portions of the area to the DMA as well as
hydrogeologic knowledge of the area possessed by the
DMA.
It should be recognized that political boundaries
frequently create problems in terms of water management.
Such a boundary may cross a major groundwater basin so
that, for example, pollutants from an adjoining area
may be entering from sources not subject to monitoring
by the DMA. Clearly, such situations should be mini-
mized as much as possible; alternatively, cooperation
among DMA's sharing common groundwater pollution
l
9-3
-------
problems will be essential to the success of their
respective monitoring programs.
(bj Co(jyite.r®y_on!- The physiograph-
ic basis for selecting monitoring areas recognizes that
groundwater basins are distinct hydrographic units con-
taining one or more aquifers. Such basins usually, but
not always, coincide with surface water drainage basins.
By establishing a monitoring area related to a ground-
water basin, total hydrologic inflows to and outflows
from the basin are fully encompassed. This permits all
pollution sources and their consequent effects on
groundwater quality to be monitored by a single DMA.
Where basins are extensive, monitoring areas become too
large to be practical. Boundaries should then be drawn
parallel to groundwater flows or where cross-flow com-
ponents are insignificant.
(c) Priority Considerations. It is recognized
that establishment of a national groundwater quality
monitoring program will in all likelihood develop grad-
ually and in a piecemeal manner because of administra-
tive, budgetary, and personnel constraints. Therefore,
the question of priorities in selecting monitoring
areas within a State becomes a practical concern. In
order to be most effective, monitoring should logically
be started in those areas (a) which have the largest
number of groundwater pollution sources and (b) where
there is a high utilization of groundwater. Applying
these two criteria, the total area of a State can be
divided into planned monitoring areas. These areas in
turn can be ranked in terms of the two criteria, so
that the ranks constitute a priority basis for imple-
menting the Statewide monitoring program.
Step 2 — Define Groundwater Usage
To evaluate the impact of pollution or potential
pollution of groundwater, the usage of the resource
becomes a key item. Thus, it is important to define
both the quantities of groundwater being extracted and
the locations of major pumping centers within a monitor-
ing area.
Pumpage will usually vary on a weekly and on a
monthly basis; however, in most cases mean annual
extractions are sufficient to evaluate groundwater
flows and pollutant movement for monitoring purposes.
These data will serve as inputs to the hydrogeological
analysis in a subsequent step. Determination of ground-
water pumpage depends on the type of use. The basic
categories of use - municipal, industrial, agricultural,
and rural — should be evaluated separately.
Step 3 - Identify Pollution Sources and Causes
The design of a monitoring program requires that
the sources and causes of groundwater pollution or
potential pollution within an area be identified.
Although the actual number of sources and causes can be
large, particularly in populated areas, they can be con-
veniently classified into 18 principal sources and
causes.
Procedures for identifying sources and causes of
pollution are described in the following paragraphs.
Only rarely will all sources and causes be found in a
given monitoring area.
(1) Irrigation Return Flow. The presence of irri-
gation return flows will occur wherever irrigation
occurs and is independent of whether the applied water
originates from surface or underground waters.
(2) Animal wastes. The organic wastes of livestock
constitute a negligible groundwater pollution source
except where large number of animals are confined within
2
a small area. This situation is generally limited
either to feedlots for beef production or to large
dairies.
(3)_Fj^rJnlijay°n. Fertilizers in modern agricul-
ture are applied to almost all crops, whether irrigated
or not. Thus, areas subject to ferti1ization will in
most cases coincide with agricultural crop lands. In
addition, garden and lawn fertilizers should be consid-
ered in urban areas.
(4) Soil Amendments. Soil amendments are applied
to irrigated lands to alter the physical or chemical
properties of the soil, and these often have residual
effects on groundwater quality.
(5) Pesticide Residues. Pesticides may be signif-
icant in agricultural areas as a nonpoint source of
pollution.
(6) Solid Wastes. As all solid-waste disposal
sites are important possible sources of groundwater
pollution, they need to be pinpointed within a monitor-
ing area.
(7) Surface Disposal of Liquid Wastes. The dis-
posal of liquid wastes into the ground can take many
forms, and in large urban areas the number of sources
can be large; therefore, identification of these
sources will take more effort than for most other types
of sources.
(8) Sewer Leakage. The locating of all sewer
leaks in an urban area is from a practical standpoint
nearly impossible. For a monitoring program it is
sufficient to define where such leaks may be occurring.
(9) Tank and Pipeline Leakage. Leaks from under-
ground tanks and pipelines cannot be located directly
within a monitoring area. What can be done, as with
sewers, is to identify installations where leakage can
occur. Lacking a comprehensive inventory of installa-
tions, efforts should be limited to locating only major
facilities.
(10) Disposal Wells. Wells which are employed to
dispose of liquid wastes into freshwater aquifers are
obvious sources of pollution and need to be fully
identified. Besides locating disposal wells, informa-
tion should be gathered on the type and average quan-
tity of wastes disposed and the depth of the well.
(11) Injection WeVU. Injection wells are deep
wells~Tn whiclTl iquuTwastes are injected into saline-
water formations. For each such well information
should be obtained on the type and average quantity of
wastes disposed, the depth of the well, and the injec-
tion pressure.
(12) Stockpiles. The piling of solid raw or
waste materials on the ground without protection from
precipitation can lead to leaching and groundwater
pollution. Stockpiles are normally associated with
industrial plants or construction sites.
(13) Mining Activities. The mining and milling
of coal andlietallTc^oreiT"whether from surface or
underground mines, can cause groundwater pollution;
therefore, these installations need to be clearly iden-
tified if they occur in a monitoring area.
(H) Saline Water Intrusion. Intrusion of sea-
water in "coastal aquifers usuaTTy results from exces-
sive pumping and creates a nonpoint source of pollu-
tion. Areas of intrusion, therefore, can be expected
where wells are located close to the coast.
9-3
-------
In inland areas saline water intrusion occurs where
heavy pumping causes underlying saline water to rise
into wells. The circumstances leading to such intrusion
are highly variable and difficult to identify. Because
saline intrusion is not a uniquely defined source of
pollutants, such as a landfill, identification for mon-
itoring purposes consists primarily of finding where
intrusion is known to be occurring and/or recognizing
where the potential for intrusion exists.
(15) J\qujfer Interchange through Wells. Pollution
by interchange of'groundwater through wells occurs where
two aquifers with differing hydraulic heads and water
qualities are connected by a well. As with saline
intrusion, such occurrences are not easily identified.
(16) Soils_ and_ Surface_Discharges_. Sources such as
spills'Tnd"surface discharges are largely unidentifi-
able; however, for large industrial plants, information
may be available on the mass balance of selected
constituents.
(17) Septic Tanks and Cesspools. In unsewered sub-
urban areas and in rural"areas, large numbers of septic
tanks and cesspools will be found. Although each con-
stitutes an individual point source of pollution, from
an identification standpoint it is sufficient to locate
only the areas in which they occur.
(18) Highway Deicing. Information on the applica-
tion of salts for deicing of highways in winter should
be collected for roads which are regularly treated with
salts and on the estimated application rate.
The above listing stresses present-day sources and
causes; however, those which are no longer present or
active, such as former refuse dumps, brine disposal
ponds, and feedlots, may still be responsible for pol-
lution today. This occurs because the slow movement of
groundwater has not removed earlier pollutants. It
follows, therefore, that an effort should also be made
to find previous sources and causes that may no longer
be visible on the landscape but may be evident under-
ground.
Define Hydrogeologic Situation
To understand where and how groundwater pollution
occurs and moves within a monitoring area, the hydro-
geologic framework must be known. Specific items
needed for the monitoring program include:
(a) Aquifer location, depths, and areal extents.
(b) Transmissivities of aquifers.
(c) Map of groundwater levels.
(d) Map of depths to groundwater.
(e) Areas and magnitudes of natural and artificial
groundwater recharge.
(f) Areas and magnitudes of natural groundwater
discharge.
(g) Directions and velocities of groundwater flows.
Step ji_ — Study Existing Groundwater Quality
In order to define the groundwater quality prob-
lems within a monitoring area, an assessment needs to
be made of the background quality situation. To do this
recent and past groundwater quality data need to be col-
lected and reviewed.
Step 6 - Prioritize Pollution Sources and Causes
Based on the inventory of pollution sources and
causes and the assessment of groundwater conditions for
a monitoring area, accomplished in Steps 2 through 5,
it is now possible to prepare a ranking of sources and
causes of pollution. This ranking will define the
relative importance of sources and causes and, subse-
quently, the priorities for groundwater quality
monitoring.
The establishment of rankings for sources and
causes of pollution implies that the primary goal of
the monitoring program is to gather information on
groundwater quality where the resource is subject to
the greatest impact by activities of man. The magni-
tude of the impact can be expressed in terms of five
variables. Evaluating these leads to a classification
system for prioritizing sources and causes of pollution
in terms of their potential effect on man.
The five parameters include:
(1) Type of groundwater use
(2) Type of pol lutant
(3) Geological structure
(4) Vertical distance to groundwater
(5) Horizontal distance from source to use.
Combining the above five parameters, a generalized
classification procedure for prioritizing sources and
causes of pollution can be formulated.
Step 7 -- Evaluate E x i s_t i n g J^n it or i n g_ P r o^r a ms
Development of a monitoring program in an area
will rarely involve starting an entirely new water-
quality surveillance program. Almost always a variety
of existing monitoring activities will be found.
Every effort should be made to incorporate these on-
going activities in a new monitoring program. In
fact, such inclusion is essential if the program is to
be both comprehensive and cost-effective.
With the review of existing monitoring programs
completed, it is now possible to determine monitoring
deficiencies. This is accomplished by noting the
availability of existing monitoring activities serving
the high priority sources and causes. Clearly, if the
surveillance systems for these are inadequate or
incomplete, monitoring deficiencies exist.
The final result of this step is a priority list-
ing of sources and causes having monitoring deficien-
cies. The sequence of this listing will be identical
to that developed in Step 6, except that sources and
causes which are adequately monitored will be
eliminated.
Step 6 — Identify Alternative Monitoring Approaches
This step involves the determination of the most
cost-effective methods for monitoring the prioritized
sources and causes of pollution listed in Step 7.
Specifically, this involves selecting for each high
priority source or cause the monitoring method(s)
which best meet the existing deficiency.
To evaluate monitoring approaches requires that
consideration be given to the following:
(a) Tht: type of pollution source or cause.
(b) The type of pollutants reaching the ground-
water.
(c) The geologic formation through which the
pollution passes.
(d) The volume of the aquifer affected by the
pollution.
(e) The direction and velocity of the pollution
movement.
(f) The methods for monitoring the pollution.
(g) The number and locations of the monitoring
facilities.
(h) The frequency of the monitoring activities.
9-3
-------
(i) The key indicator parameters that can be used
to detect the pollution.
(j) The costs of the monitoring methods.
Step 5 — Select and Implement the Monitoring Program
With the alternative monitoring approaches defined
for the key sources and causes of pollution, the moni-
toring program can be selected and implemented. The
selection process involves a series of decisions as to
which monitoring activities can be accomplished given
the groundwater situation within an area, the available
budget, and the goals of the monitoring program.
Development of the monitoring program requires
that decisions be made regarding:
(a) Monitoring of pollution sources and causes.
From the monitoring techniques identified" in Step 8,
certain high priority sources and causes with monitor-
ing deficiencies can be selected for monitoring.
(b) Background monitoring. If an area has sub-
stantial regions without groundwater quality data, it
may be desirable to establish a skeletal monitoring
network to fill these gaps.
(c) Monitoring for pollution control. Where
efforts are underway to control pollution from specific
sources, monitoring may be useful to evaluate the
effectiveness of the control measures.
(d) Audit monitoring. In areas where a substantial
portion of the monitoring is conducted by local public
agencies or by private companies, it may be necessary to
provide for occasional audit monitoring to verify
results.
(e) Data management for monitoring. The final
decision concerns the data management system to be
employed for the storage and retrieval of monitoring
data. The selection of a system for a given area
requires consideration of several factors. One is the
volume of data to be handled by the system. A second is
the use and distribution that will be made of the data.
Third, the compatibility of a system with others that
may already be in use must be considered. Fourth,
decisions may already have been made at a higher admin-
istrative level, such as by a State, which will specify
the monitoring system to be employed. And fifth, the
cost of the system is an important factor.
After the decisions outlined in the above five
categories have been made, the monitoring program can be
implemented. The principal action in establishing a
monitoring program for an area by a DMA will be to
install supplemental monitoring stations to augment the
existing system. This will involve letting contracts
for services and working with waste disposers to assist
them in developing monitoring activities.
An ongoing activity for a DMA will be the review
and approval of monitoring procedures for new potential
sources of groundwater pollution. In fact, with time
this will become the primary activity of a DMA as the
monitoring program for existing sources and causes is
established and operating on a routine basis.
Step 10 — Review and Interpret Monitoring Results
A key function of a DMA after the implementation of
a monitoring program is underway will be to collect and
to review all current monitoring data in its area. The
data should be analyzed and interpreted so as to define
quality trends, new pollution problems, regions of
improvement, and effectiveness of pollution control
activities. Assessments such as those portions of the
groundwater resource not meeting water-quality stan-
dards and predictions of future quality under projected
population and land-use conditions should be prepared.
The responsibility of analyzing monitoring data to
convert it to water-quality information will be a con-
tinuing activity for a DMA. As new data are received,
they should be studied promptly in order to detect
changes rapidly, particularly those requiring immediate
attention or action.
St££L. LLr Summarize and Transmit Monitoring Information
The final result of a monitoring program organized
in an area by a DMA is information on groundwater qual-
ity. Thus, the final task of a DMA is to disseminate
the information gained in usable forms to the agencies
and organizations concerned with such information.
Monitoring should be summarized in appropriate
forms for convenient study before distribution outside
of the DMA. This may involve preparation of tables
showing averages and/or changes in water quality. Sim-
ilarly, graphs prepared to readily display long-term
trends may be helpful. Maps showing, for example,
locations of major known sources of pollutions, area!
distributions of concentrations of key pollutants, and
regions having groundwater with qualities not meeting
drinking water standards, can also be drawn.
Monitoring information should be distributed
regularly to appropriate public agencies - local,
State, and Federal. Major industries in the area
should also receive the material as well as cooperating
agencies and organizations that contribute monitoring
data. In addition, the general public should be
informed of the results of the monitoring program;
therefore, reports should be sent to local newspapers,
citizens' groups, chambers of commerce, and conserva-
tion and environmental organizations. It can be
expected that the public awareness of groundwater qual-
ity created by such publicity efforts will indirectly
act to encourage individuals and organizations to pre-
serve and to protect the underground resources which
they, perhaps for the first time, more fully under-
stand.
Finally, a DMA has the responsibility to alert
action and enforcement agencies of critical problems or
situations which are discovered as a result of the mon-
itoring program. This may involve, for example, detec-
tion of hazardous or toxic pollutants which could immi-
nently affect a nearby municipal well field. Prompt
reporting of such instances is essential as well as
following up with specialized monitoring efforts for
documenting and controlling emergency situations.
-------
MONITORING GROUNDWATER POLLUTION
Kenneth I). Schmidt
Groundwater Quality Consultant
Fresno, California
Monitoring groundwater pollution entails source moni-
toring, evaluation of infiltration potential, and monitoring in
the topsoil, vadose zone, and aquifer. Soil, geologic mat-
erials, water, and solid and liquid wastes can be sampled.
The emphasis of groundwater pollution monitoring is water
in the aquifer, and the primary sampling tool is the pumped
well. The usefulness of high-capacity wells pumped for long
time periods has been demonstrated in the assessment of re-
gional groundwater pollution.
Introduction
The monitoring of groundwater pollution encompasses a
number of techniques, including the sampling of water wells.
Since most sources of pollution originate at or near the land
surface and most groundwater samples are taken from wells,
the entire soil-groundwater system must he analyzed. Po-
tential pollutants applied at the land surface in many cases
must pass through topsoil and the vadose, zone before reach-
ing the water table. The recharged pollutants then ordinar-
ily move through the aquifer a certain distance before they
are sampled. Thus groundwater pollution monitoring depends
heavily on a knowledge of soil and hydrogeologic factors that
control pollutant movement. An understanding of the natural
factors affecting groundwater quality is necessary in order
for pollutant monitoring to be effective.
There are five basic parts of the system: I) land sur-
face, 2) topsoil, 3) vadose zone, 4) aquifer, and 5) well.
Considerable confusion exists in some groundwater quality
investigations due to a lack of understanding of the pertinent
physical, chemical, biological, soil, and geologic factors.
Sanitary engineers, surface water hydrologists, and regu-
latory agencies have often focused attention almost exclu-
sively on the land surface. Soils scientists have usually
evaluated the topsoil and secondarily the vadose zone.
Groundwater hydrologists and geologists have usually anal-
yzed the portion of the system below the water table. Water
supply and regulatory agencies have focused on well sampling
in areas where groundwater is an important source of supply.
The entire sequence of events must often be interpreted in
specific situations in order to successfully monitor ground-
water pollution.
Groundwater pollution monitoring encompasses moni-
toring at the land surface (source monitoring), in the topsoil
and vadose zone, and in the aquifer. A number of tools are
available to monitor potential pollutants in each portion of
the system. As the parts of the system are closely inter-
related, the successful development of a monitoring program
largely depends on selection of the appropriate mix of tools
for the system. Land surface monitoring includes land use
surveys, inventories of waste amounts, and sampling of
liquids and solids. Significant storage capacity for potential
pollutants may exist in the vadose zone, and some pollutants
can be significantly attenuated in passage through the topsoil
and vadose zone. In some eases, pollutants introduced at
the land surface may travel through great thicknesses of
geologic materials above the water table. Long travel times
in the vadose zone, such as decades or centuries, may ren-
der aquifer monitoring ineffective.
The heart of groundwater pollution monitoring usually
lies in the aquifer, as this is where water is ultimately
pumped from wells for use at the land surface. Water sam-
pling of wells is a key Item, and along with source monitor-
ing is the primary technique for monitoring groundwater
pollution. Past groundwater pollution monitoring has dem-
onstrated the effectiveness of well sampling; however, a
considerable amount of experience and judgment is necessary
in most cases.
Source Monitoring
Sources of groundwater pollution can be conveniently
separated into 1) point, 2) line, and 3) diffuse. These sub-
divisions generally correspond to the sources of groundwater
recharge. The impact of a specific waste discharge on
groundwater quality depends on the areal extent and configu-
ration of the discharge at the land surface. Thus, the type
of applicable monitoring program differs considerably from
a point source to a diffuse source. Common point sources
are solid waste disposal sites, percolation ponds for liquid
wastes, solid wastes in stockpiles, leaking tanks, and con-
fined animal wastes. X,ine sources include polluted stream-
flow, wastes leaking from sewers and pipelines, liquid
wastes disposed of to dry stream beds, and materials used
for highway deicing. Diffuse sources include agricultural
return flow, polluted precipitation, septic tanks, and lawn
fertilizers.
Potential pollutants in each source can be characterized
as: 1) physical, 2) inorganic chemical, 3) organic chemical,
4) bacteriological, or 5) radiological. Physical parameters
of most importance are density and temperature. Inorganic
chemical constituents include the major constituents com-
monly fovmd in groundwater, other constituents, and trace
elements. Organic chemicals include pesticides, hydrocar-
bons, detergents, and other organic compounds. Bacterio-
logical constituents include bacteria, viruses, and pathogenic
organisms. 1 Methods of analysis have been discussed in
several comprehensive works. 2' ^
Types of Waste Disposal
There are a variety of types of waste disposal which
may affect groundwater. These include storage basins and
other ponds containing liquids which may escape the land sur-
face and percolate to the groundwater. The stockpiling of
solid materials from which pollutants may be leached is
another type. Irrigation and disposal to normally dry stream
channels are widely used types of waste disposal in the west-
ern U. S. Seepage trenches, percolation pits, seepage pits,
and wells are also used for waste disposal. The method of
disposal determines if a portion of the soil-groundwater sys-
tem is bypassed. For example, percolation pits and seepage
pits may allow pollutants to bypass the topsoil and part of the
vadose zone. Wells may permit pollutants to bypass the top-
soil and vadose zone and directly enter the aquifer.
Monitoring Techniques
Non-sampling techniques, at the land surface include a
wasteload inventory, pipeline and tank tests, testing of arti-
ficial liners, land use surveys, and aerial surveillance. An
inventory of waste loads includes collection and tabulation of
data on the volume of liquid wastes and weights of solid
wastes and their compositions. Data on the physical, chemi-
cal, bacteriological, and radiological characteristics of the
wastes should be collected. Field surveys, land use surveys,
and aerial surveillance can he used to determine sources of
pollution and methods of disposal. Testing for leakage is an
important monitoring procedure, since many of the pollutants
in tanks, pipelines, and lined ponds may present significant
pollution potential.
Sampling of polluted surface water bodies, wastewater,
and solid wastes is an integral part of a groundwater pollution
monitoring program. Rivers, lakes, estuaries, and canals
which are polluted and may recharge the groundwater should
be sampled. Surface water sampling will be conducted
through federally mandated monitoring programs.0 1-lem'
•and Brown, Skougstad, and Fishman2 have also discussed
I
9-4
-------
surface water sampling. The most important type of land
sin face monitoring will generally be wastewater sampling.
1 7
The American Public Health Association , Wells , American
Society for Testing Materials'", and the U. S. Environmental
8
Protection Agency have discussed wastewater sampling.
There are two basic categories of wastewater sampling in
relation to groundwater pollution: 1) discharge stream, anil
2) pond. The sampling of pond waters subject to percolation
is the more direct approach, but is generally more difficult
avid time-consuming. Consideration should be given to
groundwater recharge in selection of sampling sites. Sam-
pling of pond water often involves either the construction of
special walkways or the use of boats for sample retrieval.
In some ponds where water is being recycled, collection de-
vices allow for relatively easy sampling. Many solid wastes
can be categorized as to chemical composition on the basis of
past experience and references in the literature. Occasional
sampling is necessary, for example, for trace element and
radiological determinations on phosphate ore stockpiled for
use in fertilizer manufacture.
[evaluation of Infiltration
One of the key parts of a groundwater pollution moni-
toring program is the determination of the amount of waste-
water which can leave the land surface and potentially per-
colate to the water table. This amount is related to the
method of waste disposal used. For some methods, such as
wells and seepage pits, the topsoil and significant parts of
the vadose zone are bypassed. In this situation, the volume
of wastewater potentially reaching the water table is equiva-
lent to the volume of waste discharge. Factors ope rati vc in
the topsoil, such as evapotranspiration, must, be considered
for waste disposal in percolation ponds, dry stream channels,
and many diffuse sources. The major methods of disposal
discussed in the following sections are: 1) percolation ponds,
2) irrigation, 3) dry stream beds, and 4) seepage trenches.
Percolation Ponds
In this report, "percolation pond" is used to signify any
pond which permits significant movement of pollutants into
the underground. The water budget method can be used to
calculate seepage from percolation ponds. Waste discharge
and precipitation are summed as the input,, and percolation
calculated as the residual between input and evaporation.
Storage changes in the pond must also be considered. Evap-
oration from free water surfaces can be determined from
<) io
measurements for land pans or floating pans.' ' Factors
such as temperature and salinity affect both evaporation and
percolation rates. Stable isotopes have great potential for
11
application in the determination of pond evaporation.
Field measurements using lysimeters can be used to deter-
mine evaporation from soil surfaces.
Irrigation
The water budget method is also used to evaluate return
flow from applied irrigation water. Precipitation and applied
water are summed as the input, and evapotranspiration cal-
culated as part of the output. Return flow is then calculated
as the residual between input and evapotranspiration. Evap-
otranspiration in irrigated areas can be determined by a
number of methods. These methods are generally based
on several climatological parameters, and most have been
developed for application in certain geographical areas.
Applied water can be determined from records of surface
water deliveries and groundwater pumpage. "Irrigation
efficiency" is often defined as the evapotranspiration divided
by the applied water. When wastes are being disposed of by
irrigation, the waste discharge is added as part of the input;
in some cases it may comprise ail of the applied irrigation
water.
Pry Stream Reds
In many parts of the southwest, significant volumes of
wastes are disposed of into normally dry stream channels.
The part of the water that does not evaporate or leave the
area as runoff can infiltrate and eventually reach the ground-
water. Mining wastes and sewage effluent in Arizona and
oilfield wastes in California are commonly disposed of by
this method. Streamflow records at different gaging stations,
combined with records of precipitation, evaporation, and
streamflow diversions, can be used to calculate percolation.
In this case both evaporation from the free water surface
and wetted channel must be considered.
Seepage Trenches
Effluent from septic tanks and other wastes is common-
ly disposed of in trenches several feet beneath the land sur-
face. As such, these wastes are usually emplaced in the
topsoil and are subject to evapotranspiration. Determination
of evapotranspiration loss of effluent is often difficult in this
case because of widely varying soils conditions, septic tank
disposal practices, and other factors. Rooting depth of
plants growing in the area is a key factor in plant uptake of
effluent from trenches. Precipitation is combined with
waste discharge as the input, evapotranspiration is part of
the output, and percolation is calculated as the residual. In
this case evapotranspiration is the sum of evaporation from
bare soil surfaces and transpiration by crops, lawns, or
trees growing in the disposal field area. If there is irriga-
tion, this item must be added to the input.
Monitoring in the Topsoil and Vadose Zone
Monitoring the movement of pollutants in the topsoil
requires consideration of 1) soil water, 2) soil physics,
:i) soil chemistry, 4) microbiology, and !>) physical chemis-
try. Historically soil scientists were more concerned with
the soil conditions affecting growing crops than with the
quality of percolating waters. Today this situation is rapidly
changing and many studies are being conducted concerning
percolate quality. An extensive body of literature is avail-
able on water quality changes in the topsoil, particularly in
journals such as Journal of Environmental Quality and Soil
Science Society of America Proceedings. To date much of
this information has not been used in groundwater quality
studies.
The vadose zone has commonly been approached by the
"black box" method, but recent years have shown significant
advances in research, particularly by soils scientists. The
vadose zone in many areas is not comprised of soil in the
sense of the topsoil of agricultural workers. More often the
materials comprising the vadose zone are geologic forma-
tions with little or no soil development. One of the major
constraints in past groundwater pollution monitoring has been
the difficulty in determination of travel times of water from
the land surface to the water table. The approximate rate
of travel in the vadose zone must be known in order to moni-
tor groundwater pollution. If the travel time approaches
decades or centuries, there is little use in detailed, frequent
monitoring of aquifer water quality.
Travel Time
The travel time of recharged wastes through the vadose
zone is intimately related to the storage capacity and mois-
ture characteristics of the vadose zone. In parts of the
southwest, where water tables may be greater than 500 feet
in depth and the geologic materials are moisture deficient
due to lack of recharge, considerable amounts of water may
be added at the land surface before percolation to the water
table occurs. The portion of the porosity known as specific
retention must be satisfied before downward percolation can
occur in the vadose zone. In other areas where precipitation
or irrigation has supplied abundant recharge for decades or
centuries, the specific retention is generally already satis-
fied. In the case of relatively shallow water levels, there
9-4
-------
m;iy be relatively rapid percolation through the vadose zone.
In some cases, isotopes and other tracers may provide an
indication of the rate of travel in the vadose zone. 11
Attenuation Capacity
Attenuation capacity for pollutants introduced at the
land surface is due to physical and other factors. Pollutant
attenuation commonly occurs due to the following processes:
dilution, filtration, sorption, buffering, precipitation, oxi-
dation and reduction, volatilization, biological assimilation
and degradation, and radioactive decay. Recharge iron)
precipitation, streami'low1, canals, and othei sources can lead
to dilution of polluted water in the vadose zone. The capacity
for attenuation of many potential pollutants due to other pro-
cesses is often limited by the nature of the geologic materials
present. This is especially true for many organic chemicals,
trace elements, and radionuclides. This limited capacity is
in sharp contrast to the almost unlimited ability of many un-
consolidated materials to remove bacteriological pollutants.
In general the mobility of the various potential pollutants
from the land surface to the water table is highly variable.
The bacteriological constituents are least mobile and the
major inorganic chemical constituents are the most mobile.
Trace elements, organic chemicals, and radionuclides
generally have an intermediate mobility.
Monitoring Techniques
A number of techniques have been developed for moni-
toring water movement and water quality in the topsoil.
Monitoring in the vadose zone will require extension of tech-
nology developed in the topsoil and in the aquifer. Generally,
in a homogenous porous media, percolating water will pass
vertically through the. vadose zone. However, in a hetero-
genous, stratified material, such as most of the alluvial
deposits of the southwest, percolating water may become
perched above layers of low permeability. In tins situation,
lateral movement at substantial rates can occur above the
water table.1-1
Monitoring in the topsoil and vadose zone generally
requires retrieval of soil, geologic materials, and water.
Ordinarily water in the vadose zone occurs in the unsaturated
state, and pressure head is negative. Special sampling de-
vices are necessary to retrieve water samples, as this
water will not enter wells or holes open in the vadose zone.
However, local saturated zones are present in the vadose
zone as perched groundwater. Augers can be used for much
of the sampling of soils and geologic materials, although
standard water well drilling rig's may be necessary for deeper
sampling. Water content can be determined from core sam-
ples or in-place by neutron logging.1<> Ileeently the techni-
que has been used to monitor water storage in the vadose
zone and to estimate flow rates.
Tensiometers are used to measure soil water pressures
during unsaturated flow. A tensiometer consists of a porous
ceramic cup cemented to a rigid plastic tube. Negative pres-
sure is measured by means of a mercury manometer, when
properly enplaced in the soil, the pores in the cup will form
a continuum with the pores in the soil. Electrical resistance
blocks are also used to measure negative pressures. The
blocks consist of electrodes embedded in porous materials,
such as fiberglass. The principle of operation of these
blocks is that the negative pressure within the blocks res-
ponds to the suction of the soil with which the blocks are in
intimate contact. The tensiometer is generally the most
effective device for monitoring water flow to negative pres-
sures of about -1 atmosphere. In characterizing water
movement, a batteryof tensiometers must be installed, with
units terminating at successive depths in the vadose zone, ft
is necessary to install more than one battery to detect hori-
zontal flow. To measure negative pressures below about
-I atmosphere, the thermocouple psychrometer has been
developed.17 In-place negative pressure measurements are
possible with this tool down to -.'10 atmospheres. The basic
principle involved is that a relation exists between negative
pressure and relative humidity.
Monitoring in the topsoil and vadose zone includes
sampling oi soils and geologic materials. Important physi-
cal parameters include texture, bulk density, porosity,
water content, and hydraulic conductivity. Development of
a soil-water chemical characterization curve, relating
water content and pressure, is helpful in evaluation of water
movement. Analyses include soluble ions, cation exchange
capacity, exchangeable ions, pH, electrical conductivity,
and base saturation. Many constituents are often determined
on the saturation extract, because this can be generally re-
lated to field moisture conditions. Chemical analysis of
soil and geologic materials is extremely valuable where
large amounts of potential pollutants are retained above the
1 H
water table.
Water samples may be obtained from unsaturated re-
gions by the use of the tensiometer or suction cup. In this
case a suction is applied to bring the water sample to the
land surface. Suction cups are the best available tool for
obtaining water samples in the unsaturated parts of the va-
dose zone. However, the cups are operative only at nega-
tive pressures greater than about -I. 0 atmosphere. Because
of the small openings in the ceramic cups, certain bacterio-
logical and organic constituents may be filtered out. Han-
1')
son and Harris ' discussed the limitations of this tool in
monitoring in the vadose zone. Where water is perched in
the vadose zone, samples may be collected as in the aquifer,
for example, by bailing or pumping.
Ileeently salinity sensors have been developed for in-
20
place measurement of soil salinity. The principle of
operation of these devices is the relation between electrical
conductivity and total dissolved solids in the water. Elec-
trodes embedded in a porous ceramic plate in hydraulic
continuity with water in the vadose zone can be used to dir-
ectly measure the electrical conductivity of the water. The
temperature of the water must be measured carefully when
using this tool, and the sensors cannot be used at negative
pressures of less than -2 atmospheres.
Aquifer Monitoring
Monitoring groundwater quality below the water table
requires a knowledge of 1) hydrogeology, 2) well hydraulics,
.1) geochemistry and physical chemistry. Since most re-
charge passes through the topsoil and vadose zone, ground-
water quality evaluations must include consideration of the
full sequence of events above the water table. Of principal
importance in understanding groundwater quality is the
development of the hydrogeologie framework. The hydro-
geologic framework is important from two standpoints:
1) control on groundwater and pollutant movement, and
2) geoehemical considerations. Geochemical processes
impart natural characteristics to groundwater quality. To
date little of the knowledge from geochemistry or physical
chemistry has been applied in groundwater pollution studies.
An extensive body of literature is available on natural
groundwater quality, especially with regard to carbonates
and some silicates.l' Well sampling for groundwater pollu-
tion monitoring has little value unless well hydraulics are
thoroughly considered. A number of examples of aquifer
pollution monitoring were presented at the first two National
Ground Water Quality Symposiums and the 1974 National
Water Well Association Technical Division Education Pro-
gram, and subsequently published in Ground Water.
Attenuation Factors
Many of the processes which occur in the topsoil and
vadose zone occur below the water table, but to a different
degree. Oxygen poor waters are common in aquifers and
the geologic materials comprising aquifers are often devoid
of organic matter. Some geologic formations lack many of
3
9-4
-------
the common substrates for sorption. Because of these fac-
tors, if the topsoii and vadosc scone are bypassed during
waste disposal, the pollution potential is preatly increased.
However, once pollutants reach the water (able, they are
subject to a physical attenuation in most aquifer systems.
This attenuation results in the commonly observed phenom-
ena that point sources produce only local polluted zones in
the groundwater. The attenuation in most cases is deter-
mined by the following factors:
1) The volume of wastewater reaching the water table
per unit time
2) The concentration of pollutants reaching the water
table
3) The waste loading, i.e. the volume or weight of
pollutant reaching the water table per unit area of
the water table
4) Areal hydraulic head distribution
r>) Transmissivity of aquifer materials
(!) Vertical head gradients and vertical permeabilities
through confining beds, if present
7) Quality of native groundwater and recharge sources
8) Extent of recharge reaching the water table froxn
other sources at the land surface
9) The pumping of wellr, and well construction
In some eases, certain factors do not attenuate pol-
lutant concentrations, but act in an opposite fashion. An
example is the presence of a large depression cone in an
agricultural area due to irrigation well pumpage, whereby
pollutants are drawn into the area from many directions,
and are prohibited from leaving.
Non-sampling Techniques
There are several techniques for monitoring ground-
water pollution in the aquifer which do not involve sampling.
Water-level measurements and pump tests are usually
necessary in specific areas to evaluate the hydraulic head
distribution and aquifer parameters. Test drilling may be
necessary to adequately define the subsurface geological
conditions. Surface resistivity surveys can sometimes be
effectively used to delineate the extent of a polluted zone in
the aquifer.^ However, in general the polluted water must
be shallow, and the subsurface geology must be well defined.
Borehole geophysics can be used to determine the electrical
conductivity of water in the aquifer at specific depths below
22
land surface. If the chemical nature of water in an aqui-
fer is known from chemical analysis, resistivity logs can
be used to determine the approximate quantities of the ions
present.
Sampling Techniques
Sampling in the aquifer includes geologic sampling and
water sampling. Solid materials comprising the aquifer are
ordinarily sampled only in eases where significant amounts
of pollutants are retained in this zone. This generally
occurs in cases of a thin vadose zone or where the topsoii
and vadose zone are bypassed during well disposal. In un-
consolidated deposits the finer-grained deposits are gener-
ally of most interest. The texture, type of clay, cation
exchange capacity, and pH are important parameters. For
some rock formations, information on mineral and trace
element and radionuclide content may be required. Geologic
sampling can generally be done only during well drilling.
The method of drilling highly influences the usefulness of the
sampling.
Water samples can be collected from aquifers in sev-
eral ways. In cases of shallow water level, tile drains
beneath agricultural areas may provide suitable sampling
sites, however, often they tap only the upper portion of the
aquifer. Springs represent groundwater discharge, and
along with baseflow in streams can provide integrated sam-
ples of groundwater quality over large areas. For most
common situations, however, wells are required for aquifer
sampling. In cases where existing wells do not suffice, it
may be necessary to construct special monitor wells.
Well Sampling
Water well sampling is somewhat of an art, and the
usefulness of the sample often depends on the experience,
perception and hydrogeologic judgment of the person taking
the sample or supervising the program. The composition
of a water sample from a well is influenced by well location
and construction, pumping time, recharge sources, and the
three-dimensional distribution of groundwater quality.
Wells near point or line sources can also show a rapid res-
ponse to waste disposal operations. Time trends in and the
spatial distribution of groundwater quality must be consid-
ered in a monitoring program. There are a number of
reasons) for sampling water, and each may require a differ-
ent approach. For example, the main concern may be the
quality for subsequent use of the water. For drinking water,
a sample might be taken from the household tap rather than
from the well. For irrigation purposes, a sample could be
taken of water in a ditch at the entrance to a field. For
groundwater pollution monitoring, the water sample would
ideally be taken directly from the aquifer. However, for
practical purposes, a sample from the well discharge will
be collected in most cases.
A number of techniques can be used to obtain water
samples from open wells. Portable submersible pumps are
useful for a number of monitor wells near a point source
and when only periodic pumping is required. Water can be
air-lifted from a well by forcing compressed air down a
pipe lowered into the well. Water samples can also be taken
by bailing. A weighted bottle or short section of pipe capped
at the bottom may be lowered into small diameter wells in
order to collect water samples. Thief samplers have been
developed which can be closed at a specific depth inside the
well, thus allowing water quality sampling at different
depths. Mechanical or inflatable packers can also be instal-
led which temporarily isolate selected water-bearing zones
for pumping.
Well pumping is generally much more effective than
sampling at individual points in the aquifer. With high-
capacity wells and long pumping times, significant amounts
of water can be withdrawn. In general, low-capacity wells
with small pumps should be used to evaluate local conditions.
For example, the effect of an individual septic tank in a
rural area could be monitored by a nearby domestic well.
However, to evaluate regional conditions, such as pollution
due to diffuse sources, high-capacity wells with large pumps
may be more effective. A high-capacity well pumped for
several weeks or months can often provide an integrated
sample that is impossible to duplicate by other methods.
High-capacity wells can integrate vertical and areal varia-
tions in water quality. In some cases an integrated sample
is the most important; in others an integrated sample is
purposefully not desired.
Well sampling can also be used occasionally to monitor
the vadose zone. For example, in wells with perforations
above the water table, perched water may enter the well and
"cascade" to the water level. In some agricultural areas
where irrigation wells are inactive for several months of the
year, cascading water accumulates near the well during this
period. If a water sample is collected during the first few
hours of pumping such a well at the beginning of the irriga-
tion season, it may provide a valuable integrated sample of
return flow. In any case, the importance of wells cannot be
overstressed, as they act to supply source water for various
uses.
Spatial Water Quality Patterns
Zones of polluted groundwater have a three-dimensional
configuration. Large-capacity wells have been used to de-
lineate the areal distribution of polluted zones. Nitrate
patterns in groundwater near Tucson, Arizona determined
from sampling of irrigation wells were directly related to
land disposal of sewage effluent. ^ There is evidence that
nearly every well must be sampled to clearly delineate some
sources. For example, the quality of water pumped from
9-4
-------
irrigation wells southwest of Fresno, California appears to
reflect the influence of a leaking trunk sewer between the
24
urban area and sewage treatment plant. In this case
chlorides are an excellent tracer due to low background
contents. The extent of the polluted zones from recharge of
sewage effluent at the Fresno Sewage Plant has also been
delineated by sampling large-capacity irrigation wells during
24
periods of maximum pumpage. The background areal
distribution of groundwater quality must be defined prior to
monitoring pollution.
The vertical distribution of groundwater quality is often
difficult to assess. Water samples obtained during the
drilling of wells by the cable-tool method can provide valu-
able information on the vertical stratification of water quality
in the aquifer. Open-bottomed wells, which have no perfor-
ations or openings other than at the bottom of the casing, can
provide valuable data on water quality at specific depths in
the aquifer. Such high-capacity municipal wells in unsew-
ered parts of the Fresno area have provided sound data on
25
groundwater pollution from septic tanks. The open bottom
type of well construction merits attention in the case of
monitor wells. As well yield is not as important as in the
case of water supply wells, there is less need for special-
ized well screens, gravel packs, and other features devoted
to enhance production.
Time Trends
If a well penetrates an aquifer of rather uniform areal
and vertical water quality, the composition of the water may
not change much over long periods of time. However, in
the case of vertical stratification of groundwater quality,
which is common in areas of groundwater pollution, signifi-
cant changes can occur in water quality with time. Just as
water levels in a well change rapidly soon after the pump is
turned on, so can the quality of pumped water. Rarely have
water samples for chemical analyses been taken frequently
enough to adequately describe these short-term changes.
The result has often been widely varying analytical results
from one sample to the next. Just as water levels are mea-
sured on a logarithmic frequency during aquifer tests, water
samples can be collected simultaneously. That is, samples
can be taken on a more frequent basis during the early part
of the pumping period. Frequent nitrate analyses during
pump tests on high-capacity wells in the Fresno urban area
have illustrated exponential decreases in concentration with
pumping time. If water samples are to be collected from
high-capacity wells after only minutes or hours of pumping,
short-term time trends must be considered. In the case of
high-capacity wells which have been inactive for several
weeks or months, several days or weeks of continuous pump-
ing may be necessary to purge the well of atypical aquifer
water. This water may have short-circuited down the side
of the casing through perforations or the gravel pack.
Seasonal changes in water quality can also be significant
for high-capacity wells near pollution sources. Such changes
have been noted for nitrate and chloride contents during mon-
thly sampling of irrigation wells at sewage effluent disposal
23
sites near Tucson, Arizona. Wells near sewage effluent
percolation ponds near Fresno, California have shown sig-
nificant seasonal variations in nitrate, ammonia, and
pH. 2!5 Iligh-capacity municipal wells in unsewered parts of
the Fresno urban area have shown significant seasonal vari-
ations in nitrate and chloride. In all cases the seasonal
changes are in some way related to waste disposal operations.
Seasonal fluctuations must be established before effective
monitoring programs can be implemented. In an analagous
fashion, seasonal fluctuations in depth to water were com-
monly monitored in the early days of water-level measure-
ment programs.
Once the short-term and seasonal time trends are es-
tablished in an area for individual wells or groups of wells,
long-term trends in groundwater quality can be established.
In the case of diffuse sources, decades of such records will
often be necessary to adequately monitor groundwater pol-
lution. Long-term records in many areas have documented
the development of groundwater pollution sources. These
records can also bo used to illustrate the ameliorating
effects of groundwater pollution abatement activities.
Conclusions
Monitoring groundwater pollution entails source moni-
toring, evaluation of infiltration potential, monitoring in the
topsoil and vadose zone, and aquifer monitoring. The heart
of groundwater pollution monitoring lies in water well sam-
pling. Source monitoring includes inventoring and sampling
solid and liquid wastes which may eventually percolate to the
water table. Waste disposal methods partly determine infil-
tration potential and whether or not portions of the soil-
groundwater system are bypassed. Pollutants can be cate-
gorized into physical, inorganic chemical, organic chemical,
bacteriological, and radiological. In general the major
inorganic constituents are the most mobile in soils-ground-
water systems, whereas the bacteriological constituents are
the least mobile.
Monitoring in the vadose zone is especially important
in areas where significant amounts of pollutants are retained
in this zone. In arid areas with deep water tables and little
recharge, a large storage capacity for pollutants may exist
above the water table. In some cases recharged waters may
take decades or centuries to reach the water table. However,
in most humid areas and beneath lands irrigated for long
periods of time, travel times are often in the order of
weeks, months, or several years. However, significant
retention of trace elements and other constituents may still
occur. Water budgets and moisture characteristics in the
vadose zone provide information on travel time above the
water table. Both water and soil can be taken from the va-
dose zone and analyzed. Sampling in the vadose zone has
been most widespread and effective beneath point sources of
pollution. This is because relatively small amounts of mat-
erials are sampled compared to those in the entire system.
Monitoring in the aquifer can occasionally be done by
surface geophysics, borehole geophysics and other methods.
However, most cases require water well sampling. If
existing wells are not sufficient, specially designed moni-
toring wells may have to be constructed. There are a num-
ber of ways to retrieve water samples from open wells, and
geologic samples are usually taken during drilling. Pumped
high-capacity wells can provide meaningful water samples
for many sources of pollution. Pumped wells can also
occasionally provide direct information on the quality of
water in the vadose zone. Time trends and the three-
dimensional distribution of groundwater quality must be
evaluated to successfully monitor aquifer pollution. Care-
ful hydrogeologic judgment must be used in the establishment
of monitoring programs for groundwater pollution.
References
1. American Public Health Association, 1971, "Standard
Methods for the Examination of Water and Wastewater",
13th Edition, 874 p.
2. Brown, E., Skougstad, M.W., and M.J. Fishman,
1970, "Methods for Collection and Analysis of Water
Samples for Dissolved Minerals and Gases", Techniques
of Water-Resources Investigations of the U. S. Geo-
logical Survey, Book 5, Laboratory Analysis, Chapter
Al, 160 p.
3. American Society for Testing Materials, 1973, "ASTM
Standards", Part 23, Water.
4. U. S. Environmental Protection Agency, 1974, "Methods
for Chemical Analyses of Water and Wastes", National
Environmental Research Center, Cincinnati, Ohio, 298 p.
5. U. S. Environmental Protection Agency, 1974, "Water
Quality and Pollutant Source Monitoring, Proposed
Rules", Federal Register, vol. 39, no. 1(58, pp. 31500-
31505.
0. Hem, J.D., 1970, "Study and Interpretation of the
5
9-4
-------
Chemical Characteristics of Natural Water", U. S.
Geological Survey Water-Supply Paper 1473, Second
Edition, 303 p.
7. Wells, 1). B., 1975, "Monitoring Industrial Waste-
water", Deeds & Data, Water Pollution Control Fed-
eration, June issue, 8 p.
S. U. S. Environmental Protection Agency, 1971, "Hand-
book for Monitoring Industrial Wastewater", Office
of Technology Transfer.
9. Kohler, M.A., 1954, "Water-Loss Investigations:
Lake Hefner Studies, Technical Keport", U. S.
Geological Survey Professional Paper 269.
10. Harbeck, (5.1*:., Jr., Kohler, M. A., Koberg, G. E. ,
and others, 1058, "Water-Loss Investigations: Lake
Mead Studies", [J. S. Geological Survey Professional
Paper 298.
11. American Geophysical Union, 1967, "Isotope Techni-
ques in the Hydro!ogic Cycle", Geophysical Mono-
graph Series No. 11, edited by G. K. Stout, American
Geophysical Union, Washington, D. C. , 199 p.
12. Van Hylckaraa, T.E.A., 1968, "Water Level Fluctu-
ations in Evapotranspirometers", Water Resources
Research, vol. 4, no. 4, pp. 7(11-768.
13. Cruff, R. W. , andT.H. Thompson, 1967, "A Compar-
ison of Methods of Estimating Potential Evapotrans-
piration from Climatological Data in Arid and Subhumid
Environments", U. S. Geological Survey Water-Supply
Paper 1339-M, 28 p.
14. Black, C.A. and others, 1965, "Methods of Soil Anal-
ysis", Agronomy Monograph No. 9, American Society
of Agronomy, Madison, Wisconsin.
15. Wilson, L. G. , 1971, "Observations on Water Content
Changes in Stratified Sediments During Pit Itecharge",
Ground Water, vol. 9, no. 3, pp. 29-40.
IB. Van Bavel, C. H. M., Nixon, P. R. , and V. L. Hauser,
1963, "Soil Moisture Measurement with the Neutron
Method", U. S. Dept. of Agriculture, Agricultural
Research Service Publ. 41-70, 39 p.
17. Watson, K.K., 1967, "A Recording Field Tensiometer
with Rapid Response Characteristics", Journal of
Hydrology, vol. 5, pp. 33-39.
18. Walker, W. H. , 1974, "Monitoring Toxic Chemical
Pollution from Land Disposal Sites in Humid Regions",
Ground Water, vol. 12, no. 4, pp. 213-218.
19. Hanson, E.A., and A. R. Harris, 1975, "Validity of
Soil-Water Samples Collected with Porous Ceramic
Cups", Soil Science Society of America Proceedings,
vol. 39, no. 3, pp. 525-536.
20. Richards, L.A., 1966, "A Soil Salinity Sensor of
Improved Design", Soil Science Society of America
Proceedings, vol. 30, pp. 333-337.
21. Zohdy, A. A., Eaton, G. P., and D. R. Mabey, 1974,
"Application of Surface Geophysics to Ground-Water
Investigations", Techniques of Water-Resources Inves-
tigations of the U. S. Geological Survey, Book 2, Chap-
ter Dl, Collection of Environmental Data, 116 p.
22. Brown, D. L. , 1971, "Techniques for Quality of Water
Interpretation from Calibrated Geophysical Logs, At-
lantic Coastal Area", Ground Water, vol. 9, no. 4,
pp. 25-38.
23. Schmidt, K.D., 1972, "Groundwater Contamination in
the Cortaro Area, Pima County, Arizona", in Hydro-
logy and Water Resources in Arizona and the Southwest,
vol. 2, Arizona Section AWRA, pp. 95-111.
24. Schmidt, K.D., 1975, "Regional Sewering and Ground-
water Quality in the Southern San Joaquin Valley",
Water Resources Bulletin, vol. 11, no. 3, pp. 514-525.
25. Schmidt, K.D., 1972, "Nitrate in Groundwater of the
Fresno-Clovis Metropolitan Area, California", Ground
Water, vol. 10, no. 1, pp. 50-64.
6
9-4
-------
MANAGEMENT OF GROUNDWATER QUALITY DATA
Norman F. Hampton
Commonwealth Associates Inc.
Jackson, Michigan
SUMMARY
The growing concern for subsurface water resources will surely
be accompanied by an expanding groundwater data base, a data base
which is already quite large. This paper is intended to point the way
towards the efficient management of this data base which will assure
that pertinent information is available when and where it is needed.
The discussion presented here wilJ describe the requirement.', of
groundwater data management and survey some of the available
capabilities which may serve to satisfy these requirements.
DATA MANAGEMENT REQUIREMENTS
The identification of data management requirements is one
component of the management information system (MIS)
development process which also entails systerfl design, organizational
design, system procedures design, programming, implementation,
testing, debugging, documenting, and training. It is hoped that the
discussion herein will convey to the groundwater quality manager the
scope and breadth of the field of information management systems
and that it will expose him to the alternatives available to him in
structuring an information management capability suitable to his
needs.
A requirements analysis must address information content, data
collection, data communication, data organization and storage, data
processing, and data retrieval.
Information Content
A comprehensive groundwater MIS must be capable of
maintaining the following types of data:
• Station descriptions
• Water quality criteria
• Geologic information
• Hydrologic information
• Water quality parameter identifiers
• Water quality measurements
• Temporal information
• Information qualification data
• Monitoring agency status data
Information indexing
Station Descriptions: Station descriptive data consists of
information v\Tt ich specifies the station type (i.e., pumped well,
unpumped well, unsaturated zone, ambient monitoring, compliance
monitoring, etc.), the party responsible for monitoring the station, a
unique station identifier code, a unique location (three-dimensional),
a/id directions for locating the station in the field. With the
exception of the last item, all of this information should be
searchable. Information providing instructions for locating stations in
the field can be stored as narrative text along with other special
station specific information which is not required for retrievals or
computations (e.g., oil lubricated well subject to bearing leakage,
continuous-slot stainless steel well screen, etc.).
Water Quality Criteria: Information pertaining to established
quality criteria which a groundwater quality MIS should
accommodate as station specific data include current and projected
land use, current and projected water use, demographic data,
economic data, designated protected water uses, applicable permit
data (compliance dates and monitoring requirements — parameters
and frequency), and water quality criteria (either ambient or
discharge limitations).
Geologic Information: In order to uniquely identify the source
of groundwater samples, some geologic data is required, in addition
to geographical coordinates, to specify the water bearing material
(aquifer, unsaturated zone, or topsoil) from which the sample
originated. In the case where a monitoring station taps more than
one aquifer, aquifer identification is particularly essential and must
be provided as sample specific (i.e., input in conjunction with each
set of water quality analysis data) rather than station specific data.
The requirement for providing aquifer identification can be satisfied
by storing the established aquifer name, if available, or else the
geologic formation name and age associated with the monitored
aquifer. It should be pointed our that the latter form of
identification is not preferred since aquifers and geologic formations
do not necessarily coincide completely. Aquifer identification can be
codified and standardized, and search operations facilitated by
application of U.S.G.S. proposed modifications to the stratigraphic
coding system developed by the American Association of Petroleum
Geologists.^ Additionally, information regarding the physical
properties and chemical constituency of the water bearing materials
may be necessary, particularly if the synergistic effects between these
materials and introduced pollutants arc to be modeled.
Hydrologic Information: Most hydrologic information will be
station specific and can, therefore, be stored concurrently with the
establishment of station descriptions in the data base. In cases where
many stations penetrate the same homogenous medium, it may be
possible to store the characteristics of that medium under only one
station together with a list of the other stations common to that
medium. Major hydrological data elements will include the
following:
• Water bearing material depth, thickness, and area! extent
• Permeability
• Aquifer transmissivity and storage coefficient
1
9-5
-------
• Hydraulic gradient (vector)
• Water tabic elevation (sample specific)
» Area and magnitude of natural and artificial recharge and
discharge
• Station sampling device (e.g., pumped well, suction
lysimeter, neutron probe, etc.) operating characteristics
Water Quality Parameter Identifiers: Because of the large
number of candidate water quality variables involved in groundwater
monitoring, a generalized water quality parameter identification
scheme is appropriate. This would require that data identification be
independent of the MIS program logic, that is, the data descriptions
must themselves be data inputs to the system. Consequently, the list
of water quality parameters maintained can be virtually open ended.
Water quality parameter identifiers should be codified and
system specific. Since water quality parameter identifiers are system
specific, the system administrator should have responsibility lor
depositing and maintaining this type of data in the groundwater
MIS. An individual monitoring agency can establish a special para-
meter identifier by petitioning the system administrator who will
judge the validity, redundance, and applicability of the new para-
meter before including it in the data base.
Water Quality Measurements: The results of physical and
chemical analyses of groundwater, soil, and geologic material samples
can be stored as water quality measurement data which will represent
the bulk of the information to be managed by the groundwater MIS.
This information will be required for both retrieval and
computational operations and must, therefore, be stored as
searchable data. Each measurement data element is, of course,
measurement specific and must be stored in conjunction with
information which specifies the parameter measured (parameter
code), the sample analyzed (sample date), and the station sampled
(station identifier code).
Temporal Information: In order to provide reasonable utility,
a water quality information system must be capable of reflecting
trends. This would require maintaining water quality data as time
series. Water quality data updates need, therefore, to be appending
operations rather than destructive updates. When data is not col-
lected at constant frequency, which is most often the case with
groundwater monitoring, the date of sampling must be recorded as
sample specific data with each new set of water quality measure-
ments which is input. In situations where significant vertical strati-
fication of water chemistry is present it will also be necessary to
record and store the pumping time in hours prior to the collection
of cither a simple grab or composite sample. Additionally, in the
case of composite samples taken over time, it will be necessary to
record and store the duration (in hours) of the composite sampling
period.
Information Qualification Data: There is a cogent need for a
data qualification capability in any water quality monitoring
information system. To accomplish this the system should include a
comprehensive edit function, preferably computerized, which would
operate prior to data storage. The edit check can be based on
comparison of input data with previous trends, allowable data ranges,
and established units of measure. Data failing any one of these
checks should not be modified but rather flagged and reported as
suspect. The capability to compare input data with allowable ranges
imposes an additional data requirement which can be satisfied by
storing these ranges as station specific, searchable data.
Improvements in the value of a data base can also be attained
by allowing "reliability indicators" to be input and stored as
nonsearchable data. These indicators could be of the type that
reflect, for example, station performance anomalies, unusual
sampling conditions, unusual methods of measurement, measurement
precision, or which reflect qualitative judgments of the "goodness"
of data. Reliability indicators should be stored either as station
specitic (in the narrative text), sample specific (as a water quality
measurement) or measurement specific (in a special field) as
appropriate.
Monitoring Agency Status Data: A groundwater quality
monitoring program may involve the periodic inspection of facilities
to determine the "operational status" of monitoring programs and
equipment. In addition, where an agency has groundwater pollution
control functions, the "readiness status" of a control unit in terms of
its ability to respond to a pollution incident may also be evaluated.
Consequently, a comprehensive groundwater MIS should be capable
of maintaining this type of information. A monitoring agency or
pollution control unit would be regarded as a station by the MIS, an
inspection tour regarded as a sampling iteration, and status data as
water quality measurements with parameter codes being established
accordingly.
The operational status of a monitoring agency will be estimated
based upon its ability to monitor the stations, parameters, and at the
frequencies required. A "readiness index" could be formulated which
would reflect the ability of a pollution control unit to respond to a
pollution incident in a timely manner. This index would be a
function of personnel on hand, personnel training, equipment on
hand, and equipment reliability.
Information Indexing: Information indexing allows ready
access to abstracts of existing data sets. The groundwater quality MIS
should provide indexing of two major categories of data sets: water
quality data files present in the MIS data bank, and groundwater
research documentation. Information contained in the water quality
file abstracts will be station specific and would include parameters
monitored, monitoring frequency, and period of record. Research
documentation indexing requires the storage of document titles,
author names, report numbers, performing and sponsoring
organizations, report dates, textual abstracts, key words, and
geographical areas of interest all of which should be maintained as
searchable data.
Data Collection
Data collection, in the context of MIS design, is the process of
translating information into machine readable form. The primary
factors considered in selecting data collection systems are purchase
cost, operating cost, reliability, responsiveness, and minimizing the
bottleneck created by relatively high internal computer processing
speeds and low input speeds.
Total MIS expenditures are particularly sensitive to data
collection costs since data entry typically accounts for 20 to 40
percent of electronic data processing costs. ^ In addition, the data
entry process represents the single greatest source of error in a MIS.
There is a wide variety of available capabilities which will
provide automated support of the data collection phase of a
computerized MIS. These include conventional keypunch, buffered
keypunch, key-to-tape, key-to-disc, remote "dumb" terminals,
remote "intelligent" terminals, mark sensing, magnetic ink character
recognition, and optical character recognition (OCR) devices. These
nine options are listed more 01 less in order of increasing
implementation cost and, correspondingly, increasing speed and
reliability. The devices listed all have applicability to groundwater
data entry. Selection of equipment by each groundwater data
depositor will depend primarily upon the magnitude of data flow
generated by his monitoring activity.
2
9-5
-------
An additional category of devices available to the data
depositor, which has particularly attractive applicability to
groundwater monitoring is source data automation. Source data
automation is the process of capturing primary data iti machine
readable form. Examples of such equipment are automatic digital
recorders used in conjunction with Keck groundwater level recorders,
automatic laboratory chemical analysis- equipment and robot water
quality monitoring stations. The advantages of source data
automation are that it produces data which is easily converted into
other rriach ine-useable form, reduces the opportunities for
introducing errors, and lowers clerical costs.
Data Communications
User interaction with a management information system can be
segregated into four major activities: 1 j file creation; 2) file updating;
3) information requests; and 4) information reception. Computerized
management information systems accomplish these functions in one
of two modes: 1) batch; or 2) realtime/interactive.
User access to the groundwater surveillance data base should be
in the batch mode whereas access to the information index system
component should be in the real-time mode, at least for retrievals.
Although user interaction with a batch processing system allows
optional use of telecommunication links with the system,
telecommunication is mandatory for real-time processing.
A telecommunication link requires a terminal to enter data,
modems to encode (in a form acceptable to the transmission
channel) and decode data, and a transmission channel. Transmission
channels can be ordinary telephone services such as provided by
WATS (best suited to widespread, high volume data flow), dial-up
service such as provided by TWX or TELEX (best suited to
widespread, low volume data flow which is likely to be the case for
groundwater surveillance), or dedicated private line (best suited to
high volume data flow concentrated between a few points).3 The
major factors to be considered in the selection ol a transmission
service will be responsiveness, reliability, and implementation and
operating costs. Data security is not likely to be a significant
consideration in the development of a groundwater MIS.
Data Storage
The development of the data storage component of a MIS
entails the selection or fabrication of hardware devices, data
organization schemes, and data base management software packages.
Factors to be considered inclvidc cost, storage space, response time
and current and future use of information stored.
Three general classifications of hardware are available for data
storage; internal, secondary, and external. Internal storage is best
utilized for holding programs and data being immediately executed.
Internal storage media include magnetic core, thin films, magnetic
rods, and plated wire devices — all of which are characterized by high
access speeds and costs. Secondary storage is not an integral part of
but is directly connected (on-line) to the central processing unit
(CPU). Secondary storage devices include magnetic disc, drum, card,
and tape peripherals characterized by moderate access speeds and
costs. External storage is not directly connected (off-line) to the
CPU, External media include removable disc pacts, magnetic tape,
punched cards, and paper tapes all characterized by low access speeds
and costs.^
Data files arc structured using one or a combination of three
basic organizational concepts: sequential, random, and list.
Sequential files store records iti a specified sequence relative to other
records so that the next logical record is also the next physical
record. Sequential organization permits rapid access of a series of
records logically related to one another but is cumbersome for
updating and retrieving individual records out of sequence.
Random organization requires the establishment of a
predictable relationship between a record key and the direct address
ot the location where the record is stored. In most cases this will
require a "dictionary look up" process as part of each record
retrieval. Random access allows rapid retrieval of individual records
or data items where only a small portion of the data file is affected
but is not well suited to retrievals of multiple records. Random
organization is quite appropriate for files containing groundwater
measurements for which station identifiers can be used as record
keys.
List structures (simple, inverted or ring) incorporate pointers in
each record which point to other records that are logically related to
the first record. Of particular applicability to the management of
groundwater data are inverted list structures which make every data
element available as a record key. For instance, a station type code
could be used as the key to a record which contained pointers to
every station of that type. The inverted list approach allows very
rapid (and therelore inexpensive) retrievals, but requires a great deal
of storage and does not foster easy file updates. Therefore, inverted
list structures can best be used for files which are frequently accessed
but infrequently updated. Inverted list structuring can be very well
applied to the design of groundwater data files containing station
descriptive data which is used to identify stations with particular
characteristics.
Data Processing
Computerized data processing is accomplished either in batches
or on a continuous (real-time) basis. Batch processing requires the
accumulation and preprocessing of a group of transactions all of
which will be computer processed at one time. Real-time processing,
on the other hand, accepts and processes transactions as they occur.
Both processing modes can accept input data from either remote or
local terminals. The basic difference between the two processing
methodologies, as seen by the system user, is the difference in
response time with the turn-around time for real time processing
being significantly quicker.
Real-time processing should be implemented only where rapid
system response is really needed since batch processing permits more
efficient and economical hardware utilization by requiring less
system redundancy. Therefore, only accession to the groundwater
information indexing components of the groundwater MIS requires
real-time processing. This requirement is imposed by the users need
to interact intellectually (browse) with the information indexing data
base.
Batch processing associated with access to the Groundwater
Data File will be composed of editing, sorting- storing, retrieving, and
statistical operations. Input editing will examine input data for
format errors, check the validity of codes, (parameter codes, aquifer
codes, etc.) and compare water quality data with acceptable ranges.
For compliance monitoring, the input editing module can also be
used to compare water quality data with established water quality
standards and prepare violation reports as necessary. The sorting and
storing processes organize the data and update the appropriate files.
The retrieval commands access the appropriate data files, organize
the requested information, and format output reports. The statistical
processor would function in conjunction with the retrieval routines
to operate on taw data as designated by the information requestor.
The statistical processor would be required to generate extreme
values, first and second moments, regression and correlation
coefficients, logarithms, daily loading (for source monitoring), and
coordinates necessary to create plots.
3
9-5
-------
TABI.i; 1
Existing Information Management Systems
System Name
Storage and
Retrieval System
National Water Data
Storage and
Retrieval System
ORSANCO Robot
Monitor System
National Water
Data Exchange
Environmental
Data Index
Acronym
STORLT
WATSTORE
Administrator
USEPA
NAWDEX
ENDEX
Remote Control RECON
System
General Information GIPSY
Processing System
System 2000
Total
USGS
ORSANCO
USGS
NOAA
ERDA
University of
Oklahoma
MRi System
Corp.
Cine on
Systems, Inc.
Function
Surveillance
Data Storage
Surveillance
Data Storage
Surveillance
Data Storage
Data File
Index
Da (a File
Index
Document
Citations
Document
Citations
Data Base
Management
Data Base
Management
Informal ion
Water Quality
Water Quality
and Supply
Surface Water
Quality
Water Quality
and Supply
Environmental
Data
Energy and
Environmental
Water
Resources
Generalized
Generalized
Compatibility
IBM 371/58
I KM 370/155
IBM 1130
IBM 370/155
IBM 360/65
IBM 360/75
IBM 360/65
IBM 360/370
CDC 6000
UNIVAC 1106
IBM, CDC,
Honeywell
UNIVAC
Data Retrieval
Data retrieval is the process of translating information which is
meaningful only to machines into a form which is meaningful to
humans. Designing the data retrieval component of a MIS requires
identifying the information to be output, specifying the retrieval
procedures acceptable to the system, developing the required
retrieval software, determining output formats, and selecting
hardware.
The data retrieval component of the proposed MIS which
accesses the Groundwater Data File will be required to yield both
alphanumeric and pictorial output. The groundwater monitoring MIS
can offer the data user the most powerful capabilities if it can
provide a wide range of useful retrieval procedures. A retrieval
procedure is characterized by the information required by that
procedure as user input to the system to enable the system to locate
data and generate output. The user should be able to implement a
number of these procedures in conjunction with each other so that
Boolean retrieval strategies can be applied. In addition he should be
able to dictate, to some extent, the format of the output which he
receives.
Factors involved in the selection of data retrieval hardware
include considerations of speed, cost, flexibility, reliability, noise,
number of copies needed, and output format. Retrieval hardware can
be categorized according to the following distinctions:
• Impact, non-impact, cathode ray tube (CRT), digital
plotter, microfilm, or voice response
• Serial, which produces 10 to 200 characters per second
(cps) or parallel, which produces 300 to 10,000 cps^
• Full character or dot matrix
In general, impact printers produce full characters either one at
a time (serially) or a line at a time (parallel). Impact printers provide
good legibility and multiple copies (a constraining factor for many
applications) but are noisy and subject to relatively frequent
breakdowns.
Non-impact printers will, in most cases, best satisfy the
requirements of accessing groundwater monitoring data. Non-impact
printers can be either serial or parallel printers which, in a majority
of machines, produce dot matrix characters. Inkjet and electrostatic
printers are two types of non-impact printers which offer speed,
reliability, portability, competitive purchase cost, and quiet
operation. The disadvantages which are normally characteristic of
these devices are high operating costs (e.g., electrostatic printers
require special paper), the inability to produce multiple copies, and
slightly poorer image quality than is provided by impact printers.
CRT displays produce dot matrix characters, either serially or in
parallel, as well as graphics. Although CRT displays themselves are
unable to generate permanent records, they are fast, reliable, and
economical to purchase and operate. In addition, these devices afford
great flexibility by virtue of the optional peripheral equipment which
may be attached, such as hard copy output, light pens, and
information storage capabilities. CRT terminals would be most
appropriate for accessing groundwater information indexing files.
4
9-5
-------
Digital plotters which produce permanent graphic displays, are
available at a wide range of prices and, correspondingly, with a wide
range of capabilities. Microfilm systems can receive output directly
from a CPU, via either paper to film or CRT to film, and provide the
advantages of a compact, inexpensive, external storage medium.
Microfilm systems generate output in the form of microfilm
(normally 16 mm film), or microfiche (which records many pagesof
data on one frame of film).
With the exception of voice response units, which are used most
extensively by operations which interface with the public, any of the
above mentioned hardware options may find appropriate
applications in a groundwater monitoring program. The selection of
specific retrieval hardware components will depend upon the
requirements of individual data requestors and of interfacing with
the central system. The central system should be designed to be
flexible so that it represents a minimal constraint on the selection of
user output hardware,
EXISTING MANAGEMENT INFORMATION SYSTEMS
A comprehensive groundwater data management system is
composed of three major components which: maintain the data
generated by groundwater surveillance, index that data so that it can
be accessed expeditiously, and maintain concise citations of relevant
groundwater research documentation. In addition to the previously
discussed capabilities available for accomplishing these three
functions, there arc numerous existing management information
systems which may also be applicable. Table 1 presents an inventory
of existing data management capabilities (including generalized data
base management packages offered by commercial vendors as well as
developmental systems) which may provide the framework for
developing the desired capabilities.
The inventory of existing data management systems which is
presented here is not intended to be comprehensive. Rather, existing
systems were selected for inclusion on the basis of their significance
and relevance. Readers with particular interest in one of these
systems are referred to the associated users and systems
documentation.
REFERENCES
(1) Price, W. E., and C. H. Baker, Catalog of Aquifer Names and
Geologic Unit Codes Used by the Water Resources Division,
U.S. Department of the Interior, Geological Survey, Water
Resources Division, Reston, Va., 306 pages, 5974.
(2) Ferrara, R., and R. L. Nolan, "New Look at Computer Data
Entry," Journal of Systems Management, Association for
Systems Management, P 24-33, February, 1973.
(3) House, W. C., ed., Data Base Management, Mason and Lipscomb
Publishers, Inc., New York, New York, 470 pages, 1974.
(4) Lobel, Jerome, and M. V. Faritia, "Selecting Computer Memory
Devices," Automation, Penton Publishing Co., Cleveland, Ohio,
P 66-70, October 1970.
(5) Lorber, Matthew, "Evaluating Computer Output Printers,"
Automation, Penton Publishing Co., Cleveland, Ohio, P 64-67,
March 1972.
5
9-5
-------
GROUNDWATER POLLUIION MONITORING CASE STUDIES
L., G. Wi lson
Water Resources Research Center
The University of Arizona
Tucson, Arizona 85721
Summary
Investigators who were closely associated with
five groundwater monitoring programs were asked to
critically examine their studies as a guide to others
involved in similar projects. The particular ques-
tion to be answered was, "What monitoring techniques
should have or could have been implemented?" given
that time arid money were not constraints.
The case studies involved contamination of aqui-
fers from oil field brine disposal, plating waste
disposal, landfill leachate, nitrate from multiple
sources, and recharge from an oxidation pond.
Among tiie general recommendations of the inves-
tigators, resulting from the process of critical
evaluation of their associated projects, were the
following: establish interdisciplinary committees to
set up the monitoring program; maximize the density
of well network; use alternative methods to wells;
completely analyze the samples, including heavy
metals; thoroughly examine the hydrogeology of the
problem site; use tracers; develop predictive computer
models of the flow system; monitor in the zone of
aeration, where applicable; develop innovative
metnodologies; and continue monitoring until the prob-
lem is thoroughly quantified.
JiLlL?ductjon
Section 106 of trie Federal Water Pollution
Control Act stipulates that grants to State and
interstate agencies for pollution control programs
require trie establishment of procedures to monitor the
quality of groundwater. Methodology for monitoring iri
both tiie saturated zone (phreatic zone) and zone of
aeration (unsaturated zone, vadose zone) are presented
by Schmidt1 and by Meyer2. However, as pointed out by
Le Grand3, "Seldom will sufficient money and time be
available to completely delineate and monitor a con-
taminated area." Each monitoring program should
therefore be carefully planned at the outset to mini-
mize the effects of these two constraints {arid others)
on the yield of data. Often the experience of inves-
tigators who have been associated with similar pro-
grams will he 1 p in sucli planning, particulary if these
individuals are willing to discuss frankly not only
the merits, but also the limitations of their programs.
As an example of this approach, individuals
closely associated witu the monitoring programs of
five representative case studies were asked to crit-
ically examine their projects. They were requested
not only to summarize the objectives, methodology and
results, but also to indicate alternative monitoring
techniques or approaches which could have, or should
nave, been used if time, funds, or other factors, had
not been limiting. This paper is based on the
critiques of these individuals.
The case studies and related investigators were;
(1) Arkansas Brine Disposal, J. S. Fryberger'*,
Engineering Enterprises, Inc., Norman, Oklahoma; (2)
Plating Waste Pollution iri Long Island, New York, by
James J. Geraghty and N. M. Perlmutter5, Geraghty arid
Miller, Inc., Tampa, Florida; (3) Landfill Leachate
Pollution at an East Coast Location, by James J.
Geraghty and N. M. Perlmutter6; (4) Monitoring Nitrate
Pollution of Groundwater from Multiple Sources in the
Fresno-Clovis Motropolitan Area, by K. D. Schmidt7,
Fresno, nmJ (5) Arizona Oxidation Pond, by L, G.
Wilson'', University of Arizona, Tucson, Arizona.
Case Studies
Arkansas Brine Disposal
Details of this study were presented by
Fryberger'3. The pollution problem involved the con-
tamination of a shallow, alluvial aquifer during
disposal of oil-field brine in southwest Arkansas,
first, by recharge from an "evaporation" pond, and
later by leakage from a faulty disposal well. Verti-
cal distribution of alluvium consists of alluvial
clays and sand, overlying alluvial sand. The latter
unit is underlaid by sedimentary formations of Eocene
arid Cretaceous ayes. Static water table levels
average about eight ft below land surface. The
monitoring program related entirely to the saturated
zone. Monitoring facilities to supplement existing
pumping wells comprised 36 observation wells at 28
locations near the pollution source. At some loca-
tions several wells were installed at different depths
to obtain vertical salinity profiles. Observation
wells consisted of plastic well-screens attached to
2-1/2 inch pipe. Several wells were sampled at suc-
cessive depth increments during installation. The
following techniques were considered and rejected as
alternatives to permanent cased wells: surface resis-
tivity surveys, to estimate the lateral spread of the
pollution plume; and uncased rotary holes, coupled
with electric logs, to estimate the vertical distribu-
tion of salinity. Several construction alternatives
were considered: drive points; continuous flight auger;
mud/water rotary drilling; cable tool drilling; and
air-rotary drilling. The method which was eventually
selected entailed"- drilling with a continuous flight
auger through the overburden, then driving well points
to the desired depth in tiie sand aquifer. Brine and
well water were analyzed chemically to estimate mixing
of brine and groundwater. Analysis included evaluation
of botii major and minor constituents.
PIatiii£ Waste ContarnjnatjIon, Lqrij _Is 1 and._N_._Y ¦
Background information together with monitoring
procedures and results of this problem were reviewed
in detail by Perlmutter, e_t._al.10. Briefly, a pol-
lution plume, mainly comprising chromium, was monitored
in the vicinity of a former aircraft plant at South
Fariningdale. The source of the plume was a basin used
to dispose of plating wastes. A principal concern to
heal tti officials was that an auxi liary groundwater
supply at Massapequa would be contaminated by the
plume. The problem was first detected in samples from
a water supply well near the plant. Later a network
of sampling wells was installed.
The groundwater reservoir in the problem area
comprises 1300 ft of saturated consolidated deposits
on crystalline bedrock. The distribution of tne three
principal aquifer-s is: upper glacial aquifer, Magothy
aquifer, and Lloyd sand. Water table averages about 15
ft.
Tiie upper glacial aquifer formation has been
1
9-6
-------
primarily affected by contamination. In 1962, the
dimensions of the plume, detected by sampling were
4300 ft Iong, 1000 ft wide, and from land surface to
50-270 ft below land surface.
The monitoring program involved sampling entirely
within the saturated zone. Test wells, installed in
1949, consisted of 1-1/4 inch driven well points.
Water samples were obtained in five ft increments
during installation of the points. Drilling was con-
tinued to about 60 ft, i.e., until field sampling
produced an absence of chromium contamination. In
1962, 100 additional wells were constructed together
with several test holes for defining lithology and
hydraulic characteristics of the geologic formations.
Massapequa Creek and underlying formations were sampled
to evaluate heavy metal concentrations. A water bud-
get was conducted for the region. Patterns of flow of
groundwater into the stream were characterized.
Landfill Leachate Contamination, Milford, Connecticut
The third representative case study of ground-
water quality monitoring related to contamination by
landfill leachates. Specifically, the landfill site
investigated was in Milford, Connecticut. The purpose
of the study was to determine whether contamination of
groundwater and surface water by leachate would pre-
clude the development of a park on the site. A
monitoring program was established to evaluate chemi-
cal water quality and gas production and the patterns
of surface water and groundwater movement.
The landfill was constructed on a tidal marsh
bordering on Long Island Sound. Underlying the sur-
face landfill and marsh deposits are unconsolidated
deposits, 40 to 60 ft thick, consisting of glacial
till and outwash sediments. Glacial deposits are
underlain by consolidated bedrock of schist and gneiss.
The water table varies from an elevation at sea level
to about eight ft above sea level. Again, monitoring
was restricted entirely to the saturated zone.
Based on a water-budget evaluation, it was esti-
mated that the daily flow of precipitation into the
fill averages 80,000 gallons and that an equivalent
volume of leachate moves out of the fill. Several
hundred million gallons of groundwater have evidenced
contamination by leachate, which apparently is grad-
ually displacing underlying good quality water.
The monitoring program for the site involved an
initial review of background information, eg., topo-
graphy, vegetation, well data, rainfall, tides, etc.
Subsequently, seismic and electrical resistivity
techniques were employed, followed by installation of
36 test wells. Well depths ranged from 12 ft to 96
ft, and at several locations paired wells were used to
determine vertical head gradients.
Water levels and temperature profiles were mea-
sured in wells. Well water samples were collected for
chemical analysis, including a few determinations in
the field. Surface waters were also analyzed. Gas
samples were collected from shallow gas sampling tubes.
Studies were conducted by biologists to relate vegeta-
tive stress to the groundwater system.
Nitrate Pollution, Fresno-Clovis Metropolitan Area
The fourth monitoring project entailed estimating
the extent, sources and time trends in nitrate pollu-
tion of groundwater underlying a 145 square mile area
of the San Joaquin Valley, California. The area is
predominantly urban, although the surrounding area is
agricultural. The principal sources of nitrate are
leakage from septic tank fields, land disposal of
sewage effluent, sewer line leakage; deep percolation
of fertilizers during irrigation; and recharge of meat
packing and winery wastes. Aquifers are located
within unconsol idated alluvium, locally thousands of
feet thick. Water levels average 70 ft.
The study was conducted to fulfill the require-
ments of a Ph.D. dissertation11>12.
Initial steps in the program involved collecting
background data on soils, groundwater, well construc-
tion and drillers logs, pollution sources, and
chemical analysis of sources and groundwater, particu-
larly relating to the areal distribution of nitrates.
Because of limited funding, monitoring was re-
stricted to the sampling of available wells, coupled
with the use of field kits for chemical analyses. To
reflect regional conditions, only high capacity (500
to 2500 gpm) municipal and irrigation wells were
sampled. Well construction was considered an important
parameter because of the vertical distribution of
nitrate with depth. To obtain the most representative
quality of regional groundwater, principal sampling
occurred during the warmest time of the year, the
period of maximum pumping in the majority of the high
capacity wells. According to Schmidt7, a key factor
in his study was to subdivide the area on the basis of
predominant nitrogen sources together with the hydro-
geology and nitrate areal distribution in groundwater.
Sampling of wells occurred on a weekly or monthly basis
near point source, compared to seasonal sampling near
diffuse sources. In addition to nitrate, the following
were determined: chloride, potassium, ammonium and
calcium.
Schmidt15 did not monitor in the zone of aeration,
but speculated that movement of water in this zone is
rapid because of well hydrograph response. He further
assumed that there is no gross uptake of nitrate during
flow in the zone of aeration.
Arizona Oxidation Pond
The fifth case study was concerned with monitoring
seepage during the filling of a new 10-acre oxidation
pond near Tucson, Arizona. In contrast to the above
case studies, an attempt was made in this study to
monitor in the zone of aeration as well as in the
saturated region. Particular emphasis was placed on
monitoring the movement of nitrate because groundwater
in the area has manifested an increase in this consti-
tuent. Details of the study were presented by Wilson,
et. al.13, and by Sma111k.
The site abuts the Santa Cruz River, an ephemeral
stream, and is immediately downstream of two principal
tributaries of the Santa Cruz River. Unconsolidated
alluvium underlies the area to great depth. Water
levels average about 70 ft in the vicinity of the pond.
Monitoring facilities were installed at the site
during a study on land disposal of oxidation pond
effluent15. The new pond encompassed several of these
facilities, which thereby functioned for two projects.
The facilities consisted of two 100 ft deep access
wells, two PVC wells, suction cups and an irrigation
well. The access wells were used with a neutron
moisture logger for monitoring water content changes
in the zone of aeration. Each access well contained a
screened well point for sampling in the vicinity of the
water table. Depths of the PVC wells, 40 ft and 60 ft,
were based on neutron probe data showing that perched
water tables (mounds) develop in the zone of aeration
at these depths during percolation of applied surface
water. Two batteries of suction cup samplers (suction
lysimeters) were installed to permit sampling within
the soil zone down to five ft. A nearby 300 ft deep
irrigation well was used to sample an extensive region
of the saturated zone.
During filling of the pond, water was metered
2
9-6
-------
into the pond and the seepage rate was estimated.
Analysis of pond water and groundwater samples includ-
ed the entire nitrogen sequence, together with major
constituents and coliform organisms. Project results
illustrated that indigenous nitrogen, or soil nitrogen
concentrated during past irrigation and land disposal
operations, contributed a high initial peak of nitrate
which exceeded the nitrate levels which could have
been derived from pond effluent. Furthermore, deni-
trification and/or ammonium sorption effected a
control on nitrate levels.
trospective Alternativ£s
and Recommendations for
SJmi 1 ar Honitoring Studies
Each of the investigators of the above monitor-
ing case studies were, to a large extent, successful
in characterizing their groundwater contamination
problem. Nevertheless, each individual believed that
if money and time had not been constraints, additional
or alternative procedures could have, or should have
been implemented. Many of these procedures have rele-
vance only to the specific case study. Others were of
a general nature, being listed by several or all of
the investigators.
The general procedures may be regarded as recom-
mendations for other investigators to bear in mind
during establishment of monitoring programs. The more
noteworthy of these recommended procedures are sum-
marized below:
1. If possible, establish at the outset an inter-
disciplinary committee to ensure collection of inter-
related quality parameters. As an example, for the
Arizona pond study, Wilson8 indicated that such a com-
mittee could have included a soil chemist, soil
physicist, sanitary engineer, aquatic biologist, and
hydrogeologi st.
2. Obtain a thorough understanding of the hydrogeo-
logy of the project area from examination of available
reports, drillers logs, etc. Particular emphasis
should be placed on defining groundwater flow direc-
tion and vertical and lateral gradients. The latter
information is useful in determining whether the
groundwater system in the vicinity of the pollution
source is recharging or discharging. A hydro-chemical
balance should be attempted. Along the same line,
establish base-line water quality data and water
levels in the region of the pollution source. Also
obtain information on well construction.
3. Ensure that the network of sampling wells is as
dense as possible in the vicinity of the pollution
source. Where feasible, use wells for multiple pur-
poses: eg., sampling and neutron logging. Be certain
that each pollution source is defined.
4. In addition to wells, use alternative techniques
for defining the confines of the pollution plume. For
example, several of the individuals reporting on the
above case studies, suggested using surface resisti-
vity techniques. Stollar and Roux'6 outline method-
ology and limitations of earth resistivity techniques
for defining groundwater contamination.
analyzed in the field (eg. for pH, CO2, HC03, etc.).
Check that cations and anions balance in complete
analyses and prepare water analysis diagrams.
7. Trace the flow of the pollution plume using spe-
cific tracers (eg. nitrogen isotopes in nitrate prob-
lem areas) or indigenous tracers (eg. chloride).
8. Consider using available, or specially constructed,
digital models of the flow system. Available models
include the finite difference model of Prickett and
Lonnquist17; or the finite element model of Pinder and
Frind18.
9. When possible, monitor in the zone of aeration as
well as in the saturated (phreatic) zone, particularly
when the former is of considerable thickness and
perching layers may develop. Core or drill samples
should be taken in this zone and characterized for
particle size distribution, cation exchange capacity,
etc. An example of an investigation which involved
monitoring in the zone of aeration is that of Apgar
and Langmuir19, in which suction cups were extensively
employed.
10. Develop, when necessary, innovative or specialized
monitoring techniques or well construction methods
(see for example, Yare20).
11. Monitoring should be continued until the problem
is quantified. Several of the individuals reporting
the above case studies regretted that their projects
had to be shut down before long-term trends could be
defined.
References
1. Schmidt, K. D., 1975; Monitoring groundwater
pollution. Paper presented at the International
Conference on Environmental Sensing and Assess-
ment, Las Vegas, Nevada, Sept. 14-19.
2. Meyer, C. F., ed., 1973; Polluted Groundwater:
Some Effects, Controls and Monitoring. G. E.
TEMPO Report pTepared for"tTie Environmental
Protection Agency, EPA-600/4-73-0016.
3. Le Grand, H. E., 1972; Monitoring of changes in
quality of ground water. In Water Qua 1 ity in a
Stressed Environment, ed. by WTTTTettyjohn^
BurgesTTub. Co., pp 122-129.
4. Fryberger, J. S., 1975; Arkansas brine disposal.
In Report 5, Monitoring Ground-Water Quality:
H jus trative EximpTes , prepared T^^gTTE^'TEW'O
for THe~Environmental" Protection Agency (In
Press).
5. Geraghty, J. J., and N. M. Perlmutter, 1975;
Plating waste contamination in Long Island, New
York. In Report 5, Monitoring Ground-Water Qual-
ity: Illustrative Examples", prepared by G. E^
TEMPO for the Environmental Protection Agency (In
Press).
6. Geraghty, J. J., and N. M. Perlmutter, 1975;
Landfill leachate contamination in Mil ford, Con-
necticut. In Report 5, Monitoring Ground-Water
Quality: Illustrative Examples, prepared"by~"G~7~E.
TEMFOror the Environmental Protection Agency (In
Press).
7.
5. Establish a thorough regular sampling program in
wells and surface water supplies. If available, use
large capacity wells to obtain more representative
samples.
6. Analysis of samples should be as complete as
feasible, including not only determination of major
constituents, but also heavy metals, and organic
toxins. If germane to the project, microbial concen-
trations (eg. total coliform, fecal coliform) should
also be evaluated. When possible, samples should be
Schmidt, K. D., 1975; Monitoring nitrate pollu-
tion of groundwater from multiple sources in the
Fresno-Clovis Metropolitan area. In Report 5,
Monitoring Ground-Water Quality: Illustrative
ExampTes,Vrepare3~6y""G. E. TEm) forHFie Envir-
onmentaT Protection Agency (In Press).
3
9-6
-------
8. Wilson, L. G., Arizona pond study. In Report 5,
Monitoring Ground-Water Quality: Illustrative
ExanipTis~, prepared by G7~ET~TEMPl5~7or_tlTe~Envir-
onmental" Protection Agency (In Press).
9. Fryberger, J. S., 1972; Rehabilitation of a^
Brine-Polluted Aquifer. ^TPA^Y^^0T4TTfrwiron-
mentaTTrotect'ion Agency, Technology Series.
10. Perlmutter, N. M., M. Lieber, and H. L.
Frauenthal, 1963; Movement of waterborne cadmium
and hexavalent chromium wastes in South Farming-
dale, Nassau County, Long Island, New York. In
Short Papers in Geology and Hydrology: U. S.
GeoToqTcal Survey Prof. ^per 475-C, pp~C179-C184.
11. Schmidt, K. D., 1971; The Distribution of Nitrate
in Ground Water in the Fresn^cTWjT~MetropoTTTaF
AreaT^arTTIoTquTrri/al 1 ey ."TaTTTornTiT Unpub- ~
TTsKedThTBT DissertationTTRe" University of
Arizona.
12. Schmidt, K. 0., 1972; Nitrate in groundwater in
the Fresno-Clovis metropolitan area, California,
Ground Water, 10, pp 50-64.
13. Wilson, L. G., W. L. Clark, and G. G. Small,
1973; Subsurface transformations during the
initiation of a new stabilization lagoon. Water
Resources Bulletin, 9(2), pp 243-257.
14. Small, G. 6., 1973; Groundwater Recharge and
Quality Transformations~puring the Initiation and
Managenient~~of aIjievrTtaFTTTzatTon Lagoon. Un^~
publishid~M7s. Thesis, University of Arizona,
Tucson.
15. Wilson, L. G., and G. S. Lehman, 1967; Reclaiming
sewage effluent, Progressive Agriculture in
Arizona, 19(4), pp~22-247~~~^
16. Stollar, R. L., and P. Roux, 1975; Earth resisti-
vity surveys - a method for defining ground-water
contamination. Ground Water, 13(2), pp 145-150.
17. Prickett, T. A., and C. G. Lonnquist, 1971;
Selected Digital Computer Techniques for Ground-
Water Resource FvaTuation. Illinois State Water
Sur7ey7~Bu1 letin
18. Pinder, G. F., and E. 0. Frind, 1972; Application
of Galerkin's procedure to aquifer analysis.
Water Resources Research, 8(1), pp 108-120.
19. Apgar, M. A., and D. Langmuir, 1971; Ground-water
pollution potential of a landfill above a water
table. Ground Water, 9(6), pp 76-94.
20. Vare, B. S., 1975; The use of a specialized dril-
ling and ground-water sampling technique for
delineation of hexavalent chromium contamination
in an unconfined aquifer, southern New Jersey
Coastal Plain. Ground Water, 13(2), pp 151-154.
Acknowledgments
The author gratefully acknowledges the assistance
of Mrs. Sarah Schuster in typing the final manuscript,
and Mr. Vincent Uhl for reviewing and contributing to
the original manuscript.
4
9-6
-------
MONITORING OF SUBSURFACE WASTE-DISPOSAL OPERATIONS;
A GENERALIZED SYSTEMS APPROACH WITH EXAMPLES
FROM CANADIAN CASE HISTORIES
FRANK SIMPSON
Associate Professor of Geology, University of Windsor,
Windsor, Ontario, Canada N5B 3P4
Summary
Injection of industrial wastes into subsurface
strata has been widely adopted in recent years, as a
means of isolating frequently noxious fluids in geolo-
gical settings, remote from the biosphere. A simple
systems model of a typical subsurface-disposal facility
comprises 3 main subsystems:
(1) an aataatoa, the disposal well, through which
materials (fluid wastes) and energy (injection pres-
sures, exothermic heat of chemical reactions) are di-
rected into subsurface reservoirs;
(2) -in]nation-e.cte,d Mb'-'.iA-ace, Apace, of the
disposal interval, in which the growing plume of wastes
is accommodated through structural adjustment of the
aquifer framework and displacement of formation fluids;
and
(3) a tnamducuJi or monitoring device, provid-
ing a continuous record of both waste behaviour in the
subsurface environment and response of the reservoir
strata to injection.
Monitoring provides the only means of evaluating
measures of effectiveness for subsurface-disposal ob-
jectives and usually takes the form of one or more of
the following:
(1) wellhead records of injection and annular
fluid pressures;
(2) shallow observation wells, drilled to near
surface aquifers;
(3) deep observation wells, penetrating the dis-
posal formation(s); and
(4) precise geodetic surveys to detect subsidence
or increase in strain rate near fracture zones.
The current status of monitoring in Canadian sub-
surface-disposal operations is reviewed with reference
to 74 waste-injection facilities initiated to date.
Introduction
Industrial operations, designed to develop natural
resources or convert them into energy and material
goods for the needs of society, inevitably generate
wastes, which are returned in various forms to the en-
vironment. In recent years, there has been growing
realization that the industrialized ecosystem has a
limited capacity to withstand disruption through pollu-
tion!. As a result, two contrasting, though complemerv-
tary goals in environmental management have arisen:
(1) reduction of unwanted industrial output (en-
ergy and materials) through improvement in technology
and recycling of wastes; and
(2) innovation in waste-disposal technique to
minimize pollution hazard by isolating wastes from the
environment with increased efficiency.
Injection of fluid, industrial wastes into subsurface
aquifers has been widely adopted in North America as a
means of attaining the latter aim.
Decisions, as to what constitutes a desirable
level of environmental quality and how this might be
best maintained through coordination of largely private
effort, are most efficiently made in the public sector.
The methodology of decision analysis can be a signifi-
cant aid to public decision makers, concerned with sub-
surface waste disposal, since rigorous step-by-step
analysis minimizes the danger of suboptimization, that
is, temporary solution of a disposal problem through
creation of a long-term pollution problem. The main
stages of decision analysis, arising in evaluation of
legislation against air pollution2, have application to
antipollution legislation generally and are rephrased
below to relate to subsurface disposal of industrial
wastes:
(1) structuring the decision problem by (a) for-
mulation of the problem, (b) identification of objec-
tives, and (c) selection of measures of effectiveness;
(2) describing possible consequences of alterna-
tive disposal programs in terms of the effectiveness
measures;
(3) prescribing relative preferences of the deci-
sion maker for each possible consequence, that is, pre-
cise identification of trade-offs between conflicting
objectives; and
(4) rationally synthesizing information from
stages 1 to 3 above, to select a disposal program.
In industrial waste-disposal situations, the prob-
lem involves determination of the merits of alternative
modes of disposal and evaluation of various disposal
formations, if deep-well inflection is one of the dis-
posal methods under consideration. A generalized ob-
jective, assumed to be desirable to both private and
public sectors, is that of maintaining high environmen-
tal quality. Conflicting sub-objectives arise, however,
where trade-offs must be made between reduction of dis-
posal-system costs and minimization of the risk of ad-
verse effects upon health and economy. The possible
adverse effects, of course, constitute undesirable out-
put of the disposal system and provide a basis for con-
sideration of effectiveness measures. In the disposal
program finally selected, measures of effectiveness are
determined by monitoring waste behaviour in the dis-
posal setting.
Surveys of deep-well injection of wastes in
Canada^' ® have been published by workers from
the Inland Waters Directorate of Environment Canada and
have been important in drawing attention to the environ-
mental significance of this mode of disposal at a time,
when little information had been coordinated for
Canadian provinces other than Ontario. Accounts of
deep-well disposal in particular Canadian provinces
dea1 with Ontario^" ® and the northern Williston basin
region (Saskatchewan and Manitoba)9, 10.11,12, 13^
Papers dealing with individual subsurface-disposal op-
erations are by MacLeod14 and de Korompay15.
The present account is a preliminary report on
some of the findings arising from an analysis of infor-
mation, currently being compiled on all Canadian sub-
surface-disposal systems. This continuing study places
particular emphasis on analysis of decisions, taken by
disposa1-wel1 operators. The purpose of this note is
to outline the present status of monitoring in Canadian
subsurface-disposal operations, involving deep-well in-
1
9-7
-------
jectiori of wastes.
Subsurface Disposal of Industrial Wastes in Canada
The first deep-well injection of fluid, industrial
wastes in Canada was in 1951 and involved disposal of
deslater water and spent caustic from the. Husky Oil
Operations Limited refinery at Lloydim'nster, Alberta,
into a sandstone aquifer at 2,140 feet below surface in
the Husky Refinery No. 2 well (Lsd 11-1-50-1W4). Two
more disposal wells were drilled by Husky near this lo-
cation intheearly 1950's. In 1958, the first of five
disposal wells was drilled at the Imperial Oil Limited
refinery at Sarnia, Ontario, for disposal of spent caus-
tic and phenolic water into carbonate strata at about
700 feet below surface. From the early 1960's onward,
deep-well disposal of industrial wastes gained wide ap-
plication in southwestern Ontario and the Prairie prov-
inces. To the end of 1974, some 74 subsurface-disposal
systems had been initiated in Canada. This figure does
not include wells, in which oilfield brines are returned
to producing formations in waterflood projects or in-
jected into strata, other than those yielding commer-
cial hydrocarbons, as disposal programs. Only 45 of
these wells are currently in operation, the remainder
having been suspended or abandoned.
Fluids injected into subsurface space in Canada
include "natural" waste and non-waste and "foreign"
waste categories, in the terminology of van Everdingen
and Freeze3. "Natural" fluids have constituents com-
monly found in the subsurface, though not necessarily
in the disposal formation, and all others are termed
"foreign". The principal categories of wastes, injected
into deep aquifers in Alberta, Saskatchewan, Manitoba
and southwestern Ontario, are:
(1) oilfield brines, injected during salt-water
disposal and pressure-maintenance operations;
(2) waste brines from shaft and solution mining,
as well as experimental solution of potash deposits;
(3) waste brines from solution mining of caverns
in halite for subsequent storage of liquefied petroleum
gases (LPG) and dry natural gas;
(4) refinery effluent, consisting mainly of sour
water, spent caustic and hydrocarbons; and
(5) chemical-plant wastes, arising from the manu-
facture of a wide variety of chemical products.
The type of waste to be injected has considerable
bearing on both the complexity of the disposal system
and the ultimate capital investment in the injection
operation10' 13. For example, the potash-brine dis-
posal wells of Saskatchewan are the most refined sys-
tems and the most costly (ultimate capital investments
in the range $120,000 to $418,000). They are complex
in design, which must take into account continuous op-
eration, high injection rates and high injection pres-
sures during a well life of 1 to 2 decades. These dis-
posal wells commonly involve directional drilling to
deep target formations below the potash ore bodies
(Middle Devonian Prairie Evaporite) currently being
mined, in order to minimize hazard to the mine workings.
Hole size may range from Vh inches in diameter over
the first 500 feet or so to 8 3/4 inches down to total
depth in a given well and costs of casing (lined) and
tubing strings and cementing the well to the surface
from the disposal formation are correspondingly high.
Mud programs must employ fluids, which do not react
with evaporite minerals, but will not damage the dis-
posal formation. Salt-cavern brine-disposal wells and
refinery and chemical-plant waste-disposal wells are
drilled without any comparable restriction with regard
to depth and commonly cost far less than potash brine
disposal wells, a study of the costs of deep-well dis-
posal of large volumes of waste brines, generated at
high rates by the potash industry and salt-cavern opera-
tions in Saskatchewan, suggests that alternative surface
storage would be at least 2 to 3 times more costly than
subsurface disposal.
The distribution of Canadian waste-injection wells
by province (wells in operation given in brackets) is:
20 in Alberta (15), 30 in Saskatchewan (21), 1 in
Manitoba (1), and 24 in southwestern Ontario (8). Of
the 74 wells considerd, 35 dispose of refinery wastes,
28 are brine-disposal wells, and 11 inject chemical
plant wastes of diverse origin. Brines from potash
mines and salt caverns in Saskatchewan account for 25
of the brine disposal wells. Disposal of refinery
wastes into the subsurface has been carried out mainly
in Alberta (16 wells) and southwestern Ontario (16 wells).
Of the 11 wells injecting chemical-plant wastes, 5 are
located in Ontario. Depths of injection intervals in
Canada range from a minimum of 600 feet in southwestern
Ontario to a maximum of 5,199 feet in southern Alberta.
Only 7 of the wells in Ontario have been drilled to dis-
posal depths greater than 900 feet, although it should
be noted that the shallow disposal wells of this area
have mostly been abandoned and only 2 wells injecting
wastes at depths of 800 to 850 feet are still in opera-
tion. The depth distribution of disposal intervals is:
16 wells with the intervals encountered at less than
1000 feet, 15 from 1000 to 2000 feet, 14 from 2000 to
3000 feet, 16 from 3000 to 4000 feet, 12 from 4000 to
5000 feet, and 4 from 5000 to 6000 feet. These figures
include 2 values for each of 4 potash-brine disposal
wells in Saskatchewan, injecting into 2 different dis-
posal intervals. In 41 cases (4 multizone completions
included) the disposal formation is largely composed of
limestone and dolomite, while in a further 28 instances
(4 multizone completions included), waste injection is
into sandstone aquifers. In 9 systems, wastes are per-
mitted to accumulate in caverns, formed by controlled
solution of evaporite strata.
Although nationwide coverage of data on injection
rates and pressures is as yet incomplete, the picture
emerging indicates that more than half of all Canadian
subsurface-disposal systems inject wastes at average
rates of less than or equal to 250 U.S. g/m and at
pressures of less than or equal to 500 psig. The maxi-
mum injection rate encountered was 1,100 U.S. g/m and
the maximum injection pressure was 1,750 psig; both
values relate to brine injection in Saskatchewan and
are from potash-mining and salt-cavern operations res-
pectively. It is noteworthy that the shallow subsur-
face-disposal operations of southwestern Ontario?' 8
have been characterized by low injection rates of less
than 100 U.S. g/m and low injection pressures, ranging
from atmospheric pressure to 450 psig.
Monitoring of Subsurface-Disposal Operations
Deep-well disposal of industrial wastes has been
described as "final" or "ultimate" and the process of
acceptance of a stream of waste fluids by a deep aqui-
fer has been likened to a storage situation. These
descriptions carry a fallacious connotation of perma-
nence and confinement of injected wastes in a predesig-
nated part of the subsurface environment, which are not
easily obtained in practice. Gravity-controlled move-
ment of formation waters in deep and shallow aquifers
alike takes place in the hydrogeologic environment of a
geological basin on scales varying from local to re-
gional, with the flow distribution determined largely
by regional topography, climate, basin geometry and
lithologic characteristics of the basin f-ji i 18, 19
Fluid wastes introduced into the hydrogeologic environ-
ment down disposal wells tend to assume the regional
vector properties of the ground-water regime, although
9-7
-------
the latter may be modified to a marked extent locally
by injection pressures of the disposal system.
Decisions to drill disposal wells and initiate
waste injection are frequently made under considerable
uncertainty, which is a function of the reconnaissance
level of subsurface information at most proposed dis-
posal sites. This uncertainty imposes limitations on
predictions, concerning waste movement in the subsurface,
but is significantly reduced by testing of the disposal
well and disposal interval and by monitoring of waste
injection operations. The utilization of monitoring
records in mathematical models of subsurface flow, such
as that proposed by Freeze^O, would present the possi-
bility of using these models as prediction devices for
actual disposal situations. The significance of moni-
toring devices in waste-injection operations is dis-
cussed below with reference to a simple systems model
of a typical subsurface-disposal facility, comprising
3 main subsystems:
(1) an aetuatoi, the disposal well, through which
materials (fluid wastes) and energy (injection pressures,
exothermic heat of chemical reactions involving waste
components) are directd into subsurface reservoirs;
(2) ¦injzction-a()le.c.t^d ipace of the
disposal interval, in which the growing plume of wastes
in the vicinity of the disposal well is accommodated
through structural adjustment of the aquifer framework
and displacement of formation fluids; and
(3) a t>iaft4duceA or monitoring device, providing
a continuous record of both waste behaviour in the sub-
surface environment and response of the reservoir strata
to injection in addition to offering a means of checking
on the efficiency of the disposal well itself.
Monitoring of subsurface-disposal operations has
as its objective the observation of all responses to
waste injection. The entity "injection-affected subsur-
face space" is ideally restricted to the disposal inter-
val in the vicinity of where the latter is penetrated
by the disposal well. The degree, to which this norma-
tive situation is obtained, is a function of:
(1) well efficiency with particular reference to
the rates and pressures of injection favoured;
(2) compatibility of the waste components with the
rocks of the disposal interval and incorporated fluids,
as well as with each other;
(3) porosity-permeability gradients in the dis-
posal interval;
(4) the structural setting of the disposal unit;
and
(5) the dynamics of ground-water flow near the
disposal system and within the sedimentary basin as a
whole.
Monitoring of waste-inject ion operations thus involves
observation of responses to injection in both disposal
well equipment (surface and subsurface) and the subsur-
face environment generally and provides the only means
of establishing and evaluating measures of effectiveness
for the entire disposal operation. Thus the following
subobjectives of the monitoring subsystem are also mea-
sures of effectiveness for the subsurface-disposal sys-
tem:
(1) to detect variation in the rate of migration
of the injected wastes and leaks in the well equipment
by means of wellhead records of injection and annular
fluid pressures;
(2) to detect contamination of near-surface aqui-
fers by means of shallow obervation wells;
(3) to record changes in chemical composition of
fluids incorporated in the disposal interval and the
rate of flow of the waste stream, using monitor wells
drilled to the disposal interval; and
(4) to recognize increase in strain rate near sub-
surface fracture zones and changes in land elevation
near disposal systems with the aid of precise geodetic
surveys.
Record measurements of variation in injection pres-
sure, operating on a given waste stream of known injec-
tion rate, provide both an indication of the efficiency
of the disposal-wel1 subsystem and information on the
behaviour of the waste in the subsurface. For example,
a fall in pressure (and increase in injection rate) may
indicate leakeage of waste from the disposal well or ac-
celerated migration of the plume of injected waste within
or near the disposal interval on encountering rock of
improved reservoir quality. An increase in pressure
(and decrease in injection rate) usually suggests dete-
rioration of reservoir quality within the disposal in-
terval, as a result of clogging of pores by deposits,
representing either high suspended-solids content in the
waste stream or chemical reaction between wastes and
subsurface material of the disposal unit.
Clearly disposal wells in continuous use over long
time periods at moderate to high injection rates and
pressures, such as facilities serving potash mines and
salt-cavern operations, require careful monitoring of
pressure variation. Allan Potash Mines Operators Limited
in Saskatchewan, for example,employ a monitoring system
giving continuous records of pump pressure, flow rate
and injection temperature of the brine, as well as peri-
odic checks on waste specific gravity and suspend-solids
content. Kalium Chemicals Limited also mine potash ore
in Saskatchewan and in their borne-disposal operations
place emphasis on the continuous monitoring of flow rates,
which are telemetered and recorded in a computerized pro-
cess-control data-logging system. Disposal operations
of this type frequently have automatic shutdown capabi-
lity and pressure-release valves, effective when injec-
tion pressures and flow rates are abnormally low or high
with respect to a predetermined range of values. In ad-
dition, alarms are commonly connected to the flow rate
or pressure monitor and are activated by abnormal values.
In simpler and less costly injection systems, such as
that of Imperial Oil Enterprises Limited at the Maple
oilfield, Manitoba, where refinery spent caustic is in-
jected at low rates and pressures, visual checks are
made on injection pressures at regular time intervals.
Measurement of pressures affecting the fluids con-
fined in the annular area between the well tubing and
the well casing provides a check on possible leakeage of
wastes as a result of corrosion of the injection tubing.
Downhole equipment is frequently safeguarded through use
of a plastic-lined tubing string and cathodic protection
for the casing. The annulus is filled with an inhibited
fluid, usually diesel fuel or water. Corrosion rates are
generally monitored by means of corrosion coupons. Pre-
cautions against corrosion are of particular importance
where refinery and chemical-plant wastes are injected.
The measurement of injection pressure and/or flow
rate, together with annulus pressure, in addition to re-
gular checks for traces of corrosion, are minimum monitor-
ing requirements for safe deep-wel 1 injection of wastes
and are widely practiced with a degree of refinement,
which depends on the complexity of the disposal system.
The latter observation is reflected in costs of monitor-
ing disposal operations in the northern Williston basin
regionl0>135 listed by waste category: oilfield and
salt-cavern brines, $0 to 1,000; refinery and chemical
plant wastes, negligible cost to $5,000; and potash-mine
brines, $3,000 to $11,200.
Information on the direction and rate of subsurface
waste migration is obtained by drilling observation wells
to the deepest fresh-water aquifer above the disposal
interval or to the disposal interval itself. In both
3
9-7
-------
cases, measurement of formation pressure and analysis of
samples of formation fluids provide records of waste
movement in the subsurface. Experience of shallow ob-
servation wells near waste-injection systems in Canada
is largely restricted to the monitoring of groundwater
at depths of up to a few hundred feet to detect possible
contamination from nearby surface pits containing waste
fluids. S subsurface-disposal operation which incorpo-
rated 5 injection wells, disposing of spent caustic from
the Imperial Oil Refinery, Sarnia, into a carbonate
aquifer at 635 to 800 feet below surface, and an obser-
vation well, drilled to the disposal interval, was de-
scribed by MacLeod^. Injection pressures of 220 to 430
psig and injection rates in the range 6 to 90 U.S.g/m
during the first half of 1970 produced no change in the
static pressure of 52 psig noted in the observation well,
located some 9,500 feet from the nearest disposal well.
Nonetheless, relatively high phenol concentrations in
fluids reaching the surface in the Sarnia district sug-
gest that upward migration of injected wastes has taken
place". All 5 injection wells at the Imperial Oil Refi-
nery were abandoned in late 1974, as part of a program
by the provincial government to phase out injection of
"foreign" wastes into all strata, other than the basal
Cambrian sequence; regretably, the observation well was
abandoned at the same time. Another possible means of
monitoring pressure changes in a disposal interval is
provided by situations, in which 2 disposal wells have
been drilled to the same disposal formation. Both Husky
Oil Operations Limited, Lloydminster, Alberta, and Shell
Oil Company Limited, Corunna, Ontario, have disposed of
refinery wastes into the subsurface using one of two in-
jection wells, while using the other well both as an
emergency storage facility (or alternate disposal well)
and an observation well.
A recent survey of industry-induced crustal insta-
bility21 noted that well documented cases of injection
triggered earthquakes in North America (2) are few by
comparison with the number of fluid-injection wells in
operation (several hundred thousand). Neither the na-
tional surveys of subsurface disposal-^ 4 nor the more
detailed studies of waste injection in 0ntario7, 8 ancj
Saskatchewan10' 12' 33 revealed comparable phenomena
in Canada. However, K. Whitham, in the discussion of
the paper by Simpson and Lennox", noted the possibility
that the first record of an earthquake, produced by
fluid injection in Canada, is from Snipe Lake, Alberta,
during September, 1970. Precise geodetic surveys are
conducted routinely above some potash mines and solution
cavern operations in Canada, but nowhere is this done
in connection with fluid injection.
References
1. Study of Critical Environmental Problems, 1970, Man's
Impact on the Global Environment: and
Re.corme.ndotA.cm (ph. AcJ^ion: MIT Press, Cambridge,
Mass., 319 p.
2. Ellis, R.H., Keeney, R.L., 1972, A rational Approach
for government decisions concerning air pollution:
p. 376-400, in Drake, A.M., Keeney, R.L.,and Morse,
P.M. (Editors), AnaZyi-li Puhli.c. Syitem, MIT
Press, Cambridge, Mass., 532 p.
3. Everdingen, R.0. van, Freeze, R.A., 1971, Subsurface
disposal of waste in Canada: Canada Dept. of Environ-
ment, Inland Waters Branch, Technical Bull. no. 49,
64 p.
4. Vonhof, J.A., Everdingen, R.0. van, 1973, Subsurface
disposal of liquid industrial wastes: Preprints Ann.
meeting Canadian Inst. Mining and Metallurgy, Paper
9-1, 22 p.
5. Everdingen, R.0. van, 1974a, Subsurface disposal of
wastes in Canada - II. Disposal-formation and injec-
tion-well hydraulics: Canada Dept. of the Environ-
ment, Inland Waters Directorate, Tech. Bull. no. 78, 30p.
6. Everdingen, R.0. van, 1974b, Subsurface disposal of
wastes in Canada - III. Regional evaluation of po-
tential for underground disposal of industrial liq-
uid wastes: Canada Dept. of the Environment, Inland
Waters Directorate, Tech. Bull. no. 82, 42 p,
7. MacLean, D.D., 1968, Subsurface disposal of liquid
wastes in Ontario: Ontario Dept. Energy and Resour-
ces Management, Petroleum Resources Section, Paper
68-2, 91 p.
8. MacLean, D.D., 1971, Deep well disposal - A new look
at its potential in Ontario: Ontario Petroleum Inst.
Proc. Technical Sessions, Tenth Ann. Conference,
Paper no. 2, 8 p.
9. Ives, R.E., Eddy, G.E., 1968, Subiu-iftice. V-UpoiaZ
ofa JnduAtsiiat Wa&t&i: Interstate Oil Compact Com-
mission, Oklahoma City, Okla., 109 p.
10. Dennison, E.G., Simpson, F., 1973, Hydrogeologic
and economic factors in decision making under uncer-
tainty for normative subsurface disposal of fluid
wastes, northern Williston basin, Saskatchewan,
Canada: p. 879-928, jm Braunstein, J. (Editor),
Underground Uaite Management and A/iti /J iciaX Re.-
change, v. 2, Am. Assoc. Petroleum Geologists, U.S.
Geol. Survey and Internat. Assoc. Hydrol. Sci., 931 p.
11. Simpson, F., 1974a, Subsurface disposal of fluid
wastes in Saskatchewan (Abstract): Aw. Assoc. Pet-
roleum Geologists Bull., v. 58, p. 910-911.
12. Simpson, F., 1974b, Feasibility of deep-well injec-
tion of fluid, industrial wastes in Saskatchewan;
Preprints Ann. Meeting Canadian Inst. Mining and
Metallurgy, Paper no. 10, 16 p.
13. Simpson, F., Dennison, E.G., in press, Subsurface
waste-disposal potential of Saskatchewan: Sas-
katchewan Dept. Mineral Resources Report no. 177.
14. MacLeod, I.C., 1961, Disposal of spent caustic and
phenolic water in deep wells: Proc. Eighth Ontario
Industrial Waste Conference, p. 49-60.
15. Korompay, V. de, 1972, Brine storage in underground
reservoir by deep well disposal under active potash
mining area in Saskatchewan: Preprints Ann, Meeting
Canadian Inst. Mining and Metallurgy, Paper 9-2,26p.
16. T6th, 0., 1962, A theoretical analysis of ground-
waterflow in small drainage basins in central Alberta,
Canada: dour. Geophys. Research, v. 67, p. 4365-4387.
17. Tfith, J., 1963, A thermal analysis of groundwater
flow in small drainage basins: Jour. Geophys, Re-
search, v. 68, p. 4795-4812.
18. Hitchon, B., 1969a, Fluid flow in the western Canada
sedimentary basin - 1. Effect of topography: Water
Resources Research, v. 5, p. 186-195.
19. Hitchon, B., 1969b, Fluid flow in the western Canada
sedimentary basin - 2. Effect of geology: Water
Resources Research, v. 5, p. 460-469.
2Q Freeze, R.A., 1972, Subsurface hydrology at water
disposal sites: IBM Jour. Research and Development,
v. 16, p. 117-129.
21. Simpson, F., Lennox, D.H., 1974, Geodetic aids to
decision making in environmental management: Proc.
Conference "Geodesy for Canada", National Advisory
Committee on Control Surveys and Mapping, p, E1-E30.
-------
APPLICATION OF REMOTE MONITORING TECHNIQUES
IN AIR ENFORCEMENT
C. B. LUDWIG and M. GRIGGS
Science Applications, Incorporated
1200 Prospect Street, P. O. Box 2351
La Jolla, California 92037
Summary
The legal and technical aspects involved in the
application of remote monitors in air enforcement
programs are discussed. It is found that some of the
instruments presently under development or being
field tested are good candidates as enforcement moni-
tors while others are not suited. The advantages and
disadvantages of the remote sensors as compared to
the instack monitors are discussed. Recommenda-
tions are made for future developments.
Introduction
In EPA's regulatory programs for air pollution,
remote monitoring techniques can generally be applied
in the following three areas:
• enforcement programs
• research and development of regulations, and
• establishment of ambient air quality trends.
The specific purpose of this study is to evaluate
the first application of remote monitoring techniques.
In enforcement monitoring, two levels of "sophistica-
tion" are distinguished: (1) evidentiary monitoring for
case development, in which the plant owner is served
with a notice of violation and court action is initiated in
the event of non-compliance, and (2) surveillance, in
which a large number of stack emissions are screened
to assist in the determination of those stacks in pos-
sible non-compliance.
In the present study remote monitoring is de-
fined as sensing, by an electro-optical technique,
specific chemical and/or physical parameters of the
environment where the monitoring instrument and the
parameter under investigation are separated by a dis-
tance. Presently, EPA, NASA, NOAA, DOT and other
Federal agencies are sponsoring the development of
instruments and/or techniques to remotely monitor
the environment.
These instruments and/or techniques utilize
active or passive systems. Active systems consist of
two basic units: a transmitter and a receiver, while
passive systems consist only of a receiver. The
transmitter in an active system emits a beam of
energy, which interacts with the plume/atmosphere
by scattering, absorption and/or stimulated emission,
which is subsequently observed by the receiver. In
the passive system, the receiver merely observes
the radiation from the plume/atmosphere, which may
be emitted by thermal radiation, or scattered solar
radiation.
In enforcement monitoring, several modes can
be distinguished (see Figure 1): Direct observation of
the plume by passive or active monitors, perimeter
Passive/Active
Art a Monitoring
Activt Lon^j Path
Passive/AfUvt
Figure 1. Modes of active and passive remote techniques
in enforcement monitoring: Direct observation
of stack plume, perimeter monitoring (ground
and airborne) and area monitoring.
1
10-1
-------
monitoring by up-looking from van-based platforms
or down-looking from airborne platforms, and hori-
zontally by active long-path systems, although the
latter application is limited to cases where the hori-
zontal long-path is representative of a given situation.
In addition, area monitoring' using au aircraft, can be
used in surveillance. These surveys of wider geo-
graphical areas can also assist in the determination
of environmental quality degradation, plant site selec-
tion, the design of contact monitoring networks, and
the tracking of plumes to study atmospheric dispersion,
diffusion and fate of pollutants.
Legal Aspects in Air Enforcement Monitoring
A review was made of the legal aspects of air
enforcement monitoring, the specific requirements in
case development, and the existing proposed and pro-
mulgated rules for performance specifications of
sources and measurement methods. A literature
search revealed a lack of legal cases in which analyti-
cal instruments were used in environmental enforce-
ment proceedings. However, a number of cases in-
volved the visual observations of plume emissions,
which were either accepted as admissible evidence,
or rejected as too subjective to be admissible. Based
on the available material, we have concluded that
visual observations will be more and more questioned
by the courts as admissible evidence in light of the
availability of instruments, even though EPA has re-
vised and strengthened the Reference Method 9 (Visual
Determination of Opacity) in response to court rulings.
Based upon established practices for intro-
ducing admissible evidence into a court of law, it
appears that evidence obtained through the application
of remote techniques (other than visual observations)
could be introduced into court proceedings the same
way as evidence has been presented in previous (non-
environmental) cases. In this procedure, the field
enforcement officer convinces himself that the remote
sensor is producing reliable data, and accumulates a
"preponderance of evidence". The common sense
questions the courts would likely ask in deciding
whether to accept that evidence are expected to be:
Is the scientific principle underlying the instrument's
operation valid? Does the instrument successfully
embody and apply this underlying principle? Was the
instrument in proper working order and properly cali-
brated at the time of the test ? Was the person con-
ducting the test qualified to do so? Did the person
conducting the test use the proper procedures? If
different from the person conducting the test, is the
person interpreting the test's results qualified to do
so? It is believed, however, that a great percentage
of cases will not even lead to court action. Once evi-
dence of non-compliance is obtained, past experience
indicates that the plant owner/operator will probably
take steps to comply.
In enforcement monitoring, the accuracy of
the remote instrument must be known, but does not
need to be of a given absolute value. Recent court
decisions indicate that the measured departure from
the standards must merely be beyond the boundaries
of probable measurement error. Thus, there is no
requirement of an absolute lower limit on the instru-
ment sensitivity and accuracy, as long as it is better
than the departure from the standard to be measured.
However, remote instruments used in research and
development must have absolute bounds in accuracy
and sensitivity, like other instruments used in scien-
tific research.
In recent years, new legal issues have been
raised alt ar particular measurement methods were
challenged. Questions about the equivalency of mea-
surement methods used in setting the standards and
in establishing non-compliance in enforcement pro-
ceedings are involved. Thus, test methods and tech-
nical support for standards must be carefully estab-
lished and logical quantitative relationships between
standards and tests used for compliance must be de-
veloped. As a consequence, we believe performance
specifications and test procedures must be developed
that establish remote monitoring instruments as
"alternate methods" similar to the performance speci-
fications and test procedures that establish certain
manual and/or automatic monitors as "equivalent"
methods in the field of ambient air monitoring. The
right of conducting passive remote measurements by
enforcement personnel without notice was recently
affirmed by the Supreme Court in the case of Colorado
Air Pollution Variance Board vs. Western Alfalfa
Corp., where visual observations were involved.
(Cases involving "invisible pollution" and the probing
with active techniques must yet be tested in court.)
Legal opinions by environmental lawyers representing
industry stress that the Fourth Amendment is being
violated by the unannounced monitoring of sources and
by entry of premises without search warrants.
Present Development and Analysis
oFRemote Monitoring Techniques
A review of the recent development of remote
monitors sponsored by Federal and other agencies,
as well as by private industry, both in the U. S. A. and
other countries, revealed that several active and pas-
sive techniques were worthy of further analysis. The
active techniques included the pulsed laser systems
that involve differential absorption, Raman, resonance
Raman, fluorescence, and Lidar, as well as the con-
tinuous wave (CW) systems that involve laser, dis-
persive and non-dispersive components. The passive
techniques include dispersive and non-dispersive cor-
relation systems, spectrometers, interferometer-
spectrometers, radiometers, heterodyne radiometers,
photography and vidicons. An overview of these tech-
niques and their main characteristics are presented
in Tables 1 and 2.
The air pollutants (previously identified as
important to monitor) that are judged to be amenable
to remote monitoring in the immediate and near-term
time frame include particles/opacity, S02, NOg, CO,
light hydrocarbons, HC1, HF, NH3, NOx, H2S, HNO3,
O3 and vinyl chloride; and in the long-term time
frame include heavy hydrocarbons, oxides of sulfur,
certain specific trace elements, and chlorinated hy-
drocarbons. In addition, many newly identified pol-
lutants of interest can probably be monitored by remote
monitors, after their spectral characteristics are
identified.
2
10-1
-------
TABLE 1. Overview of Active Systems Under Development
Tc.-chnique7 Spectral
Instrument Region
Species?
Parameter
Mode
Application
Development
Status
Remarks
Differential Vis/UV
Absorption
IR
SO,, NO„
so2, no2
3
Many
Gases
Stack
Perimeter/ Evid/Surv Available
Area
Evid/Surv Field Tested Ground-based—Present instrumentation
as used not eye-safe.
Has not been done, but feasible? both
ground and aircraft based.
Stack
Evid/Surv
Perimeter/ Evid/Surv
Area
Under
Development
Under At present lor ozone.
Development
for A/C
Lldar Vis
Laser Doppler IR
Velocimeter
Opacity
Particles
Velocity
Mass Flow
Evid/Surv Field Tested Not eye safe yet, eye-safe system
being developed
Stack/ Evid
Perimeter
Stack Evid
Field Tested Gives 3-dimensional mapping of relative
on A/C concentrations
Field Tested Necessary for some emission standards
Under Relates opacity and mass concentration
Development
Long-Path Vis/UV
/IR
Many
Gases
Field Tested Using remote transmitter or retro-
reflector; can be laser, dispersive or
non-dispersive systems; useful mainly
for ambient air monitoring
Raman Vis/UV SO,
Stack
Field Tested Limited in range, especially during day.
Resonance Vis/UV
Raman
Many
Gases
All Gases
(?)
Stack/
Area
Stack/
Area
Theoretical Usefulness limited.
Lab Study Needs to be demonstrated in field;
possible interference due to fluorescence
by gases and other species.
Fluorescence Vis/UV
Fabry-Perot Vis/UV
Raman
Many
Gases
Some
Gases
Stack/
Area
Stack/
Area
Lab Study Looks doubtful in terms of sensitivity
and specificity.
Lab Study Provides increased sensitivity over
vibrational Raman; still limited in range,
especially during day.
TABLE 2. Overview of Passive Systems Under Development
Technique/ Spectral Species/
Instrument Region Parameter Mode Application
Development
Status
Remarks
Matched Filter UV/Vis SO,, NO,
Correlation
Stack/
Perimeter
Surv.
Field Tested Quantitative interpretation difficult due
to varying aerosols.
Gas Filter
Correlation
so2
CO
Stack Evid/Surv Field Tested Limited in concentration and tempera-
ture range, but temperature effect
reduced.
Perimeter Evid/Surv Field Tested No quantitative data reported as yet.
SO„
Evid/Surv Under
Development
Photography
Opacity
Stack
Evid/Surv Field Tested Needs further development for quantita-
Ground &
Aircraft
tive analysis; nighttime observations
feasible with image intensifier
Vidicon
Stack Surv Field Tested Quantitative interpretation difficult due
to varying aerosols; has potential as a
velocimeter.
SOo
Stack
Planned Independent knowledge of plume tempera-
ture required for quantitative analysis;
has potential as a velocimeter.
Heterodyne IR Many Stack/ Evid(?)/
Radiometer Gases Perimeter/ Surv
Area
Lab Study
Achieves high specificity; has yet to be
demonstrated in field.
Dispersive
Spectrometer
IR
Many
Gases
Stack Surv Field Tested Includes scanning spectrometer and in-
terferometer-spectrometer; requires
high spectral resolution for specificity
and requires knowledge of plume tem-
perature.
3
10-1
-------
Our general conclusion is that the applicability
in terms ol range in plume temperature and pollutant
concentration is greater for the active systems than
for the passive systems. In addition, the passive
systems are also more influenced by interfering para-
meters, such as background radiation. On the other
hand, passive systems are cheaper than the active
systems and can be made into cost-effective tools for
enforcement monitoring, as long as the measured
deviation from the standard in pollutant concentration
is larger than the error limits.
Advantages and Disadvantages
Based upon the review of the remote monitoring
techniques, the following advantages have been identi-
fied.
• Cost Effectiveness
Although the initial capital costs are higher
for remote instruments than for point samplers, the
operational costs are less because of their mobility
allowing rapid coverage of more sources and areas.
Our estimate is that for some remote monitoring ap-
plications the operational costs are less by as much
as about a factor of 10.
Remote techniques provide also a cost-effective
method to monitor and survey wide geographical areas
for the purpose of assisting in the determination of
environmental quality degradation, plant site selection,
design of contact monitoring networks, and in tracking
plumes.
• Objectivity in Opacity Measurements
Opacity measurements taken with instruments
are objective, in contrast to the subjective visual ob-
servations. The Appellate Court in Washington, as
well as several State regulations, favor opacity mea-
surements taken by instruments,
• Unannounced and Non-Interfering Monitoring
The remote technique provides a most effective
tool to monitor suspected violators, even at night,
without requiring entry on the facility premises. In
addition, remote monitoring does not interfere with
the normal plant operations.
• Rapid Response
In cases of air pollution episodes, the highly
mobile and flexible remote monitors can be used to
assess the extent, trend, and required response to
counteractions more rapidly than the stationary in-
situ sensors.
The following disadvantages have been iden-
tified:
• Possible High Initial Capital Costs
At the present time, the purchase price of all
remote instruments are more expensive than the ex-
tractive or in-situ devices. The active systems are
more so than the passive systems. However, as
more and more instruments are built, one can assume
that the purchase price will come down.
• Limited Applications Under Certain Conditions
Adverse weather conditions such as fog, heavy
rain or extremely high particulate content in the at-
mosphere can affect the measurement capability of
the remote techniques.
• Calibration Procedures More Complicated
Indications are that the remote techniques will
be more difficult to calibrate than the extractive or
in-situ devices because of the atmospheric influence.
It is anticipated that—similar to the calibrated smoke-
stack used in smoke schools—a test range with a re-
mote stack source must be provided in which several
atmospheric conditions can be reliably simulated.
• Possible Eye Safety Hazard
Caution must be exercised in using lasers so
that they comply with the Federal Regulations. Thus,
some of the laser applications which demand an in-
crease in emitted power in order to meet the sensiti-
vity and range requirements, had to be eliminated
from the list of useful systems.
Conclusions
Based upon our review and analysis of the re-
mote techniques being developed, our major conclusion
is that remote monitoring can play a significant role
in EPA's enforcement activities, commencing with a
few operational programs in the very near future, and
expanding to larger involvement in the long term.
However, not all techniques/instruments presently
developed are equally well suited for enforcement
monitoring of smokestacks.
For direct observation of hot plumes and in-
direct observation of complex sources through peri-
meter monitoring, the less expensive passive systems
will suffice, while the direct observation of cool
plumes will require the application of the more ex-
pensive active laser systems. The laser systems
that are most promising for near-term operational
use are differential absorption, Lidar and laser
Doppler velocimeter. The most promising passive
systems are correlation instruments, vidicons and
aircraft photographic techniques. For area surveys,
both active and passive systems (ground-based and
airborne) have application.
The work upon which this presentation is based was
performed pursuant to Contract No. 68-02-2137 with
the Environmental Protection Agency.
4
10-1
-------
LASER DOPPLER SYSTEMS IN POLLUTION MONITORING *
Christopher R. Miller
Raytheon Company
52.0 Boston Post Road
Sudbury, Massachusetts 01776
Dr. Charles M. Sonrienschein
Raytheon Company
528 Boston Post Road
Sudbury, Massachusetts 01776
Dr. William P. Herget
Environmental Protection Agency, ESRL
Research Triangle Park
North Carolina 27711
R. Milton Huffaker
National Aeronautics and Space Administration, MSFC
Huntsville, Alabama 35812
Introduction
Most conventional methods for monitoring
mass flow from stationary sources such as
power plants are expensive and time-consuming.
They require the vise of highly specialized
teams to install monitoring equipment on a
stack, make the required measurements, and
analyze the data. New methods for station-
ary source monitoring, based on remote meas-
urement of the electromagnetic properties of
the pollutants in the stack effluent, are now
being developed. Since the new instrumenta-
tion may be operated without requiring access
to the power plant property, a complete test
can be done with a minimum of preparation and
t ime.
Two types of data are needed to calculate
a mass flow rate by remote techniques:
species concentration and velocity. The U.S.
Environmental Protection Agency is currently
evaluating various remote measurement tech-
niques. Research over the past ten years,
funded by National Aeronautics and Space
Administration and other agencies, has shown
that remote measurement of wind velocities at
kilometer ranges are possible using a Laser
Doppler Velocimeter (LDV)The basis of the
technique is that laser radiation scattered
from particulates in the air is Doppler shift-
ed in frequency, and measurement of this Dopp-
ler frequency shift yields the velocity of the
particulates and hence of the wind. The
method has excellent potential for use in the
determination of smoke stack gas exit veloc-
ities. Additionally, it appears likely that
particulate concentration measurements can be
made with the same instrument by relating the
intensity of the scattered radiation to the
particle concentration. In this manner, a
single system could remotely monitor mass
flow .
This paper reports on a program designed
to determine the feasibility of remotely meas-
uring both smoke stack velocity and particu-
late concentration using an LDV. To accom-
plish this, a CO2 Laser Doppler Radar system
*This work was supported in part by the Envi-
ronmental Protection Agency and the National
Aeronautics and Space Administration.
was assembled into a trailer, and measurements
were made against an EPA instrumented smoke
stack at the River Bend Steam Station of the
Duke Power Company in Mt. Holly, North Caro-
lina. This facility is a coal burning power
plant with four boilers, each driving a tur-
bine generator capable of producing an output
up to 150 MW. Each smoke stack was equipped
with electrostatic precipitators which remove
over 99% of the particulate matter from the
effluent. In-stack measurements of the flue-
gas velocities and cross-stack optical trans-
mission were supplied by the EPA for compar-
ison to the remote data.
Laser Doppler Velocimeters
A Laser Doppler Velocimeter operates by
illuminating a moving target with laser radia-
tion. The motion of the target causes the
scattered radiation to occur at a different
frequency than the incident energy. This fre-
quency is proportional to the component of
target velocity parallel to the direction of
the radiation propagation vector. The Doppler
frequency is measured by superposing the
scattered energy collected by the receiver
with a portion of the transmitter radiation
and focussing the combination on a square law
detector. The detector output current con-
tains energy at the Doppler frequency. The
magnitude of the electrical power at the Dopp-
ler frequency is proportional to the scatter-
ing properties of the target.
For a target that can be represented as a
collection of aerosols (and in which multiple
scattering can be neglected), the detector
power is proportional to the sum of the scat-
tering gross section of the individual par-
ticles. The cross section of each particle
depends upon the ratio of the particle size
to the laser wavelength and upon the particle
index of refraction. Thus, a lack of detailed
effluent data will make it difficult to find
an exact quantitative relationship between
particle concentration and signal strength.
However, it is expected that a qualitative
agreement between concentration and scattered
radiation can be established. In order to
test this hypothesis both LDV signal strength
and cross-stack optical transmission data
were collected. Since a reasonable amount of
data has been collected which shows that cross
10-2
-------
stack transmission correlates with concentra-
tion for a given industry, a similar correla-
tion with LDV signal strength would demon-
strate the potential of the instrument to de-
termine mass concentration and mass flow, as
well as velocity.
A block diagram of the LDV system used to
make effluent measurement is shown in Figure
1. The optical portion of the system, shown
LASER
POWER
SUPPLY
INTER-
FEROMETER
TELESCOPE
*
DETECTOR
BIASING
CIRCUITRY
I
AMPLIFIER
-
SPECTRUM
ANALYZER
TARGET
o
FREQUENCY
TPACKER
MICROPHONE
TIME CODE
generator
7 CHANNEL
TAPE
RECORDER
FIGURE 1. BLOCK DIAGRAM OF LDV SYSTEM
schematically in Figure 2, consisted of a 20
watt CO2 laser, a Mach-Zehnder interferometer,
SCANNER MIRROR
f/8 TELESCOPE
CO LASER
2
INTERFEROMETER
4"
~ treturn pat
>
9 lens
ALIGNMENT (^)
TELESCOPE DETECTOR
FIGURE 2. OPTICAL LAYOUT OF THE LASER
DGPPLER VELOCIMETER
a 30 cm f/8 telescope, an output scanning
mirror, and a copper-doped germanium detector,
all mounted on an aluminum base and shock
mounted on a support table to protect against
vibration. The laser used was a Raytheon*
Model LS10A CQ2 laser which was water cooled
and semi-sealed, having a nominal output of 20
watts. Sealing of the system was through a
vacuum valve which permitted refilling of the
laser in the field if necessary. Refilling was
accomplished using the gas bottle and pumping
station located in the power unit, which also
*Mention of Company name does not constitute
endorsement by EPA or NASA.
contained the laser power supply, the closed
cycle laser cooling system and a high voltage
supply for tuning the laser by piezoelectric
transducer control of the cavity length.
The laser head contained a split dis-
charge tube closed off with non-hygroscopic
zinc selenide windows mounted at Brewster's
angle. The orientation of the windows deter-
mined the polarization of the output beam.
The discharge tube was mounted between a plane
output mirror and a four meter radius of cur-
vature rear reflector. The mirrors and dis-
charge tube were mounted on a frame utilizing
four Invar bars for thermal stability. The
cavity configuration and mechanical structure
assured a stable TEM output beam.
oo r
A Mach-Zehnder interferometer was used
in the LDV system. The configuration is shown
in Figure 3.
MIRROR
/L
TRANSMIT PATH
TO
TELESCOPE
/
»
BS3 '
r »¦
i 1
' RECEIVE PATH
FROM
LASER
y BSI BS2 ^
LOCAL OSCILLATOR PATH
V
TO
DETECTOR
FIGURE 3. MACH ZEHNDER INTERFEROMETER CONFIGURATION
The transmitted beam from the laser was re-
flected by the beamsplitter BS1 and the mirror
and passed through the beamsplitter BS3 to the
telescope. The signal beam, which consisted
of that portion of the scattered light which
retained the polarization of the output beam,
came from the telescope and was reflected by
BS3 and BS2 to the detector where it was mix-
ed with the local oscillator beam which came
from the laser through BS1 and BS2. The beam-
splitters, BS1 and BS2, were chosen so that
the product of their transmission yielded the
correct local oscillator level for the de-
tector being used.
The telescope was a reflecting telescope
of the Cassegrain type. The primary was a
30 cm diameter, f/8 spherical mirror and was
fed by a 2.54 cm diameter spherical secondary.
Focussing of the telescope was achieved by a
calibrated translation of the secondary on a
micrometer driven stage. Both components of
the telescope were mounted on the aluminum
shock mounted optical table. The use of the
reflecting elements in the telescope allowed
it to be aligned in the visible as well as
aimed visually, since reflecting elements
aligned in the visible were also aligned at
the laser wavelength in the infrared. The use
of reflecting elements also permitted aiming
of the system without transmitting a laser
beam.
An optical scanner was necessary to
direct the 30 cm diameter laser beam toward
the effluent from the smoke stack. The scan-
ner provided sufficient deflection in both
elevation and azimuth to accommodate various
target sighting geometries and to correct for
the rough positioning of the van. The ellip-
tical scan mirror was of sufficient aperture
2
10-2
-------
so that the entire laser beam was intercepted
at all scan angles. The mirror surface was
flat to X/10 at 10.6-um under ambient temper-
ature changes and various mirror positions.
The scanner mechanism provided a precise,
smooth adjustment in both elevation and azi-
muth so that target positioning could be
easily accomplished. In addition, once the
target had been aligned, the scantier was able
to be locked in position so that the laser
beam remained on the target. The scanner had
micrometer drives on both the azimuth and
elevation scan controls. It was possible for
the scanner to easily maintain alignment
accuracies of + 100 microradians, correspond-
ing to _+ 4cm at the 400m range.
A copper-doped germanium detector with a
quantum efficiency of 10% was used in the LDV
system. The liquid helium cooled Ge:Cu de-
tector required more than 20 milliwatts to
become shot noise limited and could be oper-
ated at local oscillator levels of 50-100
milliwatts if desired. At these higher
powers, the power reflected by the secondary
of the Cassegrainian telescope onto the de-
tector was negligible and no saturation
effects occurred. In addition, the Ge:Cu de-
tector was nearly impossible to damage if
excessive optical power was applied. The de-
tector was properly matched to the bias cir-
cuit and the preamplifier in order to obtain
the frequency response desired and to assure
that the shot noise exceeded the thermal
noise of the preamplifier and load resistor.
The detector output was connected to a
Hewlett Packard 141T/8553B/8552B spectrum
analyzer. A frequency/intensity tracker was
incorporated to allow continuous monitoring
of both the velocity of the effluents and the
intensity of the LDV backscatter signals.
This tracker had the capability of following
rapidly fluctuating, wide bandwidth, Doppler
IF signals. Furthermore, since the tracker
monitored the vertical output of the spectrum
analyzer, rather than the raw high frequency
Doppler signals, the tracker could be used
with tape recorded spectrum analyzer signals.
The ability to use a low frequency (30 kHz)
FM tape recorder in analyzing megahertz fre-
quency signals had several important advan-
tages with regard to signal processing. The
tracker could track the frequency and inten-
sity of signals in the 0 - 100 MHz range.
When sufficient signal-to-noise ratios were
available (typically 5 dB in 100 kHz band-
widths) , the tracker could follow signals as
small as -100 dBm.
The tracker worked by detecting the fre-
quency component having the greatest signal
amplitude. The tracker also integrated the
received signal between the preset frequency
limits to determine a total received signal
strength. The preset limits were used to re-
move spurious signals due to ground winds and
spectrum analyzer frequency markers. The
effects of noise were averaged out of the
tracker's intensity channel by setting the
integrated noise level to zero.
The outputs from the frequency and inten-
sity tracker channels were each passed throu^i
a low pass filter having a variable time con-
stant. This time constant could be controlled
to provide signal averaging over periods
from 10 msec to 2000 sec. Typically, 2 sec
averaging time constants were used in reducing
the data.
Data from the spectrum analyzer and the
frequency tracker, as well as a voice track
and a time code generator, were recorded on a
Precision Instrument magnetic tape recorder.
All components were located in the trailer.
"Experimental Results
The objectives of the measurement program
were to: (1) demonstrate the feasibility of
making effluent velocity measurements, (2) in-
vestigate the feasibility of measuring mass
flow, (3) produce exit velocity profiles
across the smoke stack, and (4) study the
effect of turbulence on the LDV signal spec-
trum. To achieve these objectives, measure-
ments of effluent exit velocities were made
under a variety of conditions including: (1)
different exit velocities obtained by varying
power plant load conditions from 80 MW to 140
MW, (2) different particulate concentrations
obtained by varying the precipitator operating
conditions, and (3) various elevation angles
obtained by use of a relay mirror. Data was
taken at elevation angles of 8°, 20O, 28°, and
37° with velocity profiles at altitudes of 0m,
0.4m, 1.4m, and 3.7m above the lip of the
smoke stack. All measurements were performed
at a slant range of 400m.
The EPA contracted with Environmental
Sciences and Engineering, Inc. to perform in-
stack pitot-tube velocity measurements at the
same time as the remote LDV velocity measure-
ments were being made. The in-stack monitor-
ing was done according to EPA Reference method
number 2 through sampling ports® located about
1.5m above the base of the smoke stack as
shown in Figure 4. The in-stack monitoring
also included cross-s-tack optical transmission
as-measured on an EPA_owned Lear-Seigler Model
RM-4 transmissometer.
The constriction of the smoke stack at
the top from 2.94m to 1.96m caused an increase
in velocity. The gas velocity measured at
the base was multiplied by the area ratio, 2.32
in order to obtain the exit velocity. The
use of such a factor assumed that there was -no
significant cooling or compression of the ex-
haust gases between the sampling ports and the
stack exit. Measurements of pressure and
temperature at the top of the stack confirm
this assumption.
The turbulence measurements required
Doppler data at various elevation angles
through the smoke stack plume to ensure the
turbulence data was not affected by spectral
foldover. A relay mirror was placed at
various points on top of the electrostatic
precipitators as shovm in Figure 4 to obtain
elevation angles between 20° and 40°.
Twenty-three data runs were made during
the field tests. These data runs included:
(1) nine high resolution velocity profiles at
heights above the stack lip from 2 cm to 3.7m,
(2) ten continuous runs through the center of
the plume at elevation angles from 8° to 37°,
(3) two precipitator runs to evaluate the LDV
3
10-2
-------
IASEP. I'ATlli
St ANT UANCF- - 400
1 .94
20 -40
2 .<-ih
HE I'CHT - 24.7
ELECTROSTATIC
PRECIPITATOR
[SIX BANKS)
l.b m
AMP LING PORTS
HE LAY MIRROR
COAL FCRF1D
hoilfr gases
FIGURE 4. SMOKE STACK LDV GEOMETRY
signal intensity as a function of cross-stack
optical transmission, and (4) two runs during
changing power plant load condition. Most of
these runs were calibrated to allow evalua-
tion of the backscatter coefficient.
Typical Doppler spectra of signals back-
scattered from the smoke stack effluents are
shown in Figure 5. The center frequency of
0.5MH1/DIV
aSMHl/DIV
CAl MARKER
0.5MH«/DIV
OMHl/DIV
0«5T"
FIGURE 5. TYPICAL DOPPLER SPECTRA FROM SMOKE STACK
EFFLUENTS AT VARIOUS LASER ELEVATION ANGLES.
the Doppler shifted signals, AvQ, was deter-
mined by the wavelength of the transmitted
radiation, X, the elevation angle, 0, and the
mean velocity of the effluents, v, according
to the equation:
AVp = (2v/X)sinO
The effect of varying the elevation angle can
be seen in Figure 5. The one megahertz width
of the Doppler spectra was the result of tur-
bulent velocities which were 4 to 6% of the
mean exit velocity from the smoke stack.
The Reynold's number for the smoke stack
was approximately 30,000 which was well into
the turbulent flow region. The turbulence
flattens the velocity profile in the stack.
The acceleration of the gas velocity in the
converging section at the top of the stack
further flattens the exit velocity profile un-
til it has the form shown in Figure 6. The
M/SEC FT/SEC
1 50
40-
12 cj .
H
u 30.
. 100.
3
C4
>
75
20-
50 ¦
10-
2 5 ¦
0
2 FT
APPROX
STACK DIAMETER
FIGURE 6. VELOCITY PROFILE AT THE SMOKE STACK EXIT.
flatness of the velocity profile relaxes the
LDV1s alignment requirements for making an
accurate velocity measurement and makes a
single line of sight measurement representa-
tive of the average stack velocity.
The tapes were analyzed to determine the
smoke stack effluent exit velocity as a func-
tion of in-stack velocity and the correlation
of effluent backscatter signal strength with
the cross-stack optical transmission.
Numerous data runs were used to determine
the smoke stack effluent exit velocity as a
function of in-stack velocity. During these
runs, the power load varied between 83 MW and
137 MW and both remote and in-stack effluent
velocity measurements were made. The results
of this analysis are plotted in Figure 7. It
42n
>
a
40
38-1
>< g
m o 36
§ w
i«
2 W
§ * 32-
o?
s "3o'
8 *
&
H
a
26-
24-
22-
+ x'o
V
O 8" ELEVATION ANGLE
+ 20° ELEVATION ANGLE
X 28° ELEVATION ANGLE
t. 37° ELEVATION ANGLE
1 1 1 1 1
25 30 35 40 45 5<
EXIT VELOCITY MEASURED BY A PITOT TUBE
IN METERS PER SECOND
FIGURE 7. EFFLUENT EXIT VELOCITY MEASURED BY A LASER
DOPPLER VELOCIMETER AS A FUNCTION OF THE EXIT
VELOCITY MEASURED BY AN IN-STACK PITOT TUBE.
4
10-2
-------
appears that the effluent exit velocity from the
power plant smoke stack under test is a lin-
ear function of the power load on the gener-
ating unit exhausting gases through the stack.
The standard deviation of the Doppler data
from a straight line is 1.40 m/sec, which in-
dicates that a linear equation provides a good
approximation for estimation of the exit ve-
locity .
The in-stack velocity measurements yielded
exit velocities which averaged 14% higher than
the LDV measured velocities. The standard de-
viation of the pitot-tube data from a straight
line is 1.10 m/sec again indicating a good
linear fit.
The cause for the 14% discrepancy between
the remote and in-stack velocity data is not
known. A number of possible causes are: (1)
error in measuring the LDV's elevation angles,
(2) miscalibration of the pitot tube used on
in-stack measurements, and (3) erroneous
correction factor used in calculating the gas
exit velocity from the in-stack velocity. In
any event the discrepancy appears to be the
result of a systematic error rather than a
random error.
In order to determine the relationship be-
tween the backscattered signal strength and
the cross-stack optical transmission, the
total integrated signal strength taken from
the intensity tracker was plotted as a
function of the attenuation coefficient mea-
sured in the visible by the transmissometer
in Figure 8. The apparent linearity of the
graph indicates a linear relationship between
the received backscattered signal at 10.6-pm
and effluent particle concentration.
i.o
0.9
0.8
0,7
0.6
0.5
0.4
0,3
0,2
0,1
0
2.54 Cr (N) + 0.06
0.024
o BMi TAKEN FROM TftPE #13
ON 17 JAN. 1975
+ DATA TAKEN FROM TAPE #10
ON 16 JAN. 1975
a(N)
programs.
The correlation between LDV signal
strength and pluine opacity and between plume
opacity and mass concentration indicates the
LDV has the capability of measuring mass con-
centration from a smoke stack. A combination
of the mass concentration and the velocity
data yields a measurement of mass flow rate.
Thus it appears likely that a single LDV can
remotely measure the mass flow from a power
plant smoke stack.
References
1. Coons, J. D., H. A. James, H. C. Johnson,
and M. S. Walker. "Development, Calibra-
tion, and Use of a Plume Evaluation Train-
ing Unit," Journal of Air Pollution Con-
trol Association""1!!): 199-203 , May 1965 .
2. Rose, A. H., and J. S. Nader. "Field Eval-
uation of an Improved Smoke Inspection
Guide," Journal of Air Pollution Control
Association 8: 117-119, August 1958.
3. Jelalian, A. V. et al, "Laser Doppler
Systems for Remote Atmospheric Measure-
ment", presented at NEREM, Boston, Mass.,
1971.
4. Sonnenschein, C. M. and Horrigan, F. A.,
"Signal-to-Noise Relationships for Coaxial
Systems that Heterodyne Backscatter from
the Atmosphere," Applied Optics, Vol. 10,
No. 7, July 1971, pp. 1600 - 1604.
5. Connor, W., Measurement of the Opacity
and Mass Concentration of Particulate
'Emissions by Transmissometer"] EPA-650/2-
74-128, November 1974.
6. Federal Register, Vol. 36, No. 247, Thurs-
day, December 23, 1971, pp. 21881-21885.
7. J. F. Lee and F. W. Sears, Thermodvnaini cs .
Addison-VJesley Publishing Co., Inc.,
Reading, Massachusetts, 1963, pp. 284 -
287 .
0 0.1 0.2 0.3 OA
EFFLUENT ATTENUATION COFFICIENT IN INVERSE METERS
FIGURE 8. RELATIVE INTEGRATED DOPPLER SIGNAL
STRENGTH AS A FUNCTION OF THE EFFLUENT
ATTENUATION COEFFICIENT.
'Conclusions
A laser Doppler velocimeter has conclu-
sively shown its ability to remotely measure
the velocity of effluents from a power plant
smoke stack. Velocity data was both self-
consistent, as shown from profile measurements,
and in good agreement with pitot tube data.
Deviations between the LDV and the pitot tube
were of a systematic, rather than a random,
nature and are probably associated with a rel-
ative miscalibration between the two instru-
ments. A least square fit analysis of the
velocity data indicates an accuracy of approx-
imately 1 m/sec, although greater accuracy has
been obtained in other LDV measurement
5
10-2
-------
DOWNLOOKINf. AIRBORNE LIDAR STUDIES - August 1974
J. A, Eckert, J. L. McElroy, D. H. Bundy, J. L. Ciuagliardo and S. II. Melfi
li. S. Environmental Protection Agency
Environmental Monitoring and Support, laboratory
Las Vegas, Nevada
Summary
Airborne LIDAR studies were conducted over
St. Louis, MO, during August 1974 in support of
the Regional Air Pollution Study's (RAPS) summer
intensive period. Plights were made to determine the
height of the mixed layer over the St. Louis
metropolitan area. Emphasis was on obtaining data
during the morning and evening transition periods.
The downiooking airborne LIDAR, at the present time,
has the unique capability of determining mixing
layer height over large geographical areas in a
relatively short period of time.
Instrumentation for this monitoring technique
consists of a downiooking LIDAR mounted in a fixed-
wing aircraft with on-board data capability. Final
data reduction is by ground-based digital computer.
Data taken from the February 19741 .2 RAPS intensive
peTiod demonstrated a strong correlation between
vertical profile nephelometer data obtained from
helicopter ascents and aerosol scattering as
recorded by the downiooking LIDAR. Mixed layer
height is determined by observing the LIDAR returns
from aerosols trapped at the inversion interface.
Results of the study suggest several applica-
tions for this monitoring technique:
1. Investigation of the fine structure of the
atmospheric boundary layer.
2. Determining the optimal location for ground-
based sensors for determining the mixing layer
height (e.g., acoustic sounders3).
3. Determining the dimensions of an urban plume.
4. Obtaining data on the aerosol distribution in
individual plumes.
Instrumentation and Techniques
Principal components of the LIDAR include a
Q-switched ruby laser and a 38-cm fresnel lens
receiving telescope. Return signals are detected by
a photomultiplier tube and subsequently digitized
with a fast analog to digital converter with storage
capability. On-board data output is recorded on a
strip chart. System parameters are summarized in
Table 1. The strip charts were subsequently
digitized at the home laboratory and final data
storage was on magnetic tape. Plots of the various
traverses were generated on a large digital computer.
All flights were made at an altitude of 3,000
meters mean sea level, a constraint imposed by eye
safety to a possible ground observer. Ground location
of the LIDAR pulses was determined by recording the
time sequence of laser firings and visual sightings of
landmarks beneath the aircraft. Positional accuracy
is estimated to be ± 0.5 km.
TABLE 1
SYSTEM PARAMETERS
PHYSICAL:
Laser - Q-Switehed Ruby
C
Frequency: 6943 A
Output Power: 1 joule
Pulse Length: 20 nsec
Telescope - 38 cm Fresnel Lens, Aircraft Monoque
Construction
Oetector - RCA C3100A Photomultiplier
Size - 0.S m3
Weight - 1,350 Kg
Power Requirements - 2KVI Peak, 0.6 KW Standby
OPERATIONAL:
Altitude - Minimum 3,000 m above ground level
(a.g.i.j for Eye Safety
Horizontal Resolution Element - ^750 m (75 m/sec
Airspeed)
Vertical Resolution - M5 m
Sensitivity - 0.5 Times Scattering from Clear Air
Signal Rate - 1 Pulse Every 10 Seconds
Results and Discussion
Flights were made between August S and August 20,
1974, to coincide with activities of other experi-
menters during the RAPS summer intensive study period.
Over 1S00 individual LIDAR returns were obtained from
48 flight traverses over or near the St. Louis metro-
politan area. Flight paths were chosen consistent
with predicted wind direction. Emphasis was on the
determination of the mixing layer height during the
transition periods; however, some data were obtained
on individual industrial plumes treated as targets of
opportunity during the boundary layer measurements.
One extended south to north traverse was made 30 km
east of St, Louis in an attempt to ascertain the
dimensions of the urban plume.
Two different examples of data are shown to
emphasiie the several applications for this monitoring
technique. Figure 1 is a map of the metropolitan area
showing the location of the traverses to be discussed.
The path running north to south represents a flight
over a power generation station located north of
St. Louis with data shown in Figure 2. Data shown
in Figure 3 were obtained on a series of flights
made essentially along the path labeled "West to East
Traverse" on Figure 1.
1
10-3
-------
On the morning of August 19 at 1000 CDT, a
north to south traverse was nvade across the power
generating station north of St. Louis (see Figure 1J.
The flight path was directly over and along the
direction of the plume. The LIDAR traces as plotted
in the figure have been corrected for range
dependency by multiplying the LIDAR return by R2
(range, raj and represent the actual intensity of the
backseatter signal. Vertical dimensions and details
of plume structure can be observed. Horizontal
resolution of the present LIDAR system precludes
obtaining detailed dimensions in the horizontal
plane. The 750-m to 1000-m above ground level
(a.g.l.) value for plume height is consistent
with the mixing layer height as observed in the data
obtained on a 1020 CDT west to east traverse made
immediately after the plume study (see Figure 3).
Figure 3 shows LIDAR returns from 5 west to east
traverses conducted between 0900 CDT August 19 and
0900 CDT August 20. On the morning of August 19,
winds were light and generally from the northwest.
Referring to the data obtained in the 101S CDT
traverse, note the region of increased scattering
[labeled (A)] over urban St. Louis. This represents
the top of the mixing layer and correlates with in
situ data obtained at about the same time by a
helicopter obtaining vertical profiles. Some of the
scattering in this region is due to the plume from
the power plant north of the city. The relatively
clear air indicated on the left side of the graph is
consistent with winds from the northwest and the
decrease in aerosol content upwind of the urban
environment.
The height of the mixing layer again can be
determined in the 1220 CDT traverse by observing the
region of increased scattering [labeled (B)] over
St. Louis. Note the height of the mixed layeT has
risen from 750 meters a.g.l. at 1015 CDT to 1400
meters a.g.l. at 1220 CDT.
The 2200 CDT traverse was made during the evening
transition period. Note the increased scattering at
around 1400 meters a.g.l. which defines the height of
the daytime mixing layer. The 2300 CDT traverse again
shows the height of and aerosol structure within the
daytime mixed layer [labeled (C)]. The region of
increased scattering in the vicinity of the Arch at
150 meters a.g.l. represents the urban mixing layer
and correlates with data obtained by a helicopter
sampling in this area at this time.
The final traverse of the series, taken at 0820
CDT on August 20 again shows increased scattering
in the vicinity of the Arch [labeled (D)] which
represents the urban mixing layer. The height, 150
to 200 meters a.g.l., is consistent with profiles
obtained in the area with in situ instrumentation
mounted on a helicopter.
Conclusions
1. The downlooking airborne LIDAR is able to
document the height of the mixing layer over a
relatively large geographical area in a short
period of time.
2. Results obtained by the airborne LIDAR technique
compare well with data obtained using iji situ
instrumentation mounted in an aircraft.
3. The results suggest several additional applica-
tions of the technique besides boundary layer
structure determination. The applications
include individual plume studies, studies of the
urban plume, and feasibility studies for placement
of a network of ground-based fixed sensors.
Acknowledgement
The authors wish to acknowledge the efforts of
Mr. Jack L. Peacock of the Remote Monitoring Methods
Branch, Remote Sensing Division, Environmental Moni-
toring and Support Laboratory-Las Vegas, who wrote
the necessary software for digitizing, analyzing,
and graphically displaying the data used in this
presentation.
References
1. Eckert, J. A.; J. L. McElroy; D. 11. Bundy;
J. L. Guagliardo and S. H. Melfi; "Airborne
LIDAR RAPS Studies, February 1974." (To be
published as an EPA report.)
2. McElroy, J. L. and J. F. Clarke, "Atmospheric
Diffusion during Sunset-Sunrise Transitional
Periods." Paper presented at Symposium on
Atmospheric Diffusion and Air Pollution,
American Meteorological Society, Santa Barbara,
CA, September 1974.
3. Hall, Freeman F., Jr., "Chapter 18 - Temperature
and Wind Structure Studies by Acoustic Echo-
Sounding," Remote Sensing of the Troposphere,
U. S. Department of Commerce, National Oceanic
and Atmospheric Administration, University of
Colorado, Boulder, CO, August 15, 1972,
p. 18/1 to 18/26.
2
10-3
-------
•POWER.
PIANT \
GRANITE CITY
ST. LOUIS
ARCH
SCALE OF KILOMETRES
E ST. LOUIS
i-KlURii I
Map of St. Louis Showing l.Ldar Traverses on August 19-20, 1974
1000
MISSISSIPPI RIVER STACK MISSOURI RIVER
GROUND POSITION. KM
FIGURE 2
L1DA.R Return Signals from North to South Traverse
over Power Plant North of St. Louis, MO.
10-3
-------
11015 CDT
8/19/74
1
1220 CDT
8/19/74
L
2200 CDT
8/19/74
V) 101)0
2300 CDT
18/19/74
! OmYfYi
< 1U00
0820 CDr
8/20/74
1000
ahoh
si LOUIS
'JO ?
chari.cs
GROUND POSITION KM
io if) ;>o
BLLLVIL t £
ILLINOIS
FIGURU 3
LIDAR Return Signals from 5 West to Hast Traverses
over St. Louis, MO, Made on August 19 and August 20, 1974
4
10-3
-------
VISUALIZATION OF EDDIES IN THE PLANETARY
BOUNDARY LAYER BY MEANS OF LIDAR
K.E. Kunkel, E.W. Eloranta, J.A. Weinman
Department of Meteorology
University of Wisconsin
Madison, Wisconsin 53706
Summary
A lidar system with a scanning capability has been
used to obtain pictures of aerosol distribution in the
atmospheric boundary layer. These pictures were taken
under clear conditions with no visible haze. The height
of the morning inversion and its change with time, the
wind field, and images of the structure of clear air
convection have been obtained from these pictures.
II. Introduction
The proper site selection for potential air pollu-
tion sources requires an adequate climatology of the
dispersion characteristics of an area. Ground-based
instrumentation and towers can provide this information
in the lowest part of the boundary layer. Above tower
level, information has been traditionally provided by
balloon and aircraft borne Instruments which have tem-
poral and spatial sampling limitations. A lidar system
can provide meteorological information without many of
the sampling limitations of balloon or aircraft mea-
surements .
Lidar, the optical equivalent of radar, is sensi-
tive primarily to aerosols. The density of aerosols is
very often inhomogeneous in both vertical and horizon-
tal directions. By following these inhomogeneities,
the lidar Is able to map the motion field in the bound-
ary layer.
The University of Wisconsin lidar has been used to
obtain the following:
1) Inversion height and its change with time.
2) Wind field including temporal and spatial var-
iations.
3) Nature of vertical convective mixing.
III. Lidar System
The characteristics of the University of Wisconsin
lidar are summarized in Table 1. The computer control-
led scanning ability allows Range-Height-Indicator (RHI)
and Plan-Position-Indicator (PPI) pictures of aerosol
density to be constructed from lidar data.^
IV. Time Cross-section
In the simplest configuration, the lidar is orient-
ed at a fixed position and fired at a constant rate.
This data can be used to construct a time vs height dis-
play of aerosol density.-^ Fig. 1 is an example of such
a display taken on Feb. 3, 1975, at an elevation angle
of 15.0° and a firing rate of 1 pulse/5sec. The height
of the morning surface inversion is Indicated by the
decrease in the returned signal at a height of about
200m at 1030 CST. This gradually rises as surface heat-
ing continues and convective plumes are quite apparent
in the latter part o^ the sequence. About 1115 the in-
version reached the height of the bright layer at 800m
Table 1
Lidar Specifications (1975 Model)
Transmitter
Laser rod
Wavelength (nm)
Pulse energy (joules)
Pulse length (ns)
Maximum PRF (pulses/min)
Receiver
Optics
Field-of-view (mr)
Filter passband width (nm)
Scanning and Firing
Azimuth
Elevation
Ruby (3/8 x 6 Inches)
694.3
2.0
30
60
Newtonian reflector
(8 inch)
2.0 to 7.0
2.0
0.1 - 90.0° computer
controlled
0.1 - 99.0° computer
controlled
Digital Processing/Recordlng/Dlsplay
A/D conversion
Sample intervals/range
resolution
Total number of samples
per lidar shot
Computer
Resolution: 8 bits
Bit rate: 800 mega-
bits/sec (max.)
100 nsec/15m
200 nsec/30m
1024
28K-words PDP 11/40
(16 bits/word)
and the sky quickly clouded over. The discontinuity at
a height of 1 km is an apparent upper level inversion.
The Green Bay, Wisconsin radiosonde report at 0600 CST
indicated an inversion at this height.
I
V-/
W
Eh
3
1.5
1.0
' WM1 1
¦ftp :J .y? f |
1110
TIME (CST)
Fig. 1. Time vs. height cross-section of aerosol den-
sity taken on Feb. 3, 1975 at an elevation angle of
15°.
10-4
-------
TIME
AZIMUTH ANGLE
1.0 2.0 3.0 4.0
SLANT RANGE (km)
Fig. 2. Two PPI scans taken on Mar. 10,
Indicate a feature common to both scans,
speed of seven knots.
1975 at an elevation angle of 5°. Arrows
Movement of this feature indicated a wind
This type of data can be used to determine the
height of the morning surface inversion and its change
with time. In addition, higher inversions which may
act as an ultimate lid on vertical mixing may be de-
tected.
V. Wind Measurements
By varying the lidar orientation continuously in
a scanning mode, time sequences of RHI and PPI pictures
can be produced. Individual inhomogeneities in the
aerosol density can be followed and velocities can be
calculated from the displacements. A series of PPI
scans taken on Mar. 10, 1975 at an elevation angle of
5.0° clearly shows the motion of the inhomogeneities.
Fig. 2 shows two of these scans and also points out an
individual feature common to both scans. The movement
of this feature indicated a wind speed of 7 knots at a
height of about 300 m. An anemometer at a height of
70 m also indicated a wind speed of 7 knots.
Application of lag cross correlation techniques
has generated radial wind speed profiles from lidar
data.4 A lidar-generated wind profile from Oct. 29,
1973, shown in Fig. 3, compares favorably with simultan-
eous pilot balloon soundings.
Such a system provides two-dimensional wind infor-
tlon which could be used to observe the effects of lo-
cal terrain on the wind field and could also provide
clioatologlcal wind information.
VI. Convective Structure
Since the aerosol density usually decreases with
height, buoyant rising plumes having a greater aerosol
1200
1000
800
E
H 600
x
o
UJ
x 400
200
O
Fig. 3. Radial wind speed component profiles obtained
on Oct. 29, 1973. Heavy solid line indicates the winds
computed from lidar data from 1934-1949 CST. Pilot
balloon launch times were 1920 CST (a), 1930 CST (•)
and 1943 CST (a).
0 4 8 12
VELOCITY (m/sec)
2
10-4
-------
density than the surrounding environment will be vis-
ible to the lidar. The time cross-section of Fig. 1
shows several plumes in the lower part of the boundary
layer. Fig. 4 shows an RHI scan taken on July 14, 1975,
in which plumes are clearly evident. The PPI scans in
Fig. 2 indicate a definite streeting effect of the
convective elements. Each of these days was clear and
sunny with good visibility.
These pictures can be used to easily visualize
the convective mixing processes occurring in the
boundary layer.
References
1. Battan, L.J. (1973). Radar Observations of the At-
mosphere. University of Chicago Press.
2. Naito, K., K. Takahashi, I. Tabata, and Y. Yokota
(1974). Lidar observation of the convection in the
lower atmosphere. Presented at the Sixth Confer-
ence on Laser Atmospheric Studies, September 1974,
Sendai, Japan.
3. Noonkester, V.R., D.R. Jensen, and J.H. Richter
(1974). Concurrent FM-CW radar and lidar observa-
tions of the boundary layer. J. Appl. Meteor., 13,
249-256.
4. Eloranta, E.W., J.M. King, and J.A. Weinman (1975).
The determination of wind speeds in the boundary
layer by monostatic lidar. J. Appl. Meteor., to be
published.
4.0 3.0 2.0 1.0
RANGE (km)
Fig. 4. RHI scan taken on July 14, 1975. Arrows indi-
cate convective plumes.
3
10-4
-------
REMOTE SENSING OF ATMOSPHERIC POLLUTANT GASES
USING AN INFRARED HETERODYNE SPECTROMETER
R. K. Seals, Jr.
NASA Langley Research Center
Hampton, Virginia
and
B. J. Peyton
AIL, A Division of Cutler Hammer
Melville, New York
^mmarj/
Remote sensing of the concentration and vertical
distribution of atmospheric gases using an infrared
heterodyne spectrometer (IHS) has been investigated,
and a dual C^O^ laser multichannel IHS has been
developed. Analyses of nadir thermal radiance meas-
urements from a 10-km aircraft and solar absorption
measurements from the ground indicate that initial
applications of the IHS to tropospheric measurements of
NH^ and 0^ are feasible with measurement precisions
ranging from 0.5 to 2 ppb and 20 to 30 ppb, respec-
tively. These analyses have included effects of
potential retrieval error sources and have resulted in
specifications of measurement modes, optimum signature
lines, required system parameters, and expected sensi-
tivities. Preliminary instrument performance data are
presented.
In troduct ion
A variety of atmospheric radiance and absorption
measurements have been made using direct detection
infrared radiometers.1«2 However, recent development
of heterodyne detection techniques makes such measure-
ments possible with nearly quantum-noise-limited sensi-
tivity and ultra-fine spectral resolution. These
features are of particular significance in remote
sensing of atmospheric and pollutant gases where low-
level signals exist in a background of other interfer-
ing gases. IHS spectral resolutions ranging from
-k -2 -1
10 to 6.7 x 10 cm are possible and provide the
capability of scanning individual atmospheric signature
lines. Heterodyne measurements of solar and astronomi-
cal radiances have been reported,7-12 anj heterodyne
detection of radiances from sample cells of several
pollutant gases has been demonstrated.13 Several analy-
ses illustrating the use of heterodyne detection for
remote sensing of gas concentrations have been
reported.14_18 particular, the potential of hetero-
dyne measurements for inferring pollutant gas profiles
from satellite measurements of atmospheric emissions-^
and from measurements of reflected laser (hot source)
absorption^.17 has been discussed.
This paper describes the development of a 2-GHz
bandwidth multichannel infrared heterodyne spectrometer
(IHS) for ground-based and airborne measurements of the
concentration and altitude distribution of tropospheric
NH^ and 0^ in both the atmospheric emission and solar
radiance viewing modes. Extensive new analytical
results for measurement sensitivities, optimum signa-
ture lines, effects of receiver noise and other meas-
urement error sources, and IHS system parameters are
reported. The resulting instrument, a dual laser IHS
employing discretely tunable C^O^ laser local oscil-
lators (LO's) operating in the 9- to 11-pm spectral
L-10282 1
region, Is outlined, and preliminary operating charac-
teristics are reported. Future extensions of the IHS
to measurements of other gases and to stratospheric
applications are briefly reviewed.
Signature Line Selection
A number of gases, including NH„ , 0_, C„H,, HN0„,
3 3 2 4 3
H^O, and the chlorofluoromethanes, have spectral signa-
ture lines in the 9- to 11~U® portion of the infrared.
Initially, an IHS utilizing laser LO's for sig-
nature line selection was considered. However, due to
the advantages of eliminating signal interference and
attenuation from atmospheric CO^, the use of isotope
C0„ lasers was also investigated. After comparing the
19
overlap of isotope CO^ laser transitions with calcu-
lated and measured spectra of gaseous molecules in the
9- to 11-pm region, C^oJ,^ laser LO's were chosen for
detection of NH^ and 0^ in initial applications. Both
NH^ and 0^ are Important in studies of atmospheric
processes and pollution; and tropospheric measurements,
particularly in the case of NH^, are limited. Atmos-
pheric is Involved in aerosol production2^ ancj may
constitute a significant source or sink in NO cherais-
21 x
try. Atmospheric 0 accounts for much of the pollu-
22
tion injury to vegetation, and excessive 0^ levels
are considered evidence of photochemical smog
23
format: ion.
A comprehensive study of the overlap between C^oi^
laser transitions and NH^ and 0^ spectral signature
lines was performed to determine optimum IHS operating
wavelengths. Nadir measurements of upwelling atmos-
pheric radiance (nadir radiance or NR mode) and measure-
ments of atmospheric absorption of solar radiation
(solar absorption or SA mode) from the ground were
investigated. A line-by-line computer model^-^ has been
used for all of the transmittance, absorption, and
radiance calculations. This model considers both
Lorentz and Voigt profile line shapes and includes
temperature and pressure effects on spectral line
strengths and half widths. Effects of gaseous H^O, CO^,
0 , NO, S0„, and NH were included. Line parameter
9/l ) S
data for H^O, CO^, 0^, and i^O and for NH^ were
obtained from the literature, while the data for SO,,
came from an unpublished compilation by R. F. Calfee of
N0AA. NH absorption measurements^ have been carried
13 16
out with a C O^ laser to verify the predicted
10-5
-------
signature line characteristics for NH^, and tunable diode
laser spectroscopy is being performed to verify botii the
Niland 0 predictions. A midlacitude summer model atmos-
phere has been used to provide temperature and pressure
profiles and H,0 and 0 altitude d is tributions. Esti-
27
mates of U„0 continuum absorption effects have been
included.
in selecting optimum IHS laser LO transitions, a
number of factors such as overlap with either pollutant
signature lines or "clear" spectral regions, frequency
variation of radiances within the IHS bandpass, and
interfering gas effects were considered. Figure 1
shows calculated transmittance spectra from 925 to
935 cm-1 and from 1025 to 1035 cm 1 for a ground to
10-km altitude zenith path and illustrates the spectral
overlap between C130*6 laser transitions (indicated
across the top of the figure) and some of the signature
lines of Nl^ and 03. in the 925 to 935 cm 1 region
most of the structure, not designated H.0, results from
25
Q Branch absorption of NH3. Using analytical results
such as those in Figure 1 and extensive calculations
13 16
around specific laser transitions, the C 0., R(18)
transition at 927,300 cm-1 and the R(8) transition at
920.219 cm"1 were chosen as the pollutant and reference
L0 transitions, respectively, for NH^ measurements in
both viewing modes. The role of the reference LO willbe
discussed later, in the 1025 to 1035 cm 1 region of
Figure 1, most of the non-H^O structure is due to 0^.
Because of the peculiar altitude distribution of atmos-
pheric 0^, different wavelength pairs were chosen
as LO's for the two 03 missions. For the NR mode, the
C130;^6 R(24) and R(40) transitions at 1034.838 cm-1
— 1
and 1043.473 cm were selected for the pollutant and
reference LO's, respectively. For the SA mode, the
C^o*6 R(10) and P(18) transitions at 1025.778 cm-1
and 1002.478 cm-1 were selected for the pollutant and
reference LO's.
IHS Instrument Development
A wideband, laser tunable, Dicke switched, infrared
heterodyne spectrometer (IHS) has been developed for
both ground- and aircraft-based measurements. The IHS
(see Fig. 2) utilizes two FV HgCdXe photomixera, two
grating tunable laser local oscillators (LO's),
four IF networks with RF filters which spectrally chan-
nelize the incident irradiance, two black-body sources
for measurement reference and absolute calibration,
four radiometer processing channels, and appropriate
recorders. The Dicke type IHS processor can be
employed when the source and reference temperatures are
nearly identical (NR mode); the automatic nulling gain
modulation type processor is used when the source tem-
perature is much greater than the reference temperature
(SA mode).
An optical Dicke switch alternately switches the
receiver field of view (F0V) between a collecting tele-
scope and a reference black body, and supplies syn-
chronizing signal to the processing channels. A common
collecting aperture, reference black body, and Dicke
switch are used for the matched pliotomixers to (a) pro-
vide a common receiver field of view, and (b) minimize
the effects of laser LO instabilities. A black-body
source is inserted between the collecting aperture and
the optical Dicke switch for absolute IHS calibration.
The reference channel photomixer has a 3-dB cutoff
frequency4 of 1000 MHz while the pollutant channel
photomixer has a 3-dB cutoff frequency of greater than
1500 MHz. Both infrared channels exhibit heterodyne
sensitivities of less than 1.5 x 10 W/Hz at IF fre-
quencies below 500 MHz. The bandwidths of the IHS
channels are fixed by RF filters, and the post-
detection integration time is discretely selectable
between 0.1 and 30 seconds. The NR mode black-body
sources can be varied between 280 and 300 K while the
SA mode black-body sources can be varied between 300
and 1500 K. The IHS characteristics discussed here
have been used in the computer model calculations for
predicting IHS performance.
IHS Channel Optimization
An important feature of the IHS is the capability
to accurately channelize the receiver bandwidth into
finer spectral intervals in order to obtain information
on the frequency variation of an incoming radiance.
The IHS responds to radiation over a frequency range
from (vL0 - vR) to (vLQ + vR), where vLQ is the L0
frequency and v is the frequency response of the
a
infrared photomixer. An important characteristic of
heterodyne receivers is the down-conversion of the
infrared energy to RF frequencies before channelization
occurs. In addition, energy at a frequency - v)
within the receiver bandpass is summed with energy at
frequency (v^ + v) due to the heterodyne receiver
image effect.
Spectral channelization of transmitted solar
energy around the LO transition at 927.300 cm ^ into
500 MHz channels is illustrated in Figure 3. The left-
hand part of Figure 3 shows calculated transmittance
between 927.22 cm 1 and 927.38 cm 1 for an atmospheric
aenith path with an signature line at 927.319 cm-1"
providing the structure. The 500 MHz channels are
centered at +2050, +1550, +1050, and +550 MHz from the
C13of R(18) LO transition. The right-hand part of
Figure 3 shows the altitude variation of normalized NH^
absorption coefficients for each channel (i.e., weight-
ing functions for an absorption measurement). Figure 3
also demonstrates that channel 1 (located at
vL() + 2050 MHz) is most sensitive to absorption by NH^
concentrations near ground level while channels 2, 3,
and 4 have peak sensitivities at increasingly higher
altitudes. This behavior is analogous to that dis-
cussed in analysis of vertical laser reflection absorp-
tion measurements^^'17 and provides the basis for
determining concentration profiles of absorbing gases
from IHS measurements in the SA mode. Similarly,
channelization of upwelling thermal radiances can be
used to infer concentration profiles of emitting
gasesl5 when the atmospheric temperature profile is
known. This radiance channelization provides the basis
for the NR mode IHS measurements.
The basic equations for the detected radiance in
IHS channel j (of spectral width 3 centered at
frequency v + V ) are given for the NR and SA
15 16
modes, respectively, by '
2
10-5
-------
,,NR
B . (0) x . (0) + /„ C B.(y) [IJi ,(y)/-Jyl cfy
0 J .] 0 J J
SSA *1.1 .«))
3 J J
(1)
(2)
where y - -in (p/p ), p is atmospheric pressure, the
subscript o refers to ground conditions, aad the sub-
script t refers to upper altitude limits (10 tan for
the NR mode arid through the atmosphere f:or the SA
mode). 11. (y) is Planck's radiance at \> in units of
J _ I
energy per time per area per unit solid angle per cm
for a single polarization, and r is the ground
vs»i»8ivivv. I. is the solar radiance and is assumed
;i
to be given by ft evaluated at a temperature of
5HO0 K. The quantity x . (y) is the frequency average
of the transmittance t (y) over the channel bandpass
where
- - F ff
g ¦ I y
K . Ui. e
j, i i
dy
(y)
t e
V j c
(3>
N = (bi^ 0/f|) (l^./t)
1/2
(5)
and where h is Planck's constant and t is the post-
delecti.on integration time. The effective quantum
efficiency i) for tlie l.iiS pollutant channels can be
expressed by
0.256
Ko ~ y - 10 Hz
0.2'S6{l.6Z x lO~19jvJO - ul2 - 1.35 x l10 Hz
(6)
nnil wYiere i is the transmittance related to non-
v, c
molecular absorption and other losses, K . is the
absorption coefficient at frequency ,> due to gas i,
w, is the mass mixing ratio of gas i, and g is the
gravitational constant. For heterodyne receivers, con-
verting the quantities given by Equations (1) and (2)
into energy per time units requires multiplication by
2
a factor cfj./v]0 where c is the speed of light.
IHS Retrieval Simulations
Extensive simulated radiance inversion and gas
profile retrieval calculations have been performed to
define such IHS parameters as channel locations and
bandwidths, post-detection integration times, and
measurement sensitivities. Using the selected laser
10 transitions, simulated radiances for the SA and NR
modes have been calculated from Equations (i)-(3) for
a variety of conditions and NH.j and 0 concentration
profiles. With the simulated radiances as input data,
iterative profile retrieval methods have been used
which involve an initial guess for the vertical pol-
lutant profile and successive adjustments to this pro-
file using prediction algorithms until selected conver-
gence criteria have been satisfied. The NR mode
prediction algorithm is analogous to the one described
in Reference 15. For the SA mode, successive
approximations ta the volume mixing ratio x^(y) of
gas i at level y are given by
x^' Cy) = Xj(y)f); K* ln(SjA/S^F)/ln(S^/s£EF) ]/£ R* (4)
ft
where K. is the normalized absorption coefficient of
i v
gas i for channel j and S. and S. are the
i t h
actual simulated signal and the K— approximation to
that signal-
Receiver noise effects have been included in the
simulations by perturbing the calculated radiances by
an amount related to the expected IHS noise level N
where
Errors due to temperature profile bias, uncertainties
in interfering gas concentrations, and ground bright-
ness temperature variations have also been included.
As a result of these studies, pollutant channelsof
500 MHz bandwidth centered at +550, +1050, +1550, and
+2050 MHz from the R(18) 1,0 transition were chosen for
the measurements. In the NR mode calculations only
the +550 and +1550 MHz channels were used; in the SA mode
all four channels were included. A single 1 GHz chan-
nel centered at +505 MHz from the R(8) LO transition
was chosen for the reference channel. For the 0 NR
simulations, two 500 KHz pollutant channels centered
at +350 and +1850 MHz from the R(24) LO transition
and a .1 Cllz reference channel centered at +505 Mhz
from the 8.(40) 10 transition were employed. For the
0^ SA mode, five pollutant channels centered at +305,
+925, +1175, +1375, and +1550 MHz from the R(10) 10
transition with bandwidths of 600, 250, 250, 150, and
100 MHz, respectively, were chosen. A 600 MHz band-
width reference channel centered at +305 MHz from the.
P(18) 10 transition was also selected. Post-detection
integration times of 1 and 10 seconds were used for the
SA and NR mode simulations, respectively.
Calculations also have provided evidence of the
utility of reference channels in both operating modes.
In the SA mode, the reference channel is used to
ratio the pollutant channels resulting in signals of
of the form S. = S.'' /'i and should allow nearly
J 3 KM J
complete cancellation of continuum and near-continuum
attenuations and interfering gas effects. in the NR
mode simulations, the reference channel should allow
the effects of ground brightness temperature variations
to be minimized. Use of the reference channel in the
NR mode modified Equation (1) to the form
..NR
cti
if ¥0)V0)
10
"REF^REF^
C - dy
REF
Jt
'0 "REF
REF
+ /o B.(y)3-r.(y)/3y dy
(7)
The simulated radiance and profile retrieval calcu-
lations have provided considerable insight into the
feasibility and sensitivity of the proposed IHS measure-
ments. Initial IHS experiments will consist of NH^ and
0 measurements from the ground (SA mode). Figures 4,
5, and 6 show results of some of the simulated retrieval
calculations for these measurements and illustrate the
expected sensitivities and effects of measurement errors
due to (a) measured receiver noise, (b) interfering
3
10-5
-------
gases, and (c) temperature profile bias errors;. The
channel locations and bandwidths, LO transitions, and
system integration times used in these calculations
were those listed in this and the preceding section.
Disagreement exists regarding Che concentration
and vertical distribution of atmospheric NH.^, but con-
centrations are generally expected to be in the .5 to
20 ppb range near the ground with decreasing concentra-
tion through the troposphere and rapidly decreasing
concentration above the tropopause. ^0, 2.1 Figure 4
shows simulated SA. retrieval envelopes (indicated by
the hatched areas between dotted lines) for three NH^
test profiles (indicated by the solid lines), with
ground concentrations ranging from 1.6 ppb to 12.9 ppb,
which have been chosen to incorporate expected profile
features. The retrieval envelope widths indicated in
Figure 4 are of the order of 0.5 ppb or less and are
due primarily to assumed temperature profile bias
errors of +5 K and receiver noise. The simulation cal-
culations indicate that good average mixing ratio pro-
files of NH.j in the troposphere could be obtained from
SA measurements with this degree of precision for NH^
concentrations in the 1 to 20 ppb range. A small
retrieval bias toward low concentration values is
evident near the ground and has been attributed to a
less than complete cancellation of interfering gas
effects fciy the ratio technique.
Figure 5 illustrates that the SA measurements
should be sensitive to enhanced HH^ concentrations near
the ground. The left-hand portion of Figure 5 shows a
retrieval envelope for a test NH^ profile of uniformly
decreasing magnitude throughout the troposphere. In
the right-hand portion of the figure, a retrieval
envelope for a similar NH^ profile with an enhanced
near-Earth concentration is shown. The retrieval
envelope on the right-hand side exhibits a different
shape in the Lower troposphere and a higher ground
concentration than the retrieval envelope on the left-
hand side. These results indicate that the SA meas-
urements provide a sensitivity to near-Earth concentra-
tions which is lacking in passive thermal radiance
techniques such as the NR measurements. More accu-
rate retrievals of enhanced concentration profiles
could be obtained by incorporating some prior knowledge
of profile shape into the initial guess. Retrieval
calculations for all of the simulations, both SA
and NR, used an initial guess of a constant 25 ppb
mixing ratio.
The left-hand side of Figure 6 shows typical 0^
retrieval calculation results for SA measurements of
a midlatitude summer 0 profile with a ground concen-
tration of 0.03 ppm, while the right-hand side of the
figure shows retrievals of a similar 0^ profile with an
enhanced concentration in the lower 5 km and a ground
concentration of 0.20 ppm. Good average mixing ratio
profiles are obtained in both cases with retrieval
spreads of the order of 0.025 ppm resulting from +5 K
temperature profile bias errors and receiver noise.
The right-hand side of Figure 6 also illustrates the
effect of initial guess profile shape on the retrievals
with two cases, designated by 1 and 2, being shown. In
case 1 an initial guess identical to that used in the
retrieval on the left-hand side of the figure yields a
retrieval envelope which, when compared to the normal
profile retrieval in the left-hand portion, shows a
different lower tropospheric shape and a higher ground
concentration. This indicates that the retrieval is
sensitive to the enhanced lower 5 to concentrations.
Case 2 illustrates the additional retrieval accuracy
possible when prior knowledge of the profile shape is
incorporated into the initial guess profile.
The HIS also has the capability of detecting
and 0 from an airborne platform by viewing upwelling
thermal radiances (NR mode). Typical results of
retrievaL simulation calculations for NR measurements
of NH.j and 0.^, respectively, from a 10 km platform
altitude are shown in Figures 7 and 8. For these cal-
culations, potential measurement errors due to
(a) measured receiver noise, (b) +10 K ground bright-
ness temperature variations, (c) +25% H^O eoiicentration
uncertainty, and (d) +2 K temperature profile bias have
been included. These limits should represent the
extremes of the expected error sources. As before,
test profiles are indicated by solid lines and initial
guess profiles, and in the case of Figure 8, by large
dotted Lines. Retrieval envelopes are indicated by the
hatched areas.
Figure 7 shows simulated NR retrieval envelopes
for three NH^ test profiles of the type shown in Fig-
ure 4. Results indicate that receiver noise and temper-
ature profile bias errors would limit NR measurements
of HII^ to average mixing ratio measurements of middle
and upper tropospheric concentrations with precisions
of approximately 2 ppb. Effects of ground brightness
temperature variations and interfering gases have been
minimized by using the reference channel (see Eq. (7)).
The measurements would not be sensitive to near-Earth
concentrations due to the inherent dependence of passive
tLiermal radiance measurements upon a temperature differ-
ence between the background (the Earth in this case.)
and the emitting gas.
Figure 8 illustrates NR retrieval results for a
midlatitude summer 0^ test profile with and without
enhanced low-altitude concentrations. A small measure
of 0.j altitude profiling is obtained along with some
differences between the retrievals of the normal and
enhanced concentration profiles. Generally, however,
NR measurements of 0^ from altitudes of 10 km or less
will be average mixing ratio measurements with
retrieval shapes being highly dependent upon the initial
guess profile sLiape. Measurement precisions of the
order of 25 ppb are indicated with receiver noise and
temperature profile bias being the limiting factors.
Concluding Remarks
Use of an I.HS to remotely monitor concentrations
and vertical distributions of atmospheric gases has
been discussed, and a 9- to ll-pm dual laser IMS which
lias been developed for atmospheric monitoring applica-
tions has been described. Analytical results have been
presented which simulate tl*e IHS operating in a solar
viewing mode from the ground and in a nadir thermal
radiance viewing mode from a 10 km altitude platform.
Measurement precisions of 0.5 to 2 ppb for and 20
to 30 ppb for 0^ are predicted. The results indicate
that the solar view measurements will yield good average
altitude profiles of and 0^ mixing ratios, while the
nadir radiance measurements will result primarily in
average mixing ratios.
Similar measurements for other gases such as
and 11^,0 which have signature lines between 9- and 11-pm
4
10-5
-------
also appear to be feasible using isotope CO, laser LO's.
Use of tunable laser sources,12,14,15,16 such as diode
lasers, as LO's appears possible and would extend SA
type measurements to other atmospheric gases and other
port-ions of the infrared.
Additional measurement modes such as solar oeculta-
tion and Earth limb ra iiance measurements also appear
feasible as a result of initial analyses. Solar
occultation measurements of stratospheric minor con-
stituents such as HMOy CIO, and the chlorofluoro-
methanes which have signature lines in the 9- to ll-um
region appear particularly attractive. This type of
measurement would take full advantage of the ultra-fine
spectral resolution of the IHS since stratospheric
signature lines have half widths of the order of
0.01 cm ^ or less. Preliminary calculations simulating
stratospheric solar occultation measurements of NH^ with
a satellite or shuttle IHS indicate sensitivities as
low as 0.01 ppb. Provided suitable laser overlaps (or
tunable LO's) are available, measurement sensitivities
of this order should be possible for a number of
important stratospheric minor constituents. For
example, preliminary investigation indicates that the
C130^6 R(10) or K(12) transitions at 921.675 cm ^ and
-1
923.111 cm way overlap signatures lines of CC-^F^
(Freon-12) sufficiently to provide potential solar
occultation sensitivities as much as a factor of 2
better than those of the NH^ simulations mentioned
above.
References
^Hanel, R. A., et al., "The Nimbus 4 Infrared
Spectroscopic Experiment,'" J. Geophysics Res., 77,
May 1972, p. 2629.
2 ,,
Barringer, A. R., Optical Correlation Spectrom-
etry for Remote Sensing of Air Pollution," J. Opt. Soc.
Amer., 60, May 1970, pp. 729-730.
3Teich, M. C., "Infrared Heterodyne Detection,"
Proc. IEEE, 56, Jan. 1968, pp. 37-46.
/,
Peyton, B, J., et al., High Sensitivity
Receiver for Infrared Laser Communication," IEEE J.
Quantum Elec., QE-8, Feb. 1972, pp. 252-263.
"'Arams, F., et al., "Infrared 10.6 Micron
Heterodyne Detection With Gigahertz IF Capability,"
IEEE J. Quantum Elec., QE-3, Nov. 1967, pp. 484-492.
^McElroy, J. H., "Infrared Heterodyne Solar
Radiometry," Appl. Optics, 11, July 1972, pp. 1619-1622.
^King, S. ec al., "High Resolution
Atmospheric-Transmission Measurements "riing a Laser
Heterodyne Radiometer," Appl. Optics, June 1973,
pp. 110&-U07.
g
deGraauw, Th., and H. Van de Staadt, "Infrared
Heterodyne Detection of the Moon, Planets, and Stars
at 10 Um," Nature Fhy. Sc., 246, Dec. 1973, pp. 73-75.
9
Peterson, D. W., et al., "Infrared Heterodyne
Spectroscopy of C0„ on Mars," Nature, 250, July 1974,
pp. 128-130.
i0Johnson, M. A., et al., "10 ym Heterodyne
Stellar InterferometerPhys. Rev, Letters, 33,
Dec. 1974, pp. 1617-1620.
" "DINardo, A. J., et al., "An Infrared Heterodyne
Radiometer for High Resolution Measurements of Solar
Radiation," IEEE J. Quantum Elec., QE-9, June 1973,
p. 662.
^Mumma, M., et al., "Infrared Heterodyne Spectro-
scopy oi Astronomical and Laboratory Sources at
8.5 t/m," Nature, 253, Feb. 1975, pp. 514-516.
^Menzies, R. T, , and M. S. Shumate, "Air Pollu-
tion: Remote Detection of Several Pollutant Gases With
a Laser Heterodyne Radiometer," Science, 184, April
1974, pp. 570-572.
^'Hinkley, E. D., and R. H. Kingston, "Remote
Heterodyne Detection of Gaseous Pollutants With Tunable
Lasers," A1AA Paper 71-1079, Palo Alto, California,
1971.
15
Seals, R. K., Jr., "Analysis of Tunable Laser
Heterodyne Radiometry: Remote Sensing of Atmospheric
Gases," AIAA J., 12, Aug. 1974, pp. 1118-1122.
^'Seals, R. K., Jr., and C. H. Bair, "Analysis of
Laser Differential Absorption Remote Sensing Using
Diffuse Reflection From Earth,", Second Joint Conf. on
Sensing of Environmental Pollutants, ISA JSP 6675,
Dec. 1973, pp. 131-137.
^Menzies, R. T., and M. T. Chahine, "Remote
Atmospheric Sensing With an Airborne Laser Absorption
Spectrometer," Appl. Optics, 13, Dec. 1974, pp. 2840-
2849.
IB
Peyton, B. J., "Atmospheric Monitoring Using
Infrared Heterodyne Radiometry," Proc. of the Inter-
national Telemetering Conf., X, Oct. 1974, pp. 403-411.
19
Freed, C'. , et al., Determination of Laser Line
Frequencies and Vlbrational-RotatiOijal Constants of the
12 18 1316
C 0^, C Isotopes From Measurements of CW Beat
Frequencies With Fast HgCdTe Photodiodes and Microwave
Frequency Counters," J. Mol. Spectrosc., 49, March
1974, p. 439.
20
Hitchcock, 0. R., and A. E. Wechsler, "Biologi-
cal Cycling of Atmospheric Trace Gases," NASA CR-126663,
1972.
21
McC'onnell, J. C,, "Atmospheric Ammonia," J. of
Geophy. Res., 78, Nov. 1973, pp. 7812-7821.
22
Marx, J. I,., "Air Pollution: Effects on Plants,"
Science, 187, Feb. 1975, pp. 731-733.
23
"Air Quality Criteria for Photochemical Oxidants,"
Ch. 3, U.S. Department of Health, Education, and Welfare,
Washington, D.C., 1970.
24
McClatchey, R. A., et al., AFCRL Atmospheric
Absorption Line Parameter Compilation," Environmental
Research Paper 434, AFCRL-TR-73-0096, Jan. 1973.
25
Taylor, F. W., "Spectral Data for the v2 Bands
of Ammonia With Application to Radiative Transfer in
the Atmosphere of Jupiter," J. Quantum Spectrosc.
Radiat. Transfer, 13, 1973, pp. 1181-1217.
26
Frank, and R. K. Seals, Jr., "Measure-
1 •> "It
ments of NH Absorption Coefficients With a C o"
Laser," Appl. Optics, 14, Sept. 1975.
5
10-5
-------
27
Bignell, K. J., "The Water-Vapor Infrared
Continuum," Quant. J. R. Met. Soc., 96, 1973, pp. 390-
403.
'"Peyton, B. J., et al., "An Infrared Heterodyne
Radiometer for High-Resolu trlon Measurements oi Solar
Radiation and Atmospheric Transmission," IEEE
I. Quantum Elec., QE-H, Aug. 197 5.
Ct30,16 t-ASER TRANSITIONS
D
transmittance
16 18 20 22 24 26 28
10 12 14 16 18 20 22 24
935 1025
WAVE NUMBER, cm
Figure 1. Calculated atmospheric transmittance spectra
illustrating overlap between laser transitions
and NK.J and 03 signature lines (0 -~ 10 km zenith path) .
1.0
TRANSMITTANCE
.4
112 Jl4
1: yL0 t 205OMHZ
(- 2: vL0 * 1550 MHz
3-. y l 1050 MHz
4; ulQ t 550 MHz
|J - 500 MHz
altitude;
km
J_
-j i
927.22
927.30
WAVE NUMBER, cm
927.38
-1
0 .<1 .8 1.2
ABSORPTION COEFFICIENT
(NORMALIZED!
Figure 3. tUS channelization of atmospheric transmit-
tance and the resulting channel absorption coeffi-
cients (weighting functions) for MH.j.
iMcident radiance
DICKE $VVITCH\ ill
REFERENCE
SOURCE
phqtomixerH™ Nr
m\
-------
ALTITUDE, km
TEST PROFILE
RETRIEVAL ENVELOPE
8 10 0
NH CONCENTRATION, ppbv
ALTITUDE, km
8 10
Figure 5. Simulated IHS retrievals of profiles
with nonenhanced and enhanced near-Earth concen-
trations (SA mode).
-TEST PROFILE ZXZRETRIEVAL ENVELOPE
NH3 CONCENTRATION, ppbv
12 14
Mgure 7. Simulated Niiretrievals for the HIS
(NR mode).
-TEST PROFILE . '.RnRhVAi
-INITIAL GUESS ENVELOPE
PROFILE
ALTITUDE, kin
10 .01
03 CONCENTRATION, ppmv
Figure 6. Simulated IHS retrievals of O.^ profiles with
normal and enhanced near-Earth concentrations (SA
mode) .
ALTITUDE, km
-TEST PROFILE
-INITIAL GUESS
PROFILE
- 1 RETRIEVAL
• I ENVELOPE
1 .01
0^ CONCENTRATION, ppmv
Figure 8. Simulated 0^ retrievals for the IHS
(NR mode).
7
10-5
-------
REMOTE SENSING OF ATMOSPHERIC SO? USING THE
DIFFERENTIAL ABOSRPTION LIDAR TECHNIQUE
James M. Hoail, Jr., and William R. Wade
NASA Langley Research Center
Hampton, Virginia 23665
and
R, T. Thompson, Jr.
Old Dominion University
Norfolk, Virginia 23508
Summary
This paper presents performance characteristics
obtained from the calibration of a OT differential
absorption LIDAR (DIAL) system, and describes the
application of this system to quantitative measure-
ments of sulfur dioxide emissions from a local steam
generating plant. The UV DIAL system utilized a com-
mercial flashlamp-pumped frequency doubled dye laser,
a 0.25-m-diaroeter receiver telescope, and a mini
computer for controlling data acquisition and process-
ing. The output of the laser was typically 100 uj per
pulse at 15 pps for wavelengths coincident with the
peaks and valleys in the sulfur dioxide absorption
spectrum from 296.0 to 300.1 nm. The difference in
absorption coefficient for each pair of transmitted
wavelengths was monitored using a satellite beam from
the laser.
The calibration technique utilized during this
investigation provided a convenient means for deter-
mining the systematic and random uncertainties asso-
ciated with the retrieval of known concentrations of
S02- These uncertainties reflected errors introduced
by the data acquisition system, and atmospheric fluc-
tuations occurring during the time required to acquire
the return signals for the two pairs of wavelengths
(I.e., typically 15 sec). Results obtained from the
calibration verify the linearity of ttie present system
with respect to the measurement of the total burden of
sulfur dioxide and illustrate the capability of the
present system to measure the average concentration of
sulfur dioxide out to a range of 1 km with sensitivi-
ties less than 15 ppb. Moreover, it is shown that the
performance characteristics obtained here can be used
to determine what mod Ifications to the present system
would result in sensitivities compatible with given
measurement needs.
As a demonstration of the DIAL technique, quanti-
tative measurements of the average sulfur dioxide
concentration were obtained in a region surrounding
the exhaust stack of a local steam generating plant.
These measurements, performed at night, determined 'the
average concentration of sulfur dioxide out to ranges
varying from 0.8 km to 1.9 km with sensitivities vary-
ing from approximately 10 ppb at 0.8 km to 20 ppb at
1.9 km. The maximum observed concentration was 150 ppb
over 1.2 km. Variations in the measured sulfur dioxide
concentrations correlated well with variations in wind
direction and scheduled variations in sulfur dioxide
production by the steam generating plant.
Introduction
Sulfur dioxide has long been recognized as a major
manmade atmospheric pollutant which contributes to the
formation of aerosols, damage to vegetation and build-
ing materials, and In several of the more celebrated
pollution episodes (i.e., London, 1952; Donors, Pa.,
1948; and New York, 1963) has contributed to numerous
deaths.1>2 While several techniques are presently
available for in-situ detection of sulfur dioxide,
there is a pressing need for an operational system
which will permit remote measurements of the integrated
concentrations of sulfur dioxide (i.e., column content)
or the distribution of sulfur dioxide along a given
measurement path (i.e., range resolved). One of the
more promising methods for providing both types of
measurements for sulfur dioxide (as well as other
atmospheric gases such as ozone and nitrogen dioxide)
is an active laser technique known by the acronyms
DIAL (for differential absorption LIDAR), DAS (for
differential absorption and scattering), or BASE (for
differential absorption via scattered energy). All
of these acronyms refer to a measurement technique,
hereafter referred to as DIAL, which attempts to com-
bine the high sensitivity of absorption systems with
the range resolved capabilities of LIDAR systems.
This technique utilizes a pulsed laser which is tuned
to two specific wavelengths, corresponding to the maxi-
mum and minimum in the absorption spectrum of the
particular gas of interest. As the pulsed laser
radiattow. travels through the atmosphere, Rayleigh and
Mie backscattered radiation is measured as a function
of range through the use of standard LIDAR range gating
techniques. Differences in the backscattered energy
from the two transmitted wavelengths can be related to
the column content of the absorbing gas between the
transmitter and the scattering volume. The distribu-
tion of the absorbing gas as a function of range can be
obtained from the difference between column content
measurements obtained at different ranges.
Since Schotland's original work,3 results from a
number of analytical^-9 ^ncj experimental^®-^ studies
have appeared in the literature. The analytical
studies have, in general, indicated that for range
resolved pollution monitoring the DIAL technique is
superior to other LIDAR techniques such as Raman or
fluorescence scattering and that the DIAL technique is
capable of yielding high measurement sensitivities for
concentrations and ranges useful for remote pollution
monitoring. The experimental studies appearing in the
literature have consisted of measurements of nitrogen
dioxide, sulfur dioxide, and ozone in a large calibra-
tion tank;12,13 the measurement of nitrogen dioxide
over a metropolitan area;ll and the measurement of the
vertical distribution of water vapor.3,10 general,
the results from these experimental investigations have
confirmed the potential of the DIAL technique for
pollution monitoring.
The purpose of this paper is to present results
obtained from an investigation of a UV DIAL system
which illustrates the capability of this technique for
remotely measuring sulfur dioxide in the atmosphere.
The discussion here covers two distinct, but related,
phases of the investigation. The first is a calibra-
tion technique for studying systematic and random
uncertainties associated with a measurement of sulfur
dioxide; the second is the application of a calibrated
1
10-6
-------
UV DIAL system to monitoring the exhaust emissions
from a local steam generating plant. Results obtained
from the calibration served to verify the linearity of
the present system with respect to the retrieval of
total burdens of sulfur dioxide out to various ranges,
and illustrate the capability of this system to meas-
ure the average concentration of sulfur dioxide out to
a range of 1 km with sensitivities less than 15 ppb.
It is shown that performance data obtained during the
calibration not only characterizes the capability of
the present DIAL system for measuring the total burden
of sulfur dioxide as a function of range but also can
be used to estimate the errors associated with the
measurement of the distribution of sulfur dioxide for
a variety of applications (i.e., ambient air monitor-
ing, stack plume monitoring, etc.). Furthermore, this
performance data can be used as a guide for realistic
scaling of the present system's parameters in order to
yield a measurement sensitivity compatible with a given
measurement need.
As a demonstration of the DIAL technique, quanti-
tative measurements of sulfur dioxide were obtained in
the region surrounding the exhaust stack of a local
steam generating plant. Results obtained over a period
of 4 nights correlated well with variations in wind
direction and scheduled variations in sulfur dioxide
production by the steam plant. The sensitivity of this
system for the measurement of the average concentration
of sulfur dioxide varied from 10 ppb at 0.8 km to
20 ppb at 1,9 km. These results represent the first
use of the DIAL technique for monitoring atmospheric
sulfur dioxide and, combined with the calibration
results, demonstrate the potential of the DIAL for
remote pollution monitoring.
Discussion of DIAL Technique
The DIAL technique provides two types of remote
measurements which are of interest for pollution moni-
toring. One is the total burden or column content of
a gas from the transmitter to range, Rj. The second
is the distribution of the gas along the measurement
path obtained from a set of range dependent total
burden measurements. An analytical study of the DIAL
technique for obtaining both the total burden and the
distribution of sulfur dioxide is discussed in Refer-
ence 8. In this section we will briefly summarize the
discussion given in Reference 8 for the development of
the expressions relating the signals measured by an
operational DIAL system to the total burden and range
resolved distribution of a gas.
The equation relating the total burden, or column
content, Mj, of the absorbing gas to the backscattered
laser radiation from a scattering volume at range Rj
is given by
M. =
J
1
2Ao,
2 /
J r
(In
V
R,
VVi
P2j/u2X2
In
[Af'l,2+ I
k
ei
"" i
j
(i)
where Ao'j. is the difference between the absorption
coefficients at wavelengths A-^ and A2 for gas k
with k = 1 designating sulfur dioxide; j is the
laser backscattered energy for wavelength ^ from a
scattering volume located at a distance Rj from the
transmitter; is the transmitted energy^at A
i>
3i is the total backscattering coefficient at A^
associated with the scattering volume; A£
1,2
is the
difference between the Rayleigh and Mie extinction
coefficients at A^ and A2; jf Acj^ n^ represents the
attenuation coefficient for any other atmospheric
constituents which absorb at Aj_ and A2; and n^ is
the density of gas k. (Note that
M. = f n, dr).
j ' o I
Equation (J) is exact and assumes that all the
parameters are either known or can be determined during
a measurement. The second and third terms involving
Si» 1,2 > a"d ) ^aknk' however, depend upon atmos-
k
pheric conditions (i.e., aerosol and gas concentration)
along the measurement path which, in general, are not
adequately known for the evaluation of these terms.
Further complicating the use of this expression is the
fact that, the signal measured at the receiver,
(for wavelength A-j and range Rj) is the sum of the
laser backscattered energy ^i,i and the noise energy,
Nj_, associated with background light and detector
noise. Consequently, in an operational system Equa-
tion (1) is replaced by
V
1
2Aa.
In
(S
ii.
V/(uiV!
(S,;
2j
N2>/(u2A,)
(2)
which contains parameters which are directly measurable
at the receiver of a DIAL system. The numerator and
denominator within the log term represent a background
corrected, energy and wavelength normalized back-
scattered signal. It should be noted that Mj repre-
sents an approximation to Mj. However, for wavelength
pairs which are closely spaced, the second and third
terms in Equation (1) are approximately zero and
consequently Equation (1) reduces to Equation (2). In
Reference 8 the optimum wavelengths for minimizing the
difference between Mj and Mj were determined and
are used in this study.
The expression for determining the distribution
of the gas along the measurement path is given by
.1+1/2
(Mj+l
Mj ) /AR
(3)
where AR = Rj+x
betw
Rj, and nj+i/2 the average con-
centration between range and R j. For the dis
cussion of the calibration results given below, it is
important to note that the basic quantity measured by
the DIAL technique is the column content, with I1|+l/2
being determined from a set of range dependent Mj s,
Moreover, the uncertainty associated with Hj-t-i/2'
'^nj+l/2> Is related to the. uncertainty in Mj and
Mj+1> and is given by
611
j+1/2
(<$Mj2 + <5Mj+12)1,'2/M
(4)
where ^Mj^
and
6Mj are the errors associated with
the measurement of the column content out to Ri+] and
Ri
Rj+1
respectively, and where it is assumed that the
fluctuations in
evaluation of the SM's
and Mj are uncorrelated. The
will be discussed below.
One of the goals of this study was to investigate
the systematic and random uncertainties associated
with the use of Equation (2) for a UV DIAL system using
commercial components. This would permit an estimate
of <5nj+i/2> via Equation (4) for a number of applica-
tions. In the following sections we describe the DIAL
system used in this study, and a simple calibration
technique to determine the systematic and random
uncertainties associated with the retrieval of known
column contents of sulfur dioxide.
1.0-6
-------
Description of DIAL System
Transmitter and Receiver Components
A schematic of the DIAL transmitter and receiver
section is shown in Figure 1(a). This system was
constructed from components designed for laboratory
use, and consequently was operated from a laboratory
room using an open window as an access to the atmos-
phere. The UV dye laser was a frequency doubled
flasblamp-pumped dye laser (Chromatix Model CMX-4),
For the work described here, the laser was operated at
15 pps at an output energy of 100 )jj per pulse. The
spectral bandwidth was less than 0.03 ran and the pulse
width was 1.3 (is. Tuning of the UV output required the
adjustment of two elements. The first was a birefring-
ent crystal used to select the fundamental output of
the laser. The second was an intracavity nonlinear
crystal which was angle-tuned for phase matching with
the fundamental wavelength. A calibrated stepping
motor was used to tune the birefringent crystal while
the nonlinear crystal was tuned by hand. With this
arrangement, the time required to switch wavelengths
was typically 15 seconds. The wavelengths, in nm,
used in this study and their absorption coefficients,14
given in parentheses in reciprocal atm~cm, were 300.05,
(32.7); 299.3, (7.4); 298.0, (27.1); 297.4, (8.0); and
296.25, (28.1). Using these five wavelengths, four
combinations of "on" and "off" wavelengths were used.
These were 300.05, 299.3; 298.0, 299.3; 298.0, 297.4;
and 296.25, 297.4. Through the use of the beam steer-
ing mirrors Ml and M2 the transmitted laser
radiation was alined coaxially with the receiver field
of view. Approximately 1% of the UV output was
diverted from the main beam by beam splitter BS1 with
the remaining energy being passed through the calibra-
tion cell before transmission through the atmosphere.
A detailed description of the use of this calibration
cell is given in the Discussion of Results. The 1%
diverted out of the main path was further divided by
beam splitter BS2 and directed onto detector D1
and through the sulfur dioxide absorption cell onto
detector D2 (see Fig. 1(a)). The absorption cell
contained a known concentration of sulfur dioxide
(i.e., typically 5 torr). This arrangement permitted
the energy for each pulse to be monitored via Di, and
the difference in the sulfur dioxide absorption coef-
ficients to be determined for each pair of wavelengths
by comparing the output of D.l and D2. This absorp-
tion ceil is an important feature for atmospheric
measurements, since it eliminates the need for
assuming wavelength repeatability, and consequently
minimizes the errors which would be introduced by
variations in atmospheric absorption associated with
shot-to-shot fluctuations in the transmitted wave-
lengths. The use of this feature reduces the need for
a highly stabilized dye laser and at the same time
provides a monitor which can be used for in-the-fleld
adjustment of the transmitted wavelengths.
The receiver, shown in Figure 1(a), was a 0.24-m-
diameter Da 11 Kirkham telescope with optics coated for
a total optical efficiency greater than 70% for wave-
lengths from 290.0 nm to 700.0 nm. The backscattered
energy was measured through the use of an interference
filter and PM tube mounted at the focus of the tele-
scope. The PM tube was an EMI 9S58QB with an S-20
response. The results presented here were obtained
during night operation, since a UV filter was not
used; however, daylight operation is feasible with the
use of a UV filter having a bandpass <2.0 nm. For the
night operations, a low pass color filter was used to
eliminate background light. This filter had a trans-
mission of approximately 90% from 296.0 im to 305.0 nm,
decreasing in transmission to leas than 0.1% for wave-
lengths exceeding 340.0 nm. This reduced the back-
ground noise essentially to zero leaving only PM tube
noise.
Data Acquisition and Processing
The basic components of the data acquisition and
processing system are shown in Figure 1(b). The indi-
vidual components that make up this system are a fan-
out unit, an eight-channel gated sample-and-hold
module with a time multiplexed output, an analog-to-
digital converter, and a PDP-8 mini computer with three
output units. The fan~out unit contained six variable
gain output channels which, except for the difference
in gain, provided six replicas of the input to the fan-
out. The gain of each output channel of this unit was
adjusted to minimize the dynamic range seen by the
sample-and-hold module. The range dependent back-
scattered signals were obtained by gating the six range
channels of the sample-and-hold unit at the desired
delay time, tfl. The length of the scattering volume
was determined by the gate width, which for this work
was set at 1 Us. "this combined with the 1.3 Us laser
pulse gave an effective scattering volume 300 meters
Ion g.
Except for the determination of the wavelength,
the system was designed to automatically record and
process all the basic parameters necessary for the
determination of the total burden of sulfur dioxide via
Equation (2). The first two channels were used to
monitor the output from 01 and D2, and the other
six channels monitored the laser backscattered energy
as a function of range. For each data acquisition
cycle, the analog contents of the eight channels of
the sample-and-hold unit were digitized and stored in
a PDP-8 mini computer. The gates to the system were
timed so that for each laser shot there were two com-
plete data acquisition and storage cycles. The first
cycle monitored the laser energy via Dl, the
attenuation due to the absorption cell via D2, and
the range dependent return signals plus noise, S-y.
The second cycle, which was initiated 16 milliseconds
after the laser pulse, monitored the background noise,
For a measurement of the total burden, a set of
such data was recorded for each wavelength and either
stored on a magtape unit or processed in the PDP-8
mini computer with the results displayed from the
teletype or CRT units.
Because of the time lag between recording data for
each wavelength, the column content measurements
reported here were determined from an average of the
return signals for each wavelength, which were back-
ground corrected and energy and wavelength normalized,
that is, <(S-^j - Ni)/(ujA-^)>. This average normalized
signal was computed from the backscattered signals,
S^j, and noise signals obtained from 100 succes-
sive laser firings with the laser operating at a rate
of 15 pps. The uncertainty associated with each meas-
urement of the total burden was determined by consider-
ing the propagation through Equation (2) of the stand-
ard deviation associated with the average of the
normalized return signals. The total burden calculated
in this manner represents a time average of the sulfur
dioxide concentration, which for a "reasonably stable"
atmosphere would be identical to a time average
obtained from a set of instantaneous values of the
totaL burden. The uncertainty associated -with the
present calculation of Mj reflects the same, system
errors that would be associated with instantaneous
measurements of Mj plus errors introduced by the
atmospheric fluctuations during our recording time.
3
10-6
-------
Discussion _of Results
Ca 1 ibrat ioji
in this section a description of the calibration
technique and a discussion of the results obtained
using this technique will be given. The technique
described here utilized a known concentration of sulfur
dioxide in the calibration cell illustrated in Fig-
ure 1(a) to provide a differential attenuation in the
energy of the transmitted wavelengths. The difference
in attenuation for two different wavelengths depends
on the cell length, the pressure of the sulfur dioxide
within the cell, and the difference between the
absorption coefficient for the two wavelengths. Since
the column content measured by a DIAL system is deter-
mined from the difference in the backscattered energy,
the differential attenuation produced by the sulfur
dioxide in this cell (and consequently in the back-
scattered energy) effectively simulates a column con-
tent of sulfur dioxide. In the absence of sulfur
dioxide in the atmosphere, this eeli simulates a known
column content of sulfur dioxide, which is given by
M = PcLc/2 where Pc is the sulfur dioxide pressure
in atm and Lc is the cell length in cm. The factor
of 1/2 accounts for the fact that only the transmitted
beam passes through the cell. The dimensions of this
calibration cell were 5.0 cm long by 2.0 cm in diam-
eter. The cold finger shown in Figure 1(a) was used
to control the sulfur dioxide pressure within the cell,
Typical results, illustrating the linearity of the
system, are shown in Figure 2 where the simulated
column content is plotted against the column content
retrieved through the use of Equation (2), and the
backscattered energy measured from a scattering volume
1.2 km from the transmitter. The results shown here
are for the wavelength pair 300.05 and 299.3 nm. The
procedure for determining the uncertainties associated
with retrieving the simulated column content was
described in the previous section. It should be noted
that for a fixed sulfur dioxide pressure in the cali-
bration cell, and with no atmospheric sulfur dioxide
present, the retrieved column content is constant with
range. Within the uncertainty associated with the
retrieval from each range, this was found to be true
for this system. From Figure 2 it is clear that the
linearity of the system is excellent, although there
is a tendency toward a nonzero intercept which is
indicative of the presence of a systematic error.
Such an error may be related to the use of the approxi-
mate expression for total burden (i.e., Eq. (2)) or a
small difference in the response of the present optical
system for the wavelength pairs used. The magnitude of
this error was about 1 x 10~*3 atm-cm, with similar
results obtained for other ranges and wavelength pairs.
In Figure 3 the uncertainties in the retrieval of
the simulated column content are shown with the range
of the scattering volume as a parameter. This type of
format for displaying the measurement errors illus-
trates the errors associated with the present system
for the measurement of column content, and provides a
convenient form from which the expected uncertainty
associated with the retrieval of an assumed distribu-
tion of SO2 can be calculated. For example, it is
interesting to consider how well the present system
would be able to measure the concentration in a stack
plume with boundaries at 0.8 km and 1.7 km down range.
For this example, we will assume that the column content
out. to the near boundary of the plume is 4.8x10""^ atm-
cm giving an average concentration of 60 ppb, and that
the column content within the plume is 27 x 10"* ^ atm-cm
or an average concentration of 300 ppb. This gives a
total column content of 31.8x10""^ atm-cm to the far
side of the plume. Equation (4) can now be used to
estimate ^n-j+i/2 where, the uncertainties in the
measurement of the column contents to the near and far
side of the plume are determined from Figure 3. For
the conditions assumed, the predicted uncertainty in
measuring the plume concentration is approximately
28 ppb. From this example, it is clear that the cali-
bration scheme described here provides extremely use-
ful data for realistic estimates of the measurement
capability of a DTAh system for a variety of
applications.
For some applications, the determination of the
average concent rn t i on along the measurement, path is as
important as a range resolved measurement. Figure 4
illustrates the uncertainty associated with this type
of measurement. The data for this figure were
obtained by dividing each data point: given in Figure 3
by its corresponding range to convert the simulated
column content to an average concentration in ppb.
These data indicate that for ranges out to about 1 km
the present system is capable of measurement sensitivi-
ties of less than 15 ppb for concentrations up to about
500 ppb. As expected, the measurement sensitivity
decreases as the range increases. Note also that for
a given range, the uncertainties increase as the con-
centration increases. This is due to the small signal
return for the highly absorbed wavelength. It is felt
that this effect can be circumvented by stepping the
"on" wavelength down from the peak in the absorption
spectrum thereby increasing the return signal at the
high sulfur dioxide concentrations.
The data shown in Figures 3 and 4 have been dis-
cussed in terms of the measurement capability of the
present system. It is clear that the measurement
sensitivity of this system must be improved for appli-
cations such as plume monitoring or ambient air quality
measurements. To this end, the performance data pre-
sented here provide a useful data base which can be
used to determine the improvements in sensitivity that
would result from modifications to the present system.
For example, a straightforward improvement which could
be incorporated in this system would be a higher energy
laser. A compact laser system which has been flight
testedi5 and is capable of outputs in excess of 2 mj
at pulse rates of 10 pps, a pulse width of ns, and
a bandwidth '^0.003 nm would provide an increase of
approximately 20 in output energy over the present
system. This would result in a reduction of the
uncertainties shown in Figures 3 and 4 by a factor of
about 4.5. For the plume monitoring example discussed
above, this would result in a measurement sensitivity
of about 6 ppb. For the measurement of average concen-
trations (Fig. 4) the increased output energy would
result in measurement sensitivities of less than 4 ppb
for 1 km paths. Additional modifications, such as
increased receiver area, or more efficient optics would
provide further improvements in the sensitivity of the
system.
At this point it should be noted that several
additional noise sources which are not included in the
data shown in Figures 3 and 4 must be considered. These
include the effects associated with the time lag between
the "on" and "off" wavelength, and daytime operation.
As noted previously, the present system required
approximately 15 seconds to change wavelengths. A sys-
tem having less than a 1-ms time lag would minimize the
errors introduced into the present data by atmospheric
fluctuations? and, consequently, tend to provide better
sensitivities than those predicted by the. present data.
For daytime operation, the two major sources of errors
are the increased background and interference due to
the increase in ozone concentration. It is felt the
increase in background noise will not significantly
affect the sensitivities predicted here, since through
4
10-6
-------
the use of narrow-band interference filters (i.e.,
¦ 2.0 run) conditions approaching chose experienced here
can lie achieved. The effects due Co ozone interfer-
ence will need to be considered only for predicted
sensitivities approaching 1 to 2 ppb. For example, the
error introduced by an average concentration of
100 ppb of ozone in a measurement over a 1-km path
will be less than 1 ppb for the 300.05 and 299.30 nm
wavelength pair. From the above considerations, it can
be seen that the performance indicated by the present
data represents levels which can be achieved for
general applications.
Stack Plume S0£ Measurements
The close proximity of the present laboratory-
based system to a local steam generating plant (Fig. 5)
provided an ideal situation to test the DIAL system in
a quasi-field application. Data obtained fromLangley's
Environmental Monitoring Laboratory indicate that the
ground-level concentrations of sulfur dioxide produced
by this plant sometimes reach levels in excess of
500 ppb. As indicated by the data shown in Figure 4,
the present system should be capable of quantitative
measurements of the average sulfur dioxide for these
levels.
The measurement path was oriented at an angle of
about 30° from the horizontal. The steam plant was
located approximately 130 meters down range and
190 meters southeast of this path (see Fig. 5). The
test program for monitoring the stack emissions uat
conducted over a period of 4 nights, for which time
the visibility was estimated to be typically 8 to 10 km.
Measurements of the average sulCur dioxide concentra-
tion out to ranges of 0.78, 0.99, 1.20, 1.41, 1.68, and
1.93 km were obtained every 10 to 20 minutes. The wind
direction was favorable for measuring the stack efflu-
ents during only 1 of the 4 nights of the tests.
Results from thin night (July 2, 1.975) are shown in
figure 6, while typical results from one of the remain-
ing nights (July 3, 197 5) are shown in Figure 7. In
each figure, time is plotted along the abscissa with
the measured column content in milli—atm-cm plotted
along the left ordinate and the equivalent average
concentration in ppb plotted along the right ordinate.
Data from only three ranges are shown here. The solid
curve drawn through the data in each figure serves to
emphasize its trend. in Figure 7, the measurements
were obtained with the wind from the northeast. This
tended to produce a path relatively free of sulfur
dioxide and, indeed, within the systematic error of
1 x 10~3 atm-cm Indicated by the calibration curve shown
in Figure 2, the results indicate a zero column content
at all ranges. The data shown in Figure 6 were obtained
during a period for which the wind condition proved to
be ideal for the purpose of demonstrating the measurement
capability of the. DIAL technique. At the beginning of
this night's operation, the wind was blowing the stack
effluent away from the measurement path. This is
reflected in the results prior to 2100 hours. At about
2100 hours, the direction of the wind gradually shifted
from out of the west to out of the south. In Figure 6
it can be seen that at all ranges the measured sulfur
dioxide concentration begins to increase after about
211.0 hours. At 2245 hours, the steam plant began a
scheduled cycle of operation which rapidly reduced the
sulfur dioxide emission. During the hour following
2245 hours the plant gradually returned to its original
operational conditions. There is excellent correlation
between this operating cycle and the DIAL measurements.
Notice that at about 2250 hours the measured sulfur
dioxide level began a sharp decrease which reached a
minimum 15 minutes later at 2305 hours. The sulfur
dioxide level then began to rise, reaching a maximum at
approximately 2330 hours. At this time the level
began to decrease once again. This decrease in the
sulfur dioxide level after 2330 hours is explained by
a recorded shift in wind direction such that the stack
effluent was no longer within the measurement path.
This also explains the reduced level of sulfur dioxide
at 2330 hours when compared to that measured at
2.245 hours.
For the results shown in Figures 6 and 7, It. is
important to note that the use of four different wave-
length pairs yielded similar results. This tends to
confirm that the measured gas was indeed sulfur
dioxide, since it is unlikely that an interfering gas
would have an absorption spectrum with relative maxima
and minima at all the wavelengths used here.
The uncertainties in these data are somewhat
higher than those determined via the calibration. For
example, in Figure 6 the sensitivity at 0.8 tan was
typically 10 ppb for a measurement of an average con-
centration of 90 ppb, as compared to the 6 ppb shown
In Figure 4 for the same concentration. It Is Eelt
that this is due to fluctuations in the atmospheric
sulfur dioxide concentration during the time required
to acquire hackscatterlng data for each wavelength pair.
The use of a DIAL system capable of less temporal
spacing between the wavelengths should minimize this
effect.
Concluding Remarks
An investigation of the measurement capability of
a UV DIAL .system for remote measurements of sulfur
dioxide has been described. Incorporated into the DIAL
system were two unique features. The first was an
absorption cell which permitted the determination of
the difference in sulfur dioxide absorption coeffi-
cients for the two transmitted wavelengths. The second
was a calibration cell which permitted the linearity
and performance of the system to be studied. Measure-
ment sensitivities of 10 ppb at a range of 0.8 km and
20 ppb at a range of 1.9 to were demonstrated for the
first reported application of aUV DIAL system to meas-
urements of atmospheric sulfur dioxide. From the per-
formance characteristics obtained during a calibration
of the present system, Improvements in the measurement
sensitivity to less than 4 ppb over a 1-km path are
predicted.
Acknowledgment
The authors wish to express their gratitude to
Charles B. Karpa for his expert technical assistance.
References
^Air Quality Criteria for Sulfur Oxides, National
Air Pollution Control Administration, Washington, D.C.
NAPCA Pub. AP 50, Jan. 1969.
2
Applying NASA Technology to Air Pollution: The
Sulfur Dioxide Problem, Section 2. NASA CR-100629,
1969.
"*Schotland, R. M., E. E. Chermack, and D. T.
Chang: The Measurement of the Vertical Distribution
of Water Vapor by the Differential Absorption of
Scattered Energy From a Searchlight Beam. Proc. of the
First International Symposium of Humidity and Moisture,
New York, Reinhold Book Division, pp. 569-582 (1964).
4
Measures, R. M.: A Comparative Study of Laser
Methods of Alt Pollution Mapping. University of
Toronto Institute for Aerospace Studies. Report
No. 174, Dec. 1971.
5
10-6
-------
Ahmed, Samir A.: Molecular Air Pollution Moni-
toring by Dye Laser Measurement of Differential
Absorption of Atmospheric Elastic Backscatter. Appl.
Opt., Vol. 12, pp. 901-903 (1973).
^Byer, Robert L., and Max Garbuny: Pollutant
Detection by Absorption Using Mie Scattering and
Topographic Targets as Retroreflectors. Appl. Opt.,
Vol. 12, No. 7, pp. 1496-150.5 (1973).
''schotland, Richard M. : Errors in Lidar Meas-
urements of Atmospheric Gases by Differential
Absorption. J. of Appl. Mete., Vol. 13, pp. 71-77
(1974).
g
Hoell, J. M., Jr.: Computer Simulation of an
Aircraft Based Differential Absorption and Scattering
System for Retrieval of SO2 Vertical Profiles. NASA
TN D-8077, 1975.
9
Wright, M. L., E. K. Proctor, L. S. Gasiorek,
and E. M. Liston: A Preliminary Study of Air Pollution
Measurements by Active Sensing Techniques. NASA
CR-132724, 1975.
"^Schotland, R. M. Some Observations of the
Vertical Profile of Water Vapor by Means of a Laser
Optical Radar. Proceedings of the Fourth Symposium on
Remote Sensing of Environment, Ann Arbor, Michigan,
April 12-14, 1966, pp. 273-283.
Rothe, K. W. , 1). Brlnkroarm, and H. Walther:
Applications of Tunable Dye Lasers to Air Pollution
Detection: Measurements of Atmospheric NO2 Concentra-
tions by Differential Absorption. Appl. Phys. .3,
pp. 115-119 (1974).
12
Grant, W. B., R. D. Hake, Jr., E. M. Liston,
R. C. Robbins, and E. K. Proctor, Jr.: Calibrated
Remote Measurement of N02. Appl. Phys. Lett. 24,
p. 550 (1974).
13
Grant, W. B., and R. D. Hake, Jr.: Calibrated
Remote Measurements of SO2 and O3 Using Atmospheric
Backscatter. J. of Appl. Phys. 46, pp. 3019-3023
(1975).
14
Thompson, R. T., Jr., J. M. Hoell, Jr., and
W. R. Wade: Measurements of SO2 Absorption Coeffi-
cients Using a Tunable Dye Laser. J. Appl. Phys. 46,
pp. 3040-3043 (1975).
15Davis, D. D., T. McGee, W. Heaps, A. Moriarty,
and R. Schiff: In Situ Aircraft Measurements of the
OH Free Radical: Laser Induced Fluorescence. Inter-
national Conference on Environmental Sensing and
Assessment, Las Vegas, Nevada, Sept. 1975.
CALIBRATION CLLL
V*
-•L
.?UM RECEIVER
TELESCOPE
FILTER £
PM TUBE
PM TUBE
OUTPUT
-V
l)V DYE
LASER
J 1—1 DETECTOR D1
¦z TT—H I
I
I
~
rtlSO, ABSORPT
ID CELL
[12 OUTPUT 1)1 OUTPUT SVHC TRIGGER
TO LASER
U(J I'HUl
GATED
SAMPLE
AND
HOLD
SYSTEM
T0
VARIABLE
delay
GATE
GENERATOR
OUTPUT
T0
PM TUBE.
FAN
our
SYSTEM
rl
l
rN
OUTPUT
TBiRCLR SYNCED HI [II
TIKE MULTIPLEXED OUTPUT
*l#l COHPUTER
Figure 1(a). Schematic of laboratory DIAL system:
transmitter and receiver section.
Figure 1(b). Schematic of laboratory DIAL system:
data acquisition and processing section.
6
10-6
-------
KAHf.f - ].:? KM
rfAviif.nr.rH mo', km
pAfr ru.Mm
O RANGb
A RANGE
~ RANGE
O RANGE
1.9 KM
7 KM
1.2 KM
(J.8 KM
500.05 NM
299.30 NM
WAVELENGTH
PAIR
160 2^0 320 m m 560
S1MULAFED AVERAGE CONCENTRATION (ppb)
Figure 4. The uncertainty associated with the retrieval
of the simulated average sulfur dioxide concentration.
i 9 15 '/I rl i3
SIMULATED COLUMN CONTENT Imilli-atm-cml
Figure 2. Calibration curve for the UV DIAL system.
The retrieved sulfur dioxide column content was
determined from atmospheric laser backscattered
signals originating from a scattering volume 1.2 km
down range.
O RANGE - 1.9 KM
A RANGE ¦ 1.7 KM
O RANGE - 1,2 KM
O RANGE - 0.8 KM
WAVELENGTH
PAIR
300,05 NM
299,30 NM
12 18 21) 30
SIMULATED COLUMN CONTENT (billi-atm-ch)
3
190 M
STEAM
GENERATING
PLANT
€>
STACK
LABORATORY
HOUSING
UV DIAL SYSTEM
Figure 5. Relative location of the DIAL measurement
path and exhaust stack for monitoring sulfur dioxide
stack emissions.
Figure 3. The uncertainty associated with the retrieval
of the simulated sulfur dioxide column content for
laser backscattering measured from four scattering
volumes.
7
10-6
-------
DRTE 702.1975
ON / OFF
o 236 -25/297 -40NM
~ 300-05/299- 30NM
0 298-00/299-30NM
A 298-00/297.40NM
RflNGB-1 .926 KM
O
U
2
ZT
CD
C )
RflNGE=l .205 KM
15 '
0-
5 -
0 .
-"•
C-J
1
v-
cr
LU
h~
2!
G
LJ
n
_j
o
C J
10
RANGE-1.926
KM
RANGED .205 Kh
RANGE-.783 KM
!TT^3ToTr^2T3T4"ir~F
TIME (EDT)
Figure 6. Time history of atmospheric sulfur dioxide
concentration along measurement path to three ranges
At 2100 hours wind direction shifted from out of the
west to out of the south and finally from out of the
southwest at 2400 hours (see Fig. 5). At 2245 hours
steam plant began scheduled variation in sulfur
dioxide emission.
Figure 7. Same as Figure 6 except wind direction was
out of the northwest producing a sulfur dioxide free
path.
8
10-6
-------
TOTAL ATMOSPHERIC COLUMN AND TROPOSPHERIC ABUNDANCE MEASUREMENT
OF NITROGEN DIOXIDE BY ABSORPTION SPECTROSCOPY
Watson R. Henderson
National Oceanic Atmospheric Administration
Aeronomy Laboratory
Boulder, Colorado 80302
INTRODUCTION
Nitrogen dioxide (NO2) absorbs radiation in the vis-
ible portion of the spectrum. This absorption feature
is used to identify and measure the amount of N02 in the
path between a light source and the point of observation.
Two methods have been used to observe the N02 absorption.
The first method utilized an incandescent light source
with a path length of 10 km. The second method uses the
sun as a source which allows the entire atmospheric col-
umn to be observed.
In both methods of observation a spectrum of the
light source is analyzed by least square fit to known
spectra in order to identify components of the spectra
due to NO2 absorption and measure the magnitude of this
effect. Absorption features due to water vapor and Ray-
leigh scattering must also be measured and included in
the calculation of the N02 absorption in order to remove
their interference. Absolute calibration for the number
of NO2 atoms In the observed light path is obtained by
measuring a spectrum while observing through a cell con-
taining a known N02 concentration. Measurement of N02
absorption with the sun as.a source as the sun sets al-
lows an altitude profile of N02 to be inferred.
METHOD OF OBSERVATION
Measurement of atmospheric N02 by optical absorption
has been reported by Brewer1 and later Noxon^. It has
been found that a complete spectra should be used in or-
der to identify and take account of interfering absorp-
tion features. The measurement of N02 abundance in the
observed light path therefore reduces to obtaining a
spectrum, calculating the fractional absorption in the
spectrum due to N02 and then comparing this absorption
fraction to the absorption produced by a known quantity
of N02.
Nitrogen dioxide absorbs throughout much of the vis-
ible and ultraviolet portion of the spectrum. This spec-
trum has sharp vibrational structure in the 4100 & to
4500 8 portion of the spectrum. Figure 1 is a reproduc-
tion of the spectra obtained by observing a quartz-iodide
filament lamp through a cell containing 10iB atoms/cm2
(0.037 atm. cm) of N02. A resolution of 1,5 S was used
In this spectrum. Arrows indicate the limits of scan
used in the automatic instrument described below.
Two techniques have been used to measure atmospheric
NO2. In the first technique an incandescent lamp is
used as the light source and in the second the sun is the
light source. In the first technique a quartz-iodide
lamp is columnated by a fifty centimeter diameter spher-
ical mirror and placed 10 km away from the spectrometer.
Spectra are taken in a fast scan mode and digitally av-
eraged in order to hel^ reduce the sintillation effect
caused by atmospheric turbulence from introducing added
noise to the spectral data. Each spectra is taken in 0.1
sec. and usually 10,000 spectra are averaged. Effects
which limit the precision with which a spectra can be
measured in this instrument are the variation in light
intensity during a scan due to atmospheric turbulence
along the light path and noise introduced by the detector.
In the case of the instrument described here, where the
detector is a photomultiplier, the detector noise is due
to shot noise from the photo cathode. Absorption fea-
tures of the order of 0.1% can be measured with normal
operational care. This corresponds to roughly 0.2 ppb
mixing ratio N02 in the light path described.
>-
if)
X
IU
H-
Z
o
Ul
F
£
1-
43001
4500A
WAVELENGTH
Fig. 1. Absorption features observed from
light transmitted through a cell containing
1 x 10*® N02 atoms/cm2.
The second method of observation uses the sun as a
source. The high intensity of the solar light and large
angular diameter of the solar disc increase signal to
noise ratios over that obtained in the first method but
the spectra of the sun is very complicated due to the
presence of Fraunhofer absorption features, A further
advantage of using the sun as a source comes from the
ability to observe absorption as a function of solar
zenith angle. As the sun approaches the horizon, the
total number of molecules of air in the light path bet-
ween the sun and observer increases. Multiplication,
factors (Chapman factors) for the number of molecules
in the line of sight of greater than ten as compared to
a path through zenith are easily obtained.
If the zenith sky is observed after sunset the light
will have traveled through the atmosphere from the sun
and be scattered by Rayleigh scattering one or more times
until the photon enters the spectrometer from the direc-
tion of the zenith, Whipple and Noxon^ have calculated
this effect taking into account the loss of sunlight by
scattering and N02 absorption. Their results allow post
sunset or predawn measurements of N02 absorption to be
interpreted to yield a height profile of NO2 density.
The height resolution of about five kilometers will not
reveal fine detail hut it will establish the envelope of
the NO2 profile and allow the observation of vertical
10-7
-------
motion of the layer.
DESCRIPTION OF INSTRUMENT
The basic NO2 absorption spectrometer assembly can
be divided into four sub-assemblies. These assemblies
are the scanning spectrometer, the light detector, the
analog to digital converter, and the digital averager.
If the unit is to be operated in an unattended environ-
ment, automatic gain controls and a timer must be added
in order to operate the spectrometer detection system
over a reasonable dynamic range and to allow data to be
taken when desired. The spectrometer is used to select
that portion of the light source's spectrum in a narrow
wavelength interval. An f3.5 aperture concave grating
spectrometer utilizing a holographically produced grating
is employed. The transmitted wavelength is scanned by
rotating the grating with a galvanometer type motor.
This motor is driven by a signal produced in the digital
averager which keeps wavelength synchronized with aver-
ager channel during the multi-scan averaging process.
Resolution of the instrument is set by slit width and is
20 8. Light from zenith measurements is introduced into
the spectrometer by a simple mirror but light from a dis-
tant incandescent source is collected and focused by a
fifty centimeter diameter reflecting telescope.
A blue sensitive photomultiplier is used to detect
light transmitted by the spectrometer. Cooling of the
photomultiplier was not required because dark current
only amounted to a few percent of the most sensitive
scale of the preamplifier. Signal from the photomulti-
plier is processed as an analog current and is digitized
by a voltage to frequency converter.
The digital averaging assembly is arranged to pro-
vide from one to two hundred and fifty-six possible
channels. Normally the wavelength scan is divided into
thirty-two channels corresponding to thirty-two equally
spaced wavelengths. The digital averager puts out a
voltage proportional to the channel number of observation.
This voltage drives the scan motor on the spectrometer to
a position such that the wavelength transmitted by the
spectrometer to the photomultiplier is proper for the
channel number of observation. Signal from the photo-
multiplier is averaged during the 1/32 of a scan period
allotted to each wavelength during a scan by counting
the number of pulses produced by the voltage to frequency
converter during this period. At the end of each count
interval this count is added to the value already accumu-
lated in the memory for that channel and the voltage to
the scan motor on the spectrometer is increased to bring
the next wavelength into focus. The sequence is then
repeated. After the ladt wavelength in the scan is
reached and its count period completed, the voltage is
reduced to zero. The entire scan sequence is repeated
again and again until the desired Integration time has
elapsed. When the desired Integration time has elapsed
the sum of counts accumulated for each channel is punched
out on paper tape, the sum for each channel is set to
zero and the scan sequence is started again.
Operation of the spectrometer for solar zenith angles
in the range of from eighty-five to ninety-seven degrees,
in order to observe sunrise and sunset effects, requires
that the spectrometer respond to light levels that will
change four orders of magnitude. An instrument has been
built which operates unattended by utilizing a servo
system which adjusts the intensity of light entering the
spectrometer detector so that a fixed output level to the
averager is maintained as the intensity of light changes
over more than four orders of magnitude. The servo sys-
tem consists of a motor driven neutral density optical
wedge and automatic gain change in the preamplifier which
activates when the optical wedge reaches the end of its
range. A simple time switch turns the system on for the
appropriate time interval and logic in the digital aver-
ager starts the integration sequence properly.
DATA REDUCTION
The digitized data from the scanning spectrometer
is reduced by the use of a programmed digital computer.
Data ia first normalized by dividing the magnitude of
the signal level measured for each channel by the magni-
tude of a reference signal level that would have been
obtained if no NO2 were in the path. This reference
signal level can be obtained in the case of the solar
source by extrapolating the sequence of spectra obtained
for various air path multiplication factors (Chapman
factors) to a multiplication factor of zero. Since the
solar spectra is very stable with time in this wavelength
range, this spectra need only be calculated once and
checked at intervals, as necessary, to correct for
changes in spectrometer optics. The reference spectra
for the local light source may be obtained by observing
the spectra as a function of path length and extrapola-
ting to zero path length.
The normalized spectra is then fit to a linear com-
bination of absorption spectra due to N02> water vapor,
and Rayleigh scattering. A standard least square fit
algorithm^ is used. Absolute calibration of the entire
data gathering and data reduction process is obtained by
inserting a cell containing a known density of NO2 in
the light path. If absorption of the cell is held to a
few percent, so that the linear approximation of Beer's
law holds, a linear relationship between NO2 absorption
fraction and NO2 number density in the light path can be
calculated from the increase in NO2 absorption fraction
when the cell is inserted.
RESULTS OBTAINED
The optical absorption technique of measuring NO2
in the atmosphere is proving to be a fertile new method.
Several new effects have been observed and older wet
chemical measurements appear to not have produced reli-
able measurements. Noxon2 reported tropospheric abun-
dances of NO2 below 0.1 ppb in the Colorado mountains
west of Denver when winds prevented urban produced NO2
from contaminating the local air. This NO2 density for
clean air appears to be well below earlier estimates.
NO2 density profiles as a function of height and the
vertical motion of this profile has been observed. A
large reduction in NO2 density in winter at high lati-
tude has also been discovered.
FUTURE PROSPECTS
Use of optical absorption to detect NO2 in the atmos-
phere appears to have many further applications. A com-
pact unit using a light source at the spectrometer and
a retro-reflector mounted 100 meters away could be used
to monitor urban NO2 densities in the range of 10 ppb.
A spectrometer mounted on a satellite could monitor glob-
al NO2 by observing solar light reflected from the earth's
surface. It is clear that this technique can produce
sufficient sensitivity for many applications and requires
a reasonable amount of instrument complication.
REFERENCES
1. Brewer, A. S. and C. T. McElroy, Nitrogen Dioxide
Concentrations in the Atmosphere, Nature. 246. 129, 1973.
2. Noxon, J., NO2 in the Stratosphere and Troposphere
Measured by Ground Based Absorption Spectroscopy, Science,
in press, 1975.
3. Whipple, E., and J. Noxon, to be published, 1975.
2
10-7
-------
4. Carnahan, B., H. A. Luther, and J, 0, Wilkes,
Applied Numerical Methods, pp. 571-575, John Wiley &
Sons, Inc., 1969
3
10-7
-------
FACILITATING THE MERGER OF TECHNOLOGY AND GOVERNANCE
Bernard H. Manheimer
U. S. Department of Housing and Urban Development
Washington, D. C. 20410
This session was organized to permit attending
scientists and engineers to explore methods of
individually and collectively, with emphasis on
professional societies, contributing their expertise
to the process of governmental decision making. Each
of the panelist presented examples of significant
efforts to provide objective technical advice and
considerable discussion ensued among the panelists
and the audience.
As an introduction to the session, several
postulates were presented for consideration by the
panel and the audience:
•Governmental decision makers, particularly
legislators, and particularly at state and local
levels, often are unable to obtain objective technical
advice and have difficulty judging the objectivity and
quality of technical presentations.
•Professional societies, because of their
broad coverage of subject matter, geographic spread,
and representation from the public and private sector,
can be excellent sources for locating and providing
objective technical material.
•Further they should provide such material,
even if it means foregoing some other of their
voluntary activities - meetings and publications
especially - in return for the largesse they enjoy
from the general public in the form of tax relief and
financing of members meeting and publication activities.
•Scientists and engineers are often as
ignorant of governmental processes and opportunities
for meaningful contributions to decision making
having technical ramifications as legislators are of
science and engineering. Appropriate merging of
governance and technology, therefore, requires a
willingness on each side to learn the processes and
some of the knowledge of the other.
1
11-1
-------
PUBLIC SCIENCE ADVISING
Joel Primack
Physics Department
University of California, Santa Cruz
Santa Cruz, California 95064
Until recently, the interaction of the American
scientific community with government on matters of
technology was almost exclusively confined to service
by scientists on confidential advisory committees to the
executive branch of the federal government. This
relationship began to change in the late 1960's as many
scientists—some of them former government advisors-
took their concerns about an increasing variety of issues
to the public. The advisability of government financing
of the U. S. supersonic transport project, the strategically
destabilizing impact of antiballislic missile systems, the
ecological and health impacts of persistent pesticides ,
and the safety of nuclear reactors are some of the techno-
logical issues which have become the focus of such public
debates in the past several years. 1
Of course, there had earlier been significant public
debates over the creation of a non-military Atomic Energy
Commission and the desirability of a nuclear test ban treaty.
But these had been isolated fights, limited in duration and
carried on by a relatively small segment of the scientific
community. The contemporary "public interest science"
movement has had much broader participation by scientists
of many disciplines, and it is even beginning to show signs
of permanence through the institutionalization of a variety
of public interest science activities and through the increas-
ing number of scientists who have made career commitments
in this area,
One aspect of the new mood among scientists—as
among other citizens—has been a decreasing willingness
to leave technological issues to be decided by the federal
executive branch. This is partly a reaction to the excesses
of the Nixon White House, no doubt, but it also appears to
reflect a new awareness of the responsibilities and possibil-
ities of the judicial and legislative branches of government
at both the national and state level. For example, the
Environmental Defense Fund, in their long fight to ban DDT
and other persistent pesticides, began with local campaigns
and court fights in New York, Wisconsin, Michigan, and
other states before taking the issue to the Agriculture
Department, the Environmental Protection Agency, and—
most decisively—the federal judiciary.
The new Congressional Scientist-Fellow program of
the AAAS, the American Physical Society, and other pro-
fessional societies, demonstrates the opportunities of the
new focus on Congress. The first six Congressional
Scientist-Fellows served in 1972-4, and this fall the third
group will begin their year of serving on Congressional
staffs. It is a measure of the weakness of Congressional
staff capabilities in science and technology, that the
Congressional Scientist-Fellow program has more than
doubled the number of Ph .D. 's working for the Congress.
Their impact has been considerable. For example, the
recent redirection of the Energy Research and Develop-
ment Administration toward greater emphasis on research
and development for increased energy conservation rather
than increased energy production was partly inspired by
a detailed critique of ERDA's first budget, prepared for
the House Science and Technology Committee by Jon Veigel,
an AAAS Congressional Fellow at the Office of Technology
Assessment.
Besides the Congressional Scientist-Fellow program,
additional examples of the institutionalization of public
service activities in my own professional society—the
American Physical Society—include the creation of the
Forum on Physics and Society and a series of energy
studies including a major report on nuclear reactor safety.
Other scientific societies have undertaken similar programs.
These public science advisory activities are clearly serving
a useful function, as are the legal and political advocacy
activities of such public interest science groups as the
Environmental Defense Fund. I expect that these programs
will continue to provide a valuable counterbalance to the
more formal science advisory apparatus of the federal
executive branch, which is presently under repair.
Democratic decision-making on increasingly complex
technological issues demands such a plurality of inputs.
1. Scientist involvement in these and other technological
issues of recent years is explored in some detail in a
recent book by Frank von Hippel and myself, Advice
and Dissent: Scientists in the Political Arena (New
York: Basic Books, 1974) .
11-2
-------
A SCIENTIFIC SOCIETY'S APPROACH TO SCIENCE AND GOVERNMENT
Dr. Stephen T. Qulgley
Director, Department of Chemistry and Public Affairs
American Chemical Society
The approach to facilitating the interaction
of science and technology with the governmental
process involves two parallel mechanisms of implementa-
tion. Scientists and engineers have historically been
involved as individuals indirectly through the advisory
process whereby individual scientists and engineers
have been invited to provide government officials,
executive agencies and Congressional committees with
their personal views on technical issues based upon
their own research and experience.
The American Chemical Society, which is a non-
profit scientific and educational society of more than
110,000 member chemists and chemical engineers, oper-
ates under a National Charter granted by the 75th
Congress in 1937. This charter, in addition to list-
ing the broad scientific and educational objects of
the Society, also imposes a responsibility on the
Society to provide assistance to the federal govern-
ment in the solution of public problems which are
related to the areas of its scientific and technical
competence.
In the middle 1960's the Board of Directors of
the American Chemical Society established the Commit-
tee on Chemistry and Public Affairs to act as a focal
point in attempting to institutionalize the Society's
approach to more effectively fulfilling its obligation
under its National Charter. From its beginning the
Committee has planned the public affairs program with
great care and deliberation. It has been particularly
sensitive to the need for the involvement of a broad
spectrum of Society members' views on the different
aspects of the program and in its approach to the
investigation and evaluation of public issues and
policies. The Committee has long stressed the import-
ance of maintaining the Society's credibility as an
intellectual resource for the country which is as free
of personal and institutional bias as possible.
In fulfilling its responsibilities the Committee
on Chemistry and Public Affairs has concentrated the
Society's efforts in certain specific areas for study
and consideration. In the development of the Society's
public affairs program as an effective communication
link to the public, the government and the chemical
community to help meet our major national needs and
goals, the Committee has been providing leadership
and coordination in the following governmental and
other public affairs activities.
* Study projects on public Issues ox problems
—the results of which may provide informa-
tion and influence in the development of
national policies by delineating projections
of future areas of research and development,
along with future scientific manpower require-
ments and opportunities.
* Advisory and supportive programs for other
ACS units—developing local section and
divisional public affairs, governmental,
legislative, and related programs.
The Committee on Chemistry and Public Affairs
is continuously evaluating and examining other major
public concerns and problems for consideration and
action by the American Chemical Society.
* Technical or socio-economic-education action
programs.
* Attitude or policy formulation and determina-
tion for the Society and its members.
1
11-3
-------
TECHNOLOGY AND THE LEGISLATURE:
A PROBLEM IN COMMUNICATIONS
Margaret M. McNamara
New York State Assembly Scientific Staff
Albany, New York 12224
Technology and legislation is a fairly new field.
With the exception of a few attempts at organization
several years ago, which did not materialize, the first
full-time state legislative science advisory staff in
the country is the Assembly Scientific Staff of the
New York State Legislature. It is now about five years
old. The Staff is small, consisting of three profes-
sional scientists, a research and administrative assis-
tant, and two secretaries. About a dozen state legis-
latures now have some formal mechanism for obtaining
scientific advice, but the total number of full-time
people directly involved is only about a dozen or so,
so we all have a long way to go.
The Scientific Staff is funded by the Assembly
and functions in conjunction with other Assembly staffs
such as Program, Legal, and Ways and Means. Its pri-
mary activity is to anticipate and respond to inquiries
from the Members for technological information and
perspective. It works closely with the Speaker's office
and Standing Committee chairmen. It also has received
grants from the National Science Foundation enabling
it to work with professional societies and universities
to bring additional capability to the legislative
process.
The Assembly, through the Scientific Staff, has
recently taken the initiative in joining with a west-
coast restaurant chain and their equipment supplier
and the State University of New York under funding
from the U.S. Energy Research and Development Adminis-
tration, in an energy conservation demonstration pro-
ject on a restaurant now being built in Albany. It is
a complex engineering project, but briefly it will con-
sist of installing a variety of energy conservation
and reclamation equipment, together with a complete
instrumentation system (temperatures, air, water and
heat flows, etc.). There will be heat exchangers,
water storage, a heat pump, solar collectors, and other
equipment. The project will show how much energy can
be conserved at what price so that return-on-investment
can be determined. All this will be in a regular com-
mercial enterprise using the energy conservation tech-
niques now available. It is expected that the data
output from the project will impact the restaurant
industry nationally, and other commercial industries
as well, therefore not only saving fuel, but increasing
employment.
The significance of such an innovative project
being done by a state legislature, which of course is
interested in new public issues as well as pertinent
legislation relating thereto, is indicative of some
measure of success by an as yet rather small group of
scientists around the country who have been giving some,
attention to decision-makers at the state level. This
small group has an impressive record of accomplishment
in the field of science, technology and public policy
most likely because they have recognized that the real
problem here is communication.
Everyone knows that to achieve communication,
as distinct from merely sending messages, requires
a transmitter, a receiver, ana a feedback device*
Therefore writing reports, giving lectures, and
showing viewgraphs do not meet the criteria. You need
two-way, face-to-face discussion.
Much has been written and said about the com-
plexities of interaction between scientists and public
policymakers, and there seems to be a great deal of
mythology about it. Discussions usually center on the
concept of providing the information as to what the
policy should be technically. But the technical policy
is relevant to those who govern only when it is con-
sidered along with the economic, social, and political
aspects of a given problem.
The difficulty of communication between scien-
tists and decision-makers cannot be overcome easily.
It requires more than casual contacts occasionally.
Credibility must be established and it requires real
effort over a long period of time. Advocacy may be a
good thing, but legislators are surrounded by advocates
most of the time—from industry, constituency, univer-
sities and public interest groups—and they are very
skeptical of all advocates. Legislators have begun
to think in technical terms only recently. What they
need is perspective on potential alternatives.
Scientists have no basis for incorporating their
personal value judgments, whether consciously or sub-
consciously, about what ought to be done, since those
decisions are the prerogatives of policymakers, but
they do have a b.isis for presenting consequences of
alternatives so those policymakers can make informed
and enlightened decisions.
In January 1974 the Assembly and AISLE, An
InterSociety Liaison committee on the Environment, a
consortium of about two dozen professional societies,
held a workshop conference in Albany on Energy and
the Environment. Mr. Bernard Manheimer was the con-
ference chairman. A workshop between legislators and
scientists had been done only once prior to that time,
on a much smaller scale, when members of the Institute
of Electrical and Electronics Engineers had met with
several New York Assembly committees on the same sub-
ject. The Assembly/AISLE Conference results can be
measured in a number of ways. First, over three
dozen energy-related bills were introduced in the
Assembly during the 1974 legislative session, eight
of which eventually became law. Although the particu-
lar bills signed into law were not the most significant,
during the 1975 session more bills were introduced.
One which is now law creates a New York State Energy
Research and Development Authority. I believe these
actions illustrate what professional societies can do
in the way of providing technical advice, since it is
very unusual for a new concept, such as the energy
problem, to result in legislation in such a brief
time after it is brought to the attention of the
decision-makers. A new concept typically takes from
two to five years and more to result in legislative
action. Remember that several hundred people must
agree on even every punctuation mark in a bill before
it can be passed.
The AISLE Conference was different, and that is
why it is possible to measure its impact to some ex-
tent. It was not a series of lectures, but it involved
working groups of scientists and legislators. They
worked as equals, each becoming aware of the other's
expertise and capability, and dedication, in their
respective fields. The Conference established the
1
11-5
-------
channel and the par Lieipants achieved the communicative
act. It is not so surprising that the conference plan
worked well when one considers that the scientists and
engineers went much more than half way to bridge the
cultural gulf between themselves and the legislators.
Once the Members recognized this, they would do no less
in establishing communication.
The Scientific Staff benefits from another very
tangible result of the AISLE Conference. The Members
of the New York State Assembly are far more comfortable
with scientists than they were prior to AISLE. When
the fluorochlorocarbon/ozone issue surfaced early this
year, a bill to restrict the sale of, and eventually
ban, aerosol sprays was introduced. The Assembly
Committee on Environmental Conservation decided to
hold hearings and notices were sent to professional
societies, government agencies and experts in univer-
sities and industry. The response from the scientific
community was excellent. A piece of model legislation
was drafted and passed the Assembly without opposition.
During the time the bill was being considered by the
Senate, considerably more hard evidence on the actual
presence of fluorochlorocarbons in the stratosphere
was developed. However the Senate introduced a com-
promise bill, which passed both houses during the final
days of the 1975 session and was signed into law in
August. It seems quite certain that the Assembly bill
reflected the additional channels of communication open
to Members of that body.
In the final analysis, what is really important
is that communication does begin to take place between
the disciplines of science and politics. Every group
has its own priorities and value systems, but the re-
cognition that there are economic, social, political,
and technical ramifications to any major public-policy
decision is essential. The professional societies par-
ticipating in this conference are notable for their
advance-guard action and it is to be hoped that the
entire scientific community will follow their lead.
2
11-5
-------
HEALTH EFFECTS OF SULFUR DIOXIDE AND
SULFURIC ACID AEROSOLS
K. A. Bustueva
Department of Communal Hygience
Central Institute for Advanced Medical Training
Moscow, USSR
The effect of sulfur dioxide on health has been
the subject of numerous experimental investigations
and observations on occupationally exposed groups, and
on the general population living in polluted areas.
Yet, the peculiar ways in which low concentrations of
sulfur dioxide influence human health are still not
clear. In our opinion, this is due to the peculiari-
ties of the resorptive action of sulfur oxides charac-
terized mainly by its effects on enzymes and metabollic
processes and not leading to any specific disease.
For a long time, sulfur dioxide was considered as
a substance with a purely local action on the respira-
tory tract. In fact, a number of investigations con-
ducted in various countries on test animals and of
observations of human beings indicate that the inhala-
tion of sulfur dioxide causes a local reaction in the
respiratory tract with pathohistological changes in
the pithelium of the trachea, bronchi and elastic lung
tissue. Nasopharyngitis, chronic bronchitis and
changes in the muscosa of the upper respiratory tract
were observed in workers with long term exposure to
sulfur dioxide.
Investigations conducted to elucidate the direct
effect of SO2 on the ciliary activity of the respira-
tory tract mucosa (T. Dalhamn, 1956; Y. Rhodin, T.
Dalhamn, 1956), confirmed the local action of sulfur
dioxide. However, the investigations carried out dur-
ing the past two decades revealed that besides local
action sulfur diojcide has also a general resorptive
action.
The merit for discovering the resorptive action
mechanism of sulfur dioxide goes to I.V. Sidorenkov
(1955, 1957) and his colleagues who performed a series
of experiemental Investigations Involving different
species of animals. They verified their conclusions
under real conditions by studies on workers exposed to
this industrial hazard.
In Sidorenkov's opinion, the resorptive action of
sulfur dioxide is based on the inactivation of enzyme
systems regulating the course of various biochemical
reactions. The following points are significant: the
interaction of sulfur dioxide with thiamine and enzymes
containing thiamine, with dithiol groups in proteins
and protein enzymes, as well as with aldehydes and
ketones appearing as a result of metabolism. During
the period when sulfur dioxide was inhaled it was pos-
sible to observe a depressed respiration of tissues,
inhibition of glycolysis in certain tissues, accumula-
tion of pyruvic acid in the blood, hyperglycemia and
intensification of gylcogenolysis.
Under the acute and chronic action of SO2 there is
a compensatory reaction resulting in an increase in the
general protein content, an intensified oxidizing
deamination of amino-acids, simultaneously with a less
intensive oxidization of pyruvic acid.
In later investigations the findings of Sidorenkov
and his associates were confirmed. At the same time,
studies pursued in the 1960's supplemented our under-
standing of the effect of sulfur dioxide on immunobio-
loglcal activity indices,
Navrocij (1955), in a chronic experiment involving
rabbits, studied the effect of sulfur dioxide on agglu-
tination and the blood complement titer after triple
immunization with typhoid vaccine. There -were three
investigations: immunization only; preliminary expo-
sure to sulfur dioxide for a month followed by immuni-
zation; and, finally, simultaneous exposure and immuni-
zation. The author thus concluded that the chronic
action of sulfur dioxide sharply despressea the forma-
tion of agglutinins in the test animals preceeded by
a shorter period in which their titers are high, while
the blood complement reaction, being an evolutionary
function, was stable and the changes in its indices
were neither well-defined nor typical.
Without denying Sidorenkov's opinion that the site
of sulfur dioxide action is the depression of enzyme
processes, Navrocij attached a particular role to the
powerful irritant properties of sulfur dioxide. He
considered that sulfur dioxide circulating in the blood
can cause intensive irritation of the Interoceptors, as
a result of which there is a reflectory impairment of
enzymes, general metabolic processes and a sharp depres-
sion of immunobiological activity. This point of view
was confirmed by the results of the experiments per-
formed by Litkins and Saknyn (1955), who studied the
sulfur dioxide content (probably sulfites) in the blood
of rabbits that were exposed to sulfur dioxide at dif-
ferent concentrations. The amount of sulfur dioxide
content in the blood was found to be directly propor-
tional to the concentration and duration of exposure.
For Instance, when the sulfur dioxide concentration
was 50-60 mg/m3, the content was 4.8 rog/100ml; at
80-120 mg/m-' the content in the blood was as high as
8 mg/100ml. Sulfur dioxide was discovered in the
blood already after the first 2-3 minutes of exposure,
although its content continued to grow for one hour.
The suppression of immunobiological activity was
demonstrated in a number of other investigations. For
example, Erba and Korzhinen (1960) pointed to the reduc-
ed formation of antibodies in rabbits under chronic ex-
posure to sulfur dioxide.
The results of investigations on the effect of
sulfur dioxide on occupationally exposed persons agree
with the above-mentioned mechanism of action.
For instance, Sterehova (1955), when examining
workers exposed to sulfur dioxide, obtained data indi-
cating a change in the carbohydrate and pigment function
of the liver, as well as changes in the state of blood
colloids. She emphasized that the workers complained,
first of all, of pains in the epigastric area and the
right hypochondrium; coughing and difficulty in breath-
ing were secondary complaints.
Litkins, Saknyn and Sterehova (1957) indicated
that the morbidity with temporary disability, due to
gastrointestinal diseases and pneumonia, was much higher
among workers in copper smelters where they were expo-
sed to sulfur dioxide. The authors also investigated
the sulfur dioxide content in the blood of workers,
which fluctuated within 0.025-0.063 mg/lOOml. (It
would be more correct to consider this the content of
sulfites).
1
12-1
-------
Kushman and Sidorenkov (1955) pointed out the ac-
cumulation of pyruvic acid in the blood of workers ex-
posed to sulfur dioxide.
All these investigations under industrial condi-
tions have definite and well-known drawbacks: as a
rule, there is a combined action of chemical and phy-
sical factors; the range of fluctuations in the concen-
tration of sulfur dioxide during the work day is too
wide to relate the results of the observations to defi-
nite concentrations, and so on. Nevertheless, the re-
sults are of definite, interest as regards typical
features of the effect of sulfur dioxide and seem to
confirm the hypothesis about the mechanism of its
action.
The evaluation of the effects of sulfur dioxide
on the population at large is even more difficult. As
was pointed out earlier, there have been a great num-
ber of investigations on the effect of sulfur dioxide
on the general population. Of particular interest in
this respect are the investigations on children, which
to a certain extent, confirm the above-mentioned mecha-
nism of local and resorptive action of sulfur dioxide.
Not a single investigation discovered an effect that
was specific only for sulfur dioxide. All changes that
were observed in the health of children can be reaated
to a reduction in the immunobiological activity and to
the influence on enzyme and metabollic processes: higher
incidence of diseases of the respiratory tract, lower
indices of respiratory functions and physical develop-
ment, changes in the activity of certain blood enzymes,
retarded skeleton development and so on.
One must keep in mind yet another aspect of the
effect of sulfur dioxide on the organism. At present,
we know a number of substances with co-carcinogenic
properties, i.e. the ability to intensify the develop-
ment of the tumoral process in tissues, that were pre-
liminarily treated with carcinogen. There were investi-
gations conducted in recent years indicating that the
action of sulfur dioxide, in combination with carcino-
genic substances, stimulated the process of carcino-
gensis in the lungs.
In experimental conditions, the potentiating car-
cinogenic action of sulfur dioxide treated with benz(a)-
pyrene, was demonstrated by the American researchers (M.
Kushner, L. Laskin, 1968, 1971) and Soviet scientists
(Skvortsova et al., 1972),
These experimental investigations correlate very
well with observations carried out recently, The inves-
tigation of Irodova, finished in 1974, illustrates that
there is a dependency between lung cancer morbidity and
mortality of the population and the level and nature of
air pollution: lung cancer morbidity and mortality of
the population was twice as high in cities with a signi-
ficant pollution of the air by complex toxic (the main
one being sulfur dioxide) and carcinogenic substances.
This aspect of the problem, concerning control of
atmospheric sulfur dioxide, is very important since it
is quite possible that it may play a significant role
in the prevention of malignant tumors.
The experimental investigations involving both
test animals and volunteers (amdur et al., 1952, 1958;
Bustueva, 1957, 1966; Pathle et al., 1956; Nevskaja,
Kocetkova, 1961; Bustueva, 1964, and others) make it
possible to regard sulfuric acid aerosol as a typical
Irritant, the respiratory system being the target. It
is well-known that the indices for the threshold for
irritant resorptive effects of irritants are quite
close to each other. The irritating action threshold
for sulfuric acid mist is 0.5-0.6 mg/m for sensitive
individuals (Bustueva, 1957; Dorsch, 1913). At concen-
trations of about 0.4 mg/m in certain persons it is
possible to observe changes in breathing (Amdur et al.,
1952) and in electroencephalograms (Bustueva, 1960).
Lower concentrations do not induce any changes in
volunteers.
Under ordinary community conditions, where the
source of sulfuric acid aerosol is the oxidation of
sulfur dioxide, one hardly ever encounters concentra-
tions of the order of tenths of mg/m^. In view of this,
it is difficult to presume there is an independent
effect of sulfuric acid mist on the population. How-
ever, this may not be the case when there are unfavor-
able meteorological conditions, promoting the accumula-
tion of pollutants and the oxidation of SO2 to SO3, as
was the case in some acute pollution episodes.
For higher concentrations of sulfuric acid mist
(1.0 mg/m^ and higher), one may expect serious conse-
quences to human health both through the direct irri-
tating action of the mist, which may be accompanied by
a reflectory spasm of the bronchi and of the rima glot-
tisis, as well as by secondary spasmatic reactions in
the damaged connective lung tissue, which leads to the
release of histamine (Bustueva, 1966).
It is noteworthy that a critical organ for sulfuric
acid mist is the lung, resulting in changes such as
the proliferative intermediate pneumonia.
No doubt, there is a need for further investiga-
tions in order to elucidate the nature of the resorptive
action of low concentrations of sulfuric acid mist,
especially in combination with sulfur dioxide and sus-
pended particulates.
The effect of sulfuric acid aerosol on man's health
is an even more complicated problem. Sulfuric acid mist
is an air pollutant just as widespread as is sulfur di-
oxide, since it always accompanies the latter, although
it is only a small percentage of the SO2 concentration
(3-10%). This is the main reason why it is practically
impossible to study the effect of this compound on the
general population. A limited number of studies have
been performed to investigate the isolated influence of
sulfuric acid mist under occupational conditions.
2 12-1
-------
Bibliography
Dalhamn, T.
Muccous flow and ciliary activity in the trachea
of healthy rats and rats exposed to respiratory
irritant gases (SO2, NH , HCHO)
Stockholm, 1956, p. 161
Rhodin, Y. and Dalhamn, X.
Brit. J. Industr. Med., 1956, 13, p. 110
Sidorenkov, I. V. In Russian
1. The action of sulfur dioxide on brain respira-
tion
2. The role of vitamin intoxication by sulfur
dioxide
Tr. Ckalovskogo Gos. Med. Inta [Transactions of
the Ckalovskij State Medical Institute], 1955,
No. 4, pp. 25-82; 46-52
Sidorenkov, X. V. in Russian
On the mechanism of the' action of sulfur dioxide
on the organism Vestnik Ckalovskogo otd. Vsesojuz-
nogo Himiceskogo Gbscestva im. Mendeleev [Bulletin
of the Ckalovskij Branch of the All-Union Mende-
leev Chemical Society], 1957, No. 7, pp. 65-67.
Sidorenkov, X. V. , Slobodina, K. V.
Susarova, A. M. and Kunsman, M. I.
Tepljakov, N. 1.,
in Russian
On the mechanism of the action of sulfur dioxide
on the organism (In: The Proceedings and Papers
of the Volga Conference of Physiologists, Bio-
chemists and Pharmacologists,"Kuibysev, 1957,
p. 225)
Navrockij, V. K, in Russian
The influence of small concentrations of sulfur
dioxide in the presence of chronic poisoning, on
the immunobiological reactivity of rabbit.
J. Gigiena i Sanltari.la, 1959, 8, pp. 21-26
Amdur, M., Schuler, R. and Drinker, P.
Toxicity of sulfuric acid mist to guinea pigs
Arch. Industr. Hya- Occup. Med., 1952, _5,
pp. 318-329
Amdur, M.
The respiratory response of guinea pigs to sulfuric
acid mist
Arch. Induatr. Hlth, 1958, 18, pp. 407-414
Bustueva, K. A. in Russian
On the toxicology of sulfuric acid aerosol
J. Glaliena i Sanitaria a. 1957, 2, pp. 17-22
Bustueva, K. A.
in Russian
The toxicity of sulfur dioxide under conditions of
continuous exposure
(In: The biological effect and hygienic signifi-
cance of atmospheric pollutants, 1966, pp. 142-172)
Pattie, R, and Cullumbine, H.
Toxicity of some atmospheric pollutants
Brit. Med. J., 1956 (20 October), pp. 913-916
Nevskaja, A. I. and Kocetkova in Russian
On che combined effects of ozone and sulfuric acid
aerosol
2. Gigiena Truda i. Prof. Zabolevannij [J. Labour
Hygiene and Occupational Diseases] 1961,22, pp.20
pp. 20-29
Dorsch, R.
Uber die Verunreinigung der Luft durch Schwefel-
saure in Akkumulateurraumen und derer Umgebung
Dissertation, Wurzburg 1913
Bustueva, L. A. in Russian
The threshold of the reflective action of a com-
bination of sulfur dioxide and sulfuric acid
(From: Maximum concentrations of air pollutants,
1960, 4_, pp. 92-102
Litkens, V. A. and Saknyn, A. V. in Russian
On. the hygienic evaluation of the overall toxic
effect of sulfur dioxide
(In: Questions of occupational health, prophy-
laxis and toxicology in factories of the Sverd-
lovsk region. Sverdlovsk, 1955, pp. 160-172)
Litkens, V. A. , Saknyn, A. V. and Sterehova, N. P,
in Russian
The chronic toxicological effect of sulfur dioxide
Proceedings of the Jubilee Scientific Meeting,
Leningrad, 1957, pp. 382-388
Erba, L. and Korjinek
Prispevek ke studiu pusobeni SO2 na organismus
Ceskovslovenska Kygiena, 1960, 2-3, pp. 121-127
Sterehova, N. P. in Russian
The general toxic effects of sulfur dioxide, as
indicated by clinical and experimental investi-
gations
(In: Questions of occupational health, prophy-
laxis and toxicology in factories of the Sverd-
lovsk region. Sverdlovsk, 1955, pp. 173-178)
Kunsman, M. I. and Sidorenkov, I, V. in Russian
The content of pyruvic acid in the blood of o
workers in an atmosphere containing sulfur dioxides
Tr. Ckalovskogo Gos. Med. Inta [Transactions of the
Ckalovskij State Medical Institute], 1955, No. 4,
pp. 59-64
Irodova, E. V. in Russian
On the influence of industrial emissions, contain-
ing carcinogenic and toxic substances, on the in-
cidence of lung cancer.
J. Gigiena i Sanitarija, 1974, T_, pp. 6-9
3
12-1
-------
THE INDOOR ENVIRONMENT - A NEW FRONTIER?
John E. Yocom
TRC - The Research Corporation of New England
Wethersfield, Connecticut 06109
INTRODUCTION
The primary thrust in defining air quality and
its relationship to human health effects has been
directed at the outdoor ambient atmosphere. In fact,
air pollution is defined in many air pollution control
laws as "... the presence in the outdoor atmosphere
of one or more contaminants, or combination thereof,
in such quantities and of such duration as may be or
tend to be injurious to human, plant or animal life
or property...". This empahsis on the outdoor atmos-
phere is surprising in view of the large percentage of
time which the average person in this country spends
indoors where the quality of air can differ consider-
ably from that outdoors. There are no accurate esti-
mates of the percentage of time which people spend
indoors. Depending on where they live, their occu-
pation, mobility and other factors, this percentage
will vary widely. Forest rangers and construction
workers may spend 1/2 to 2/3 of their time indoors
depending on the season, while city dwellers, the old
and infirm, the very young, and young mothers may
spend almost 100 percent of their time indoors, espe-
cially in the winter.
Therefore, it is clear that indoor air quality in
terms of the duration of exposure is far more impor-
tant than outdoor air quality in determing health and
welfare. The indoor exposures to air pollutants are
made up of those discrete periods (and associated con-
centrations) spent by individuals in the workroom,
home, transportation vehicles, public buildings,
stores, schools, restaurants and barrooms, theaters,
hotels and other enclosed inhabited spaces. Aside from
the work room atmosphere whose air quality is
governed by standards in the form of Threshold Limit
Values (TLV's), the indoor environment is generally
a "no-man's-land" about which there is only limited
knowledge.
This paper summarizes some of the available data
and information on indoor air quality in enclosed
spaces. The predominant data base is in indoor air
quality in homes and office buildings.
FACTORS INFLUENCING INDOOR AIR QUALITY
The quality of the indoor atmosphere varies
widely and can be influenced by a variety of factors
which are discussed generally below.
1. Outdoor Air Pollution
Outdoor ambient air surrounding a building pene-
trates indoor spaces through natural or controlled
ventilation, leakage, and diffusion. The type of
pollutant will influence the character and degree of
penetration, e.g., particulates are in part removed
in the penetration process, while CO, an unreactive
gas, readily penetrates.
2. Indoor Generation of Pollutants
Indoor pollutants are generated by cooking, clean-
ing, smoking, painting, cosmetic application, and
simply moving around and stirring up particulate mat-
ter. Pollutants produced inside a building diffuse
throughout the closed space and build up to equilibrium
concentrations based on leakage out of space, leakage
into the space from the same pollutant present in the
outdoor atmosphere, and pollutant removal processes
such as adsorption and absorption on interior surface,
and atmospheric reactions.
3. Building, Permeability
The ability of outdoor pollutants to leak in and
indoor pollutants to leak out as a result of building
permeability has a strong influence on indoor air
quality. The principal factors are numbers of open
windows and doors and general tightness of building
construction.
4. Ventilation and Air Conditioning Systems
A positive ventilation system in a building can
have a profound effect upon indoor air quality. One
which provides outdoor ventilation will produce inte-
rior concentrations approaching those outdoors, while
one which recirculates indoor air will produce indoor
air quality dominated by pollutants generated indoors.
An air conditioning system which includes filtration,
air washing, humidifIcation or de-humidification, and
activated carbon adsorption can be expected to have an
effect on indoor concentrations of particulates and
reactive gaseous pollutants (e.g., SO2).
5. Meteorologic and Geographic Factors
Weather conditions help determine indoor air
quality in a number of ways.
a. Local meteorology can determine outdoor air
quality, e.g., temperature inversions can
cause outdoor concentrations to become ele-
vated. Under these conditions, outdoor air
penetrating the building will degrade indoor
air quality.
b. Indoor-outdoor temperature relationships can
determine tendency for penetration of outdoor
pollutants. In cold weather, a heated build-
ing exhibits a considerable stack effect which
produces significant pressure differences be-
tween inside and outside tending to draw in
outdoor air at the lower levels and exhausting
indoor air near the top of the structure.
c. Wind can also produce pressure differences
between inside and outside and enhance the
tendency for outdoor pollutants to penetrate
or indoor pollutants to dilute more rapidly.
6. Locations with Respect to Outdoor Sources
The location of a building will determine the
general ambient outdoor concentrations, but beyond
this, nearby sources whose impact on the generalized
ambient air may not be well defined, can strongly
influence the quality of the building's indoor air.
Nearby stacks, recirculation of combustion products
from the building's heating system, and a truck loading
dock near the building's air intake are examples of
this factor.
7. Energy Conservation Measures
A number of energy conservation measures can influ-
ence indoor air quality.
a. Heavy insulation and tight construction of
homes can prevent penetration of outdoor
pollutants, but will permit build up pollu-
tants generated indoors plus CO2 and moisture
from human occupancy.
1
12-2
-------
b. Maximum recirculation of heated or conditioned
air in winter and summer will reduce impact of
outdoor air quality but will maximize the in-
fluence of indoor pollutants.
c. Use of outdoor ventilation when outdoor tem-
peratures and humidities permit substitution
for recirculated heated or conditioned air,
will produce indoor air quality approaching
that of outdoors.
It is clear from the above listing that indoor air
quality is a complex function of many factors. Most
of what is known about indoor air quality is in terms
of indoor/outdoor relationships and the indoor genera-
tion of pollutants. This paper is directed principal-
ly at these two factors.
INDOOR/OUTDOOR AIR QAULITY RELATIONSHIPS
The indoor environment is basically an extension
of the outdoor environment. Depending on the type of
structure, outdoor temperature and humidity and type
of building and its ventilation system, the indoor
environment may be similar to or quite different from
the outdoor environment. Nevertheless, except for
enclosed spaces such as submarines and space vehicles,
indoor air quality depends on outdoor air quality and
a review of what is known about indoor/outdoor air
quality relationships provides some insight into the
nature of indoor air quality. EPA reviewed the liter-
ature on indoor/outdoor air quality relationships.1
1. Gaseous Pollutants
Since gaseous pollutants are mixed with the out-
door atmosphere, they penetrate the interiors of
buildings as easily as air. Except to the extent
that the pollutant gases are reactive and disappear
through various removal processes or are produced
inside, the concentrations indoors will depend upon
the degree of penetration of outdoor air.
The results showed that during the winter the stack
effect of the building strongly influenced indoor
air quality. Contaminated air from the streets and
roadways was brought into the building at the lower
levels and distributed throughout the building.
Since outdoor concentrations fall off exponentially
with height above the roadway, indoor concentrations
during the winter exceeded those outdoors at the upper
floors of the building. During the summer, wind
direction and speed were the principal factors influ-
encing indoor CO levels.
Legand
O Day i City Library
• Ntflht 1 (non-ftir conditioned)
| 4000
B
S
~ Day l Office building
• Nfflhl J (air conditioned)
& Day 1 Private home
4 Niflht I (non-air conditioned)
\
m—
I Window
u?
Far Near Near Far
Outiid* outtide tntirie imidi
Summer
Window
-L-
-I—
Far Near Neat Fbt
outside outsido inside inside
Winter
fij. i. Jndoor/ourdoor profile* for c.irbon monoxide in
tevcral type* of building* • summer and winter,
{ref. 2)
Carbon monoxide, because it is an extremely' unreac-
tlve gas, can be expected to penetrate the indoor
atmosphere without loss except through dilution.
In a comprehensive study of indoor/outdoor air
quality relationships for 6 buildings of various
types in Hartford, Connecticut, indoor/outdoor ratios
for CO were found to be about 1.0 or slightly above.3
Figure 1 shows indoor/outdoor profiles for three of
the buildings studied both summer and winter. The
spacing of sampling points in the diagram is arbitrary.
The "far outdoor" sample was taken generally about 20
feet from the building, the "near outside" and "near
inside" sampling stations were taken on each side of
a window in the building and the far inside sample was
taken in an interior room of the building. These data
show that indoor levels of CO are slightly higher than
those immediately outside, probably as a result of in-
door sources such as smoking. The appreciably higher
levels of CO at the "far outdoor" station at the City
Library is the result of this being an air-rights
structure, built over a four-lane freeway connector.
Giles, et al, measured indoor,/outdoor concentra-
tions of CO at two types of high rise buildings in
New York City."4 One was a modern apartment house
built over a freeway with central heating and option-
al window-mounted air condtioners. The other build-
ing was an older "canyon structure" with central
heating and minimum air conditioning by a few window-
mounted units. CO concentrations were measured inside
and outside at a number of elevations above the street.
Sulfur dioxide is a reactive gas which is sus-
ceptible to oxidation, and in the presence of moisture,
can be absorbed or adsorbed on surfaces and react with
them. Therefore, it is not surprising to find that
indoor/outdoor ratios for S02 are less than 1.0 with
considerable scatter in the data. Figure 2 shows
results of several studies5>6>7'8 which Indicate that
typically indoor SO2. levels are about half those out-
doors9'10. But when outdoor concentrations were above
130 yg/m3, indoor concentrations remained near their
normal levels, tn this same study, indoor/outdoor
relationships were determined for a building which
housed both offices and laboratories. During one
period while this building was being sampled, indoor
concentrations exceeded those outdoors.
/
/.
• Hotp ui, CincinnMi, U.S.A. (ill, 6)
• Houtnf, Mokow, USSR |r«l. 6)
* Hout h n«w VltcoM PUM, USSR (r*l. 7)
* Koum. Hottirtfam |«f. I)
laoo !w5 iw
Outfeor
fit. 2.
-ite—far
2
12-2
-------
Particulate Matter
INDOOR GENERATION OF POLLUTANTS
Particulate matter behaves quite differently than
gaseous materials. One would expect that some parti-
cles (especially the larger ones) would deposit out
in the process of penetrating buildings. The data
bear out this contention.
Figures 3 and 4 show indoor/outdoor profiles for
total suspended and soiling particulate matter, res-
pectively. Note in Figure 3 the relatively lower
outdoor concentrations of total suspended particulate
during the summer as compared with winter. Also,
note the steeper gradient from indoors to outdoors
during winter as opposed to summer. This is a func-
tion of both higher outdoor concentrations and the
more tightly closed mode of the buildings (except for
the air conditioned building) during the winter.
Initial investigations of indoor air quality con-
centrated on the identification of the penetration
pattern of outdoor air pollutants. Only recently has
some effort been devoted to quantitatively assessing
the impact of indoor generated pollutants as distinct
from the penetration of the same pollutants from out-
doors. Some of the most important activities that
result in the indoor generation of pollutants are:
Smoking
Use of Personal Products
Cleaning
Cooking
Maintenance
Hobbies
Electrical Appliances
Lagrf
O Day \ C'ty Ubniv
• Night 1 (non-air conditioned)
~ Day 10ffice building
¦ Night j (air conditioned}
A Day j Private home
Jk Night I (non-sir conditioned)
i Window
Far Near Near Far
ouUide oottide inside inside
Summer
Far Near N«ar Far
outside outside inside inside
Winter
Fig. 3. Indoor/outdoor protilei tor total impended particulate matter in
several types of building • summer and winter, (ref. 2}
Window |
\!
O Day ¦> City Library
* Night J (non-air conditioned) j
~ Day ) Office building i Window
¦ Night 1 (air conditioned)
6 Day ) Private home
A Night I (non-air conditioned) I
J
-J-
-0-
Far Nut Near Far
outside outside Inside inside
Summer
_2_
t
J-
Fat New New Per
outside outside Inside Inside
Winter
Fig. 4. Indoor/outdoor profiles fur soiling particulate matter in several type*
of buildings - summer and winter, (ref. 2}
This section reviews the rather limited work on the
effect of indoor generation of pollutants on indoor
air quality.
1. Gaseous Pollutants
Carbon Monoxide - Carbon monoxide is generated by
smoking, heating, cooking and by automobiles (attached
garage). Gas heating systems rarely affect indoor
CO concentrations, but gas stoves and attached garages
can have a significant effect on indoor CO levels.
The effects of stoves and garages on indoor CO concen-
trations is shown in Figure 5, which shows the CO con-
centrations in a house in Hartford having a gas range
and an attached garage. The family room is between
the kitchen and garage. For this house, CO concentra-
tions are generally much higher than and unrelated to
outdoor levels. Peak concentrations in the kitchen
correspond to the periods when meals are being cooked,
and concentrations in the family room generally follow
those in the kitchen rather than those outside. For
two periods in the record, when the car was being put
into or taken out of the attached garage, the emis-
sions from the garage are the controlling influence
on both the family room and kitchen concentrations.'
CM IttNQ tAKEN fftOM GARAGE
CM WING HUT IN GARAGE
KITCHEN
«... FAMILY DOOM
—OUtSIDE
UARI
Mil. travi*
Figure 5s Carbon Monoxide ConeentTatione in House with
Cas Range end Furnace and with Attached Garage'1
The data in Figure 4 show similar profile for
soiling particulates but not nearly as pronounced as
for particulate weight concentrations. In general,
the indoor/outdoor ratios for soiling particulates
were higher than for total suspended particulate
matter. This finding is logical since soiling effects
of particulate matter are created largely by small
particles which penetrate more readily than large
particles. Furthermore, sampling simultaneously
indoors and outdoors with cascade impactors showed
that indoor particles were smaller than those collect-
ed outdoors.
Oxides of Nitrogen - As part of a study to show
the impact of gas stoves and other indoor sources on
indoor air quality, extensive indoor and outdoor air
quality measurements were made for nitrogen oxides.
(N02 and NO) using several residential structures. 1
Figure 6 is a graph of two-hour average NO2 levels
for a typical two-day period at one of the structures
used in the study. This graph shows the rapid re-
sponse of N02 levels to stove use, followed quickly by
the response at the other two indoor locations. The
oven, in particular, seems to have the greatest influ-
ence on the indoor NO2 concentrations. Examination
3
12-2
-------
of five-minute averaged data showed that the genera-
tion of N02 by the oven is greatest during the initial
oven start-up. Average NO2 concentrations in the
kitchen during inactive periods were slightly greater
than N02 levels at the other two indoor locations which
closely approximated the outdoor NO2 concentrations.
The higher kitchen concentrations are undoubtedly
caused by the stove's pilot lights.
I
I Burner on . . Z Burner* on
to »1n. —-"j p— 30 inlii.
3001* Oven OnJ ^
) Burner on —J U- ., 1 Burner on
ill 1 45
1 (JIt,,,,,,
c I I **- * mi*
Kitchen
— — — Uvlng room
-1 1,1-11 Bedroom (upper level)
— — 0«wi
As part of this same study, a diffusion experiment
was carried out in a home unoccupied at the time. All
stove burners were operated until one or more of the
sensing instruments read maximum values, and then the
stove was turned off and NO2, NO and CO were allowed
to diffuse through the rest of the house. CO, in un-
reactive gas, showed a half-life of 2.1 hours. This
dilution rate is considered to be the natural venti-
lation air change rate for this house under the test
conditions. NO showed a half life of 1.8 hours indi-
cating little loss through processes other than simple
dilution. On the other hand, NOz had a half life of
only 0.6 hour indicating a considerable amount of loss
by processes such as adsorption, absorption and atmos-
pheric reaction.1J Figure 7, based" on this experi-
ment, shows concentration-time relationships for both
N02 and NO at three locations in the house in compar-
ison to outside concentrations.
5 Burners on
5 Burners off
Burner min,
• • Kitchen
— — Living room
bedroom (upper level)
Outside of structure
0600 0800 1000
120U 1400 1600
Time, [hrs]
1800 2000 2200 2400
Figure 7b: oilutlon and Loss with Time for NO
diffusing through a House (Ref. 11}
Sulfur Dioxide - Interior generation of SO2 is
probably limited to faulty heating systems burning oil
or coal.3 Biersteker, et al, reported that indoor S02
concentrations were not generally affected to a signi-
ficant extent by the heating method used. However, in
one 30 year-old home presumed to have a faulty heater
indoor concentrations averaged 3.8 times the outdoor
levels.8 Table I shows a comparison of S02 concentra-
tions for new and old coal heated houses in Hartford.
The exceptionally high indoor concentrations for the
old coal heated house are presumed to be caused by a
faulty heating system. Indoor concentrations at this
house were found to be unrelated to outdoor concentra-
tions; peak values were related instead to the stoking
periods of the furnace. Indoor concentrations at the
new coal heated house were much lower than at the old
house, even though outdoor concentrations were slight-
ly higher at the new house.3
TABLE I
1400
•'Kitchen
— Living room
— Bedroom (upper level)
— Outside of structure
e 1000
600
400,
200 -
0600 0600 1000 1200 1 400 1600 1800 2000 2200 2400
71ine, [hrs]
Figure 7a; Dilution end Loss with time for NO *
diffusing through a House (Ref. 11)
SULFUR DIOXIDE CONCENTRATIONS FOR
TWO COAL-HEATED HOUSES3
Concentration, ppm Indoor/Outdoor
Type of Building Indoor Outdoor Percent
New House 5 14 36
Old House 78 10 780
Little data are available on the indoor concen-
trations of aerosol propellants and the materials dis-
pensed by these propellants. In a study by Cot£, et
all11 propellent emission^ were estimated and indoor
air quality was projected in relation to specific
aerosol product usage. Table II, based on a user
survey, shows emission estimates for propellents (pri-
marily fluorinated hydrocarbons) associated with use
of various household products. These emission fac-
tors, together with use patterns established by the
survey, provided the basis for estimating indoor pro-
pellent concentrations in rooms of various sizes and
at various times after product use.
12-2
-------
TABLE II
EMISSION ESTIMATES OF PROPELLANTS
FOR AEROSOL PRODUCTS
Product Category
Deodorant Spray
Hair Sprays
Shaving Foam
Air Fresheners Spray
Disinfectant Spray
Furniture Polish Spray
Dust Spray
Oven Cleaners
Particulate Matter
Propellant Emission
Estimates (a/use)
1.0
4.9
0.3
5.6
7.5
8.4
4.2
20
1.2
6.5
0.4
11.2
8.4
25
Particulates are generated by Heating, cooking
and smoking. Smoking has been found to significantly
increase particulate concentrations indoors.®'12
According to Lefcoe and Inculet, smoking just one
cigar raised particle counts by a factor of 10 to 100.
Elevated counts persisted for a period of 1 to 3
hours.12 Yocom, et al,2 noted that the higher concen-
tration of organic particles indoors may result in
part from interior generation of pollutants from cook-
ing or smoking, although the greater penetration of
smaller organic particles accompanied with the larger
inorganic particles is also partly responsible for
the difference.
Based on limited measurements for air-conditioned
office buildings in Hartford, it was concluded that
internal generation of suspended and soiling particu-
lates was a significant factor in the estimation of
interior concentrations. For these buildings, the
ratio of internal generation to exterior concentration
was estimated to range from 0 to 0.6 for indoor/out-
door ratios of 30 to 116 percent as determined by a
simple model relating indoor and outdoor concentrations
and indoor generation rates.13 Internal generation may
also contribute to the varying indoor/outdoor ratios
and to the indoor/outdoor ratios greater than 1.0
measured in some instances in the Hartford study.2
CONCLUSIONS
Until recently the indoor environment (other than
the work room atmosphere) has been largely overlooked
as an important part of the total human exposure to
air pollutants. The assumption appears to have been
made that ambient outdoor air quality standards are
sufficient to protect human health. Since the average
person spends most of his life indoors, the indoor
environment represents a new frontier for analysis,
study and determination of its importance in affecting
human health and well being.
The results of limited studies carried out thus
far permit the following conclusions:
1. Indoor concentrations of CO can be assumed to be
equal to or higher than outdoor concentrations.
2. Indoor concentrations of particulates are gener-
ally lower than those outside, but the particles
inside are smaller in size and have a higher
organic content than those outside.
3. Present day air conditioning systems have no
effect on CO concentrations and little effect on
other pollutant gases and fine particulate matter.
4. Many activities carried out in the home contribute
to the degradation of indoor air quality. Homes
10.
11.
12.
13.
with gas stoves can have indoor concentration-time
relationships which exceed certain of the ambient
air quality standards for NO2 and CO if such
standards were applicable indoors.
REFERENCES
Benson, F. B., J. J. Henderson and D. E. Caldwell,
Tndoor-Outdoor Air Pollution Relationships,
Vol. I - A Literature Review and Vol. II - An
Annotated Bibliography, Publication Nos. AP-112
and AP—112b, II. S. Environmental Protection Agency,
Research Triangle Park, No. Carolina, August 1972.
Yocom, J. E., W. L. Clink and W. A. Cotg, Indoor/
Outdoor Air Quality Relationships, J. Air Poll.
Contr. Assoc. ^21:251-259, May 1971.
Yocom, J. E., W. A. Cot£ and W. L. Clink, Summary
Report of A Study of Indoor/Outdoor Air Pollution
Relationships to the National Air Pollution Control
Administration, Contract No. CPA-22-69-14, TRC -
The Research Corporation of New England, Hartford,
Connecticut, Vol. I and II, 1970.
Giles, W. H., R. E. Lee and L, H. Dworetzky, The
Influence of Traffic-Generated Carbon Monoxide
on Indoor Air Quality of an Air Rights and a
Canyon Structure, Proceedings - Institute of
Environmental Sciences.
Phair, J. J., R. J. Shepard, G. C. R. Carey and
M. L. Thompson, The Estimation of Gaseous Acid
in Domestic Premises, Brit. J. Ind. Med. (London)
12:283-292, October 1958.
Kruglikova, Ts. P. and V. K. Efimova, Residential
Indoor Air Pollution with Atmospheric Sulfur
Dioxide, Gig. i Sanit. [Hyg. and Sanitation]
(Moscow), 2_3: 75-78, March 1958. /
Tomson, N. M., Z. V. Dubrovina and M. I.
Grigor'eva, Effect of Viscose Production Dis-
charges on the Health of Inhabitants. In: U.S.S.R.
Literature on Air Pollution and Related Occupa-
tional Diseases, Levine, B. S. (trans), 8/.140-144,
1963.
Biersteker, K. H. De Graaf and Ch. A. G. Nass,
Indoor Air Pollution in Rotterdam Houses, Int. J.
Air Water Poll., 55:343-350, 1965.
Field Study of Air Quality in Air Conditioned
Spaces, Second Season (1969-1970). Arthur D.
Little, Inc., Cambridge, Mass., RP-86, February
1970.
Field Study of Air Quality in Air Conditioned
Spaces. Arthur D, Little, Inc., Cambridge, Mass.,
RP-86, March 1969.
CotS, W. A., W. A. Wade, III and J. E. Yocom, A
Study of Indoor Air Quality. Final Report,
Contract 68-02-0745, Publication No. EPA 650/4-
74-042, Environmental Protection Agency, Research
Triangle Park, North Carolina, September 1974.
Lefcoe, N. M. and I. I. Inculet, Particulates in
Domestic Premises; 1. Ambient Levels and Central
Air Filtration, Arch. Environ. Health, 22^230-238,
February 1971.
Holcombe, J. K. and P. W. Kalika, The Effects of
Air Conditioning Components on Pollution in Intake
Air. Presented at the Semiannual Meeting of the
American Society of Heating, Refrigerating and
Air-Conditioning Engineers, Philadelphia, January
14-28, 1971.
5
12-2
-------
AN EXAMPLE OF THE USE OF MONITORING AND MODELING
FOR EVALUATING AIR POLLUTION CONTROL PLANS
F. L. Ludwig and R.T.H. Co Ilis
Stanford Research Institute
Menio Park, California
Summary
Several alternative strategies for improving
air quality in Sarajevo, Yugoslavia, have been eval-
uated through the use of diffusion modeling. Daily
S02 and soot concentrations had been monitored at 7
to 10 locations for nearly a decade, providing his-
torical data for evaluating air quality and demon-
strating the need for control measures. The data
also provided a basis for application of the Climato-
logical Dispersion Model for evaluation of the
prospective results of several air quality control
strategies: an oil-fired central heating plant -with
a tall stack, natural gas distribution, and the use
of solid fuels (lignite and coal), preprocessed to
reduce ash and sulfur content. The study demon-
strates the value of a simple monitoring program
and simple modeling techniques iiv providing an
Objective basis for rational evaluation and planning
of pollution control strategies,
introduction
Air quality simulation modeling provides useful
inputs to the study of air pollution, but it is
seldom incorporated into practical decision making
because of the complexity of the pollution problem,
or the lack of suitable data. Successful applica~
tion of modeling to facilitate the choice among air
pollution control strategies for Sarajevo, Yugoslavia,
was possible, however, through the use of the basic
Climatological Dispersion Model1(CDM), adapted to
Sarajevo's topographic setting, and the existing
pollution data, acquired by relatively simple moni-
toring techniques. This brief account should en-
courage others with similar operational problems,
Sarajevo is an historic city located in an up-
land valley of the Bosnian Mountains. Its population
has grown rapidly to some 300,000 with resultant
serious environmental problems. Major steps are
being taken to rehabilitate the environment including
improvement of air quality. In this process Sarajevo
sought assistance to complement the locally available
expertise during the application of objective methods
used to describe current conditions, and to show how
the situation would change as a result of available
alternative control strategies. The results of that
study, which provided a basis for assessing the need
for remedial measures and for evaluating the possible
options, are summarized below.
The Problem
The heating and cooking fuels used in Sarajevo
are lignite and oil burned in many small, individual
units. In winter, burning of dirty fuels during
frequent stable, stagnant meteorological conditions
in the valley produces severe S02 and particulate
pollution. The magnitude of the pollution was deter-
mined from daily weaaurements collected by Sarajevo's
Institute of Public Health over a 10-year period at
7 to 10 locations. Using the sampling equipment
shown in Figure 1, 24-hour averages of SOg concen-
trations are determined from hydrogen peroxide
bubbler collections, where the S02 is oxidized to
HgSO^ and analyzed by titration. Soot concentra-
tions are estimated from reflectance measurements
on the exposed filter paper. This simple sampling
system provides the data to assess the air quality
problem in Sarajevo and to use in simple dispersion
modeling.
Aifl (NtAKt
PHOXH4A
A.BOV
HOUNO
O O
Figure 1.
system.
Schematic diagram of air pollution sampling
Modeling also requires \find and atmospheric
stability summaries, such as joint frequency dis-
tributions of stability, wind speed, and direction.
Sarajevo data were not in an appropriate form for
computer processing, so we used wind roses and cloud
frequencies to estimate the required joint frequency
distribution. Uncertainties arise because daytime
cloud cover is not statistically independent of wind
conditions, but the results were adequate for
evaluating proposed control measures.
Current pollutant emissions provide a baseline
against which the various control plans can be
evaluated. Figure 2 shows the current distribution
of S02 emissions In Sarajevo. The heavy emissions
in the eastern part of town come from older residences
burning lignite. Apartments with centralized heating
are located in the central part of the map.
As shown in Figure 3, heavy emissions and fre-
quently poor atmospheric ventilation produce winter
average SOg concentrations exceeding 600 (Agm m"3
in the densely populated areas. Figure 4 shows the
annual frequency at which different concentrations
are exceeded. At the worst location, Maslesa, the
24-hour SOg concentrations exceed the U.S. primary
12-3
-------
Figure 2. Annual SO emmissions (tons per year).
standard about 80 days per year. The 24-hour
_ •)
average concentration exceeds 1000 pg about 10
days per year. Even at the cleaner sites, e.g.,
Bjelave, the U.S. standards are exceeded 20 days per
year. Obviously, wintertime air pollution problems
are acute and they are expected to worsen if no
action is taken.
To assess future conditions we used the CDM
model that was developed for calculating average S02
concentrations from elevated point sources and from
ground-level distributed sources. The CDM uses
Gaussian plume approximations and requires calibra-
tion against existing measurements. Calibration
with winter average concentrations at ten sampling
sites yielded a correlation of 0.84.
Using emissions projected for 1985, without
remedial measures, CDM forecasts the very high
•Bjelave
Tucovic
Dobf j j
• Parovic
• \ •Lenjmova
Hrosno
1000
vidjKov.ac,-
Scole Kilometers
concentrations shown in Figure 5. The highest
average concentrations will increase about 50%
and the area where the maximum occurs will shift
westward toward the areas planned for new apart-
ments. The area and the population that will be
exposed to high concentrations will be larger than
at present.
3000
2000
1000
800
600
z 400
o
I—
a 200
p
z
tu
o
o
z
o
h-
<
cc
z
LU
o
z
o
100
80
60
40
3000
2000
1000
800
600
400
1 1
I 11 I i i i
1 1
—
S°2
BJELAVE
Particulate or
—
Standards
Soot Standards
I
—
N
\so,
U.S. Primary
nL
—
U.S. Secondary
U.S. Primary
--
USSR, Serbia
X
USSR
\
U.S. Secondary
—
California
¦v
California
SOOT^^v
—
1 1
1 1 1 1 1 1 1 "
i i
I I II TT
MASLESA
100
80
60
U.S. Primary
U.S. Secondary
SOO I
USSR, Serbia
California
U.S. Primary
USSR
U.S. Secondary
California -
4 20 40 80 150 250
10 30 60 100 200 275
DAYS/YEAR EXCEEDING ORDINATE VALUE
Figure 3. Winter (December, January, February) average
SO„ concentration (ngm m~3).
Figure 4. Annual frequencies of SOa
trations that exceed standards.
and soot concen-
12-3
3UTMIR
TREBEVlC
X
-------
Figure 5. Forecast 1985 winter average S02 concentrations (|j,gm m~3) with no controls.
Alternative Solutions
Ozm«
Tucovicj/
Uobrovoljoilw
• PorovicQ
Since air pollution control measures are
expensive, it is prudent to estimate the effects
of alternative control measures before they are
implemented. The CDM model -was used to evaluate
three alternative measures: (1) Use of low sulfur
oil and solid fuels with reduced sulfur and ash
content, (2) A large heating: plant (270-m stack)
serving about 50% of the population—with the
remainder using higher quality fuel as in (1), .
(3) Natural gas service to 80 or 90 percent of the
population—with the remainder using the higher
quality fuel.
The use of specially processed fuel (1) is
relatively inexpensive and can serve as an inter-
mediate measure even if another alternative is
adopted. It is not as effective as (2) and (3),
but Figure 6 shows that it would reduce the highest
concentrations to about one-third of the values
projected if no measures were taken.
An expensive central heating plant could only
serve areas with the population concentrated in
apartment complexes; others would use alternative
fuels. In modeling the effectiveness of this
alternative, the plume was assumed to follow a path
midway between a level one—intersecting the higher
terrain—and one parallel to the underlying surface—
at always the same height above the ground. The
projected 1985 winter average SOg concentrations
(Figure 7) show improved air quality only in the
areas served by the central heating system. This
limited improvement (compared with Figure 6) must be
weighed against the cost. In addition, these
projections are for average conditions and do not
indicate the occasional severe ground-level concen-
trations that would undoubtedly occur under anomalous
wind and stability conditions.
Natural gas, acquired from the USSR, would be
less costly and more effective than the central
heating plant. Figure 8 shows expected 1985 winter
so
',°p\
100-
SARAJEVO
!00
¦200-
IS 0
¦too
Figure 6. Forecast 1985 winter average S0a concentrations (jigm m"3) if light oil and smokeless fuel are used.
3 12-3
-------
Figure 7. Estimated 1985 winter average S02 concentrations (M-gm m 3) with central heating plant (see text for
assumptions).
average SC>2 concentrations if 80% of the population
is served; 90% service reduces the concentrations
only slightly more. The expense of the extra service
would be unwarranted because the recipients would be
in the hilly eastern parts of the city where
installation costs would be greatest.
Conelus ions
Simple measurements and simple modeling provide
valuable inputs to practical decision making in
air pollution control. Certainly to a first approx-
imation, they provide measures of conditions and
possible developments that could otherwise only be
assessed qualitatively and subjectively.
Acknowledgments:
We wish to acknowledge the contributions of
the many Sarajevans who participated in this study:
in particular, Mr. Aganovic, Ms. Arnautovic ,
Ms. Arifoyic , Ms. Sofilja, and Ms. Danon. Professor
Paradiz (of Ljubliana) provided valuable experi-
mental data, especially of stability conditions.
Reference:
1. Busse, A.D. and J.R. Zimmerman, 1973: User's
Guide for the Climatological Dispersion Model.
EPA Report No. R4-73-024, 131 pp.
Figure 8. Forecast 1985 winter average S02 concentrations (M-gm m"3) with 80 percent gas usage.
m
Ozme \
SARAJffVO
Tucovicc
-------
DETAILED CHEMICAL ANALYSIS OF AIRBORNE PARTICULATES
D. A. Levaggi, J. S. Sandberg, R. E. DeMandel and M. Feldstein
Bay Area Air Pollution Control District, San Francisco, Calif.
Summary
Extensive chemical analysis of suspended airborne
particles was undertaken by the Bay Area Air Pollution
Control District so that more meaningful interpretation
of total mass loading data could be made and more se-
rious problem areas differentiated. From high-volume
samples taken with cellulose paper as the collecting
nedium, analyses were performed for nitrate, sulfate,
chloride and silicon. The data viere derived from over
1600 individual samples, taken at 8 monitoring stations
with diverse geographical backgrounds, although within
a common air basin.
Since the silicates and chlorides are primarily
non-anthropogenic, they were subtracted from the total
suspended particulate (TSP) values to give total an-
thropogenic suspended particulate (TASP) concentra-
tions, which were 24 to 41 °L lower than the TSP values.
The sulfate and nitrate components of TASP were
studied in more detail, since their finer size range
implies greater visibility reduction and health effects.
In the Bay Area, the five-year mean for sulfates is
2.68 pg/m3, only slightly above the remote nonurban
background level; however, the five-year mean for ni-
trates is 2.78 pg/m^, well above national urban average.
The sulfate means over 3 pg/m3 were limited to
stations in a northern industrial arc. The relatively
high nitrate levels coincided both geographically and
temporally with high photochemical oxidant levels, and
occurred in a different subregion than the higher sul-
fate values. Both the nitrate and the sulfate data
appear to offer more promising indices of air quality
than do TSP data, for both visibility and health effects.
Introduction
In order to derive more specific and useful infor-
mation from total suspended particulate (TSP) data, the
Technical Services Division of the Bay Area Air Pollu-
tion Control District in 1969 began a program of ex-
tensive chemical analysis of filters exposed in an
eight-station network of high-volume samplers. The use
of cellulose rather than glass filters enables a deter-
mination of the important silicate component of TSP so
that a successful mass balance became possible. The
silicate and chloride constituents were determined, and
a first approximation to a total anthropogenic suspend-
ed particulate (TASP) Value was then derived by exclud-
ing these components from the original TSP value.1
Results on the more critical TASP components of
sulfates and nitrates were then analyzed, so as to de-
fine problem areas and provide a better focus for con-
trol strategies.2 Finklea et_ ad.3 have recently stress-
ed the importance of acid-sulfate aerosols in terms of
health effects, and, in a more conservative vein,
Hemeon^ has emphasized the sulfate species In a ration-
al particulate standard. The importance of nitrate
aerosols has been pointed out by Robinson and Robbins,^
Moreover, the role of nitrogen oxides in photochemical
smog gives the nitrogen compounds special interest for
California, as in the study of Pasadena aerosol by
Novakov et al.^
Analytical Procedure
For this study the use of a cellulose (Whatman ^1)
filter was critical based on the following analytical
- 1
requirements: 1) the necessity of silicon analysis and
2) the absence of chemical impurities associated with
other filter media. Obtaining total mass collected is
more cumbersome using a cellulose filter due to its hy-
groscopicity, but proper handling, humidity control, and
care can overcome the problem.
Silicon analysis is performed by ashing a portion
of the filter in a nickel crucible. The ash is then
fused with sodium carbonate, dissolved and the molybde-
num blue method applied to an aliquot for total silicon
content. The critical factor used for converting the
resultant total silicon (expressed as SiOj) to the to-
tal weight of silicates collected is 1.25. This factor
was partially derived using data developed by John et_
al.? from Hi-Vol samples taken in the Bay Area.
Chloride was determined by water extraction of
one-quarter filter and subsequent analysis by the tur-
bidimetric silver nitrate procedure. It is assumed all
the chloride is present as sodium chloride, and a factor
of 1.67 was used to convert the chloride ion to the
corresponding weight collected.
The sulfate and nitrate analyses were performed on
water extracts from one-quarter of the collected hi-
volume filter. The use of cellulose allows a medium
containing zero backgrounds of these cations as well as
a neutral surface, eliminating any possible gaseous sul-
fur dioxide conversion to apparent "particulate sulfate".
The sulfate analysis was performed by the classical ba-
rium sulfate turbidimetric procedure. Nitrate determi-
nation was performed by reduction to nitrite and subse-
quent analysis by the well known Griess reaction.
Data Collection Procedures
High-volume filter data were collected at eight
stations in the Bay Area (see Figure 1). These stations,
as abbreviated in the tables to follow, reading clock-
wise from the San Francisco (SF) central station are
San Rafael (SR), Richmond (RI), Pittsburg (PT), Fremont
(FR), Livermore (LI), San Jose (SJ), and Redwood City
(RC). As the map suggests, complex topography and vary-
ing sources might well result in a different particulate
mix for each station. From April 1969 through April
1972, the samplers were operated three days a week, and
every Thursday's data were used in this study. From
May 1972 through December 1973, the sampling schedule
was changed to every third day, so as to randomize the
day-of-week factor, and every fourth sample was che-
mically analyzed. Total samples analyzed exceeded 1600.
Results
detailed consideration of seasonal and "meteorologi-
cal factors associated with variations in the components
data were given in earlier papers for silicate and chlo-
ride^ and for sulfate and nitrate?. In this synthesis,
the five-year averages for each component are given in
Table 1, and the highest and second highest values for
each component are given in Table 2. Both may best be
examined In association with Fig. 1. Not unexpectedly,
the highest chloride values are found nearest the Pa-
cific Ocean and the more saline waters of the Bay. The
highest silicates are in the more inland valley stations,
adjacent to the dry tawny hills during our long dry sea-
son. The widely separated crescents for highest sul-
fates and nitrates, the major anthropogenic components,
are less expected but readily explained. The northern
crescent of sulfates coincides with the Bay Area's
12-4
-------
Table 1. Five-year average component values
(pg/m3) at each station for Si02, CI, SO4 and NO^
Sta.
Si02
CI
S04
NO3
SF
5.18
4.04
3.02
1.96
SR
11.54
2.26
2.14
1.88
RI
9.89
2.58
3.34
2.22
PT
15.28
1.43
3.26
2.94
FR
13.31
2.47
2.64
3.44
LI
21.18
1.56
2.24
3.04
SJ
17.35
2.16
2.44
4.02
RC
10.17
2.27
2.40
2.72
Avg.
12.99
2.35
2.68
2.78
heavy industry, primarily petrochemical. The southern
crescent of nitrates encompasses suburbanized shelter-
ed valleys with greatest motor vehicle miles travelled.
Silicate Component: Silicate values are highest at the
two most inland stations (Livermore and Pittsburg) and
lowest at the two most maritime stations (San Francisco
and Richmond). Seasonal variation is very large, with
maximums in the dry, dusty, fall (typically following
five or six months of no rainfall), and minimums in the
in the wet winter months. The extreme values (Table 2)
are of particular interest in that the silicate compo-
nent alone at inland stations on adverse fall days may
exceed the California AGM particulate standard of 60
jig/m3.
MAP Of THE
SAN FRANCISCO BAY AREA
Figure 1. Map of San Francisco Bay Area showing sta-
tion locations: SF, San Francisco; SR, San Rafael; RI,
Richmond; PT, Pittsburg; LI, Livermore; FR, Fremont;
SJ, San Jose; and RC, Redwood City. Areas with five-
year mean nitrate or sulfate values >3 jig/m3 are shaded.
To examine the relationship of silicate loading to
surface wind direction for each station, the percentage
departure of silicate concentration for each wind octant
from its station mean value was calculated. All sta-
tions experience below average levels with ESE, SSE and
SSW winds. These directions are usually associated
with winter storms and with moist, well-mixed, unstable
or neutral maritime air flowing over damp soil. When
winds are NNW, NNE and ENE, above average levels are
usually observed. Wind from the NNE is often associated
with cold air drainage out of the Central Valley in
winter. This shallow layer of stable air is often high
in entrained particulate matter. Even more Important
are the strong, warm, dry north and northeast winds of
autumn which are ideal for mechanical entrainment and
transport of large dust particles. Near normal dust
levels are found for WSW and WNW which are the most
common wind directions, usually reflecting onshore flow
of marine air during the dry summer season.
Chloride Component: As shown in Tables 1 and 2, the
chloride results demonstrate and quantify the expected
decrease with distance from the Pacific Ocean, modified
by the terrain-induced airflow trajectories. The high-
est average value occurs at San Francisco, which is
nearest the ocean. The lowest averages at Pittsburg and
Livermore are recorded at the two most inland stations.
The only noteworthy reversal is Richmond, more inland
than San Rafael, but more directly exposed to marine
trajectories through the Golden Gate.
Since extreme values are of particular concern both
in evaluating episodes and determining limits, the high-
est and second highest chloride component values are
also included in Table 2. Typical high values are 7-10
pg/m3 across the network. As expected, the high chlo-
ride values are associated primarily with westerly wind
components and with relatively strong airflow.
Sulfate Component: The District's five-year sulfate
average of 2.68 pg/m3 (Table 1) reflects both the
BAAPCD's stringent control program and the extensive use
of natural gas in the Bay Area. The remote nonurban
sulfate average (for 1966-67) has been cited by McMullen
e£ al.8 as 2.51 |ig/m3. Comparison with the average for
217 urban stations of 10.1 pg/m3 places the Bay Area much
Table 2. Highest and second highest component values (jig/m3)
over five years at each station (and dates of occurrence) for Si02, CI, SO4 and NO3
Sta. Si02 (Date) CI (Date) SO4 (Date) NO3 (Date)
SF
34.4
8-23-73
13.9
6-11-70
16.5
11-27-69
9.8
1- 6-72
29.5
8-11-73
13.5
5-28-70
13.1
1- 1-70
9.4
1-13-72
SR
64.7
9-24-70
9.8
5-31-73
22.0
12-14-72
25.3
12-14-72
38.0
5-29-69
8.5
7- 1-71
19.0
12- 9-73
19.0
1- 7-73
RI
53.2
3-26-70
11.2
7- 1-71
14.8
12- 9-73
22.6
12-14-72
40.6
2-10-72
10.3
5-19-73
14.1
7- 3-69
15.1
1- 7-73
PT
72.0
10- 8-70
4.8
5-19-73
16.7
9- 4-69
27.9
12-14-72
62.3
10- 1-70
4.7
2-12-70
13.5
10-15-70
21.6
1- 7-73
FR
54.3
7-17-69
11.0
4-15-71
17.7
10-15-70
24.3
12-14-72
45.4
7-31-69
10.3
5-20-71
8.2
9-15-72
20.1
9-16-73
LI
67.9
9- 3-70
7.8
5-28-70
8.9
9-15-72
22.6
12-14-72
57.2
10- 8-70
7.1
6-11-70
8.7
10-15-70
19.4
9-28-73
SJ
54.0
10- 8-70
8.7
2-12-70
11.6
10-15-70
14.4
10-16-73
46.9
10- 1-70
7.9
2- 5-70
7.7
9-11-69
14.3
9-28-73
RC
52.4
10- 1-70
9.8
8- 7-69
9.5
9-15-72
20.0
12-14-72
34.8
3-12-70
9.7
5-20-71
9.0
10- 1-70
14.9
10-16-73
- 2 -
12-4
-------
nearer the rural than the urban category with respect
to sulfates. The maritime background level for sulfate
aerosol has been estimated^ at 2 to 5 pg/m3, a value
which meteorologically could be supported by the San
Francisco 5-year August average of 2.1 pg/m3. Extreme
values, which are of particular concern to a control
program, are shorn In Table 2.
Despite this generally favorable picture, a "sul-
fate belt" of higher concentrations exists (Fig. 1).
Richmond at 3.34 pg/m3 and Pittsburg at 3.26 pg/m3 are
anchorpoints of an industrial arc that stretches along
the shores of a northeast extension of the bay. This
arc is responsible (as of 1973) for 210 of the District*
260 tons/day SOx emissions. Percentage departures from
the mean for each wind direction provide a useful pic-
ture of each station in relation to the sources affect-
ing it. The most significant positive departures are
+105% for ENE at San Rafael, and +55% for NNE at San
Francisco, both pointing towards the Contra Costa in-
dustrial complex.
Studies of the size distribution of Bay Area sul-
fur aerosols by Cahill' show the sulfur species to be
in the smallest size range of 0.65 microns or less.
Since this size range is the most respirable and visi-
bility-reducing, the sulfates measured in our study
may be important to air quality even at the low levels
observed.
Nitrate Component: Unlike sulfates, the 5-year Bay
Area nitrate average of 2.78 fig/m3 (Table 1) is well
above that of eastern urban centers. The remote non-
urban nitrate average as given by McMullen et_ al_. ^ is
0.85 iig/m3, and the average for 217 urban stations is
2.4 f)g/m3. As with sulfates, a tionurbati maritime back-
ground level for nitrates is supported by the San Fran-
cisco 5-year August average of 0.7 fig/m3.
The geographic distribution of high nitrate values
is very different from that of sulfates with a "nitrate
belt" across the southern portion of the District (Fig.
1), again defined as exceeding an arbitrary 3 pg/m3.
It is not unexpected that the high nitrate area coin-
cides with that of high photochemical oxidant and ni-
trogen dioxide values. Nitrates are one assumed end-
product of the photochemical process.
The highest nitrate values (Table 2) make several
important points with differing control implications.
The highest values at six stations occurred on 12-14-
72, when a strong nocturnal inversion and strong di-
rectional wind shear produced restricted vertical dis-
persion. Even more important, meteorology affected
sources in that the coldest December of the century for
many Bay Area stations caused a shortage of natural gas
and a switch by "interruptible" sources throughout the
Bay Area to fuel oil. These 12-14-72 extremes (in a
nonphotochemical period) indicate the importance of
stationary sources of N0X, particularly under energy
crisis circumstances.
More typical high nitrate days occur on days with
high photochemical oxidant and N0j and with the low
afternoon temperature inversions most conducive to pho-
tochemical development. Automotive sources, which ac-
count for 63% of the District's N0X emission inventory,
are clearly indicated on these days, and there are many
more days of this class than there are "interruptible"
fuel-switch days.
General Consideration of TASP
Although particle size distribution was not mea-
sured in this study, reasonable inferences from the
chemical species involved can be drawn from other
studies. Whitby's recent work^O has emphasized the
Table 3. Mean total anthropogenic suspended
particulates (TASP) by station as adjusted
for S102 and CI components (1969-1973)
Si02
CI
% Non-
TSP
Mean
Mean
TASP
anthro-
Sta.
Mean
(xl.25)
(xl.67)
Mean
pogenic
SF
53.6
6.5
6.7
40.4
24.6
SR
54.1
14.5
3.7
35.9
33.6
RI
56.3
12.4
4.3
39.6
29.7
PT
63.9
19.1
2.3
42.5
33.5
FR
69.1
16.6
4.2
48.3
30.1
LI
70.6
26.5
2.7
41.4
41.3
SJ
74.4
21.6
3.5
49.3
33.7
RC
54.0
12.7
3.8
37.5
30.6
Avg.
62.0
16.3
3.8
41.9
32.4
multimodal nature of atmospheric aerosol size distribu-
tions: 1) the coarse particle mode, generally at diame-
ters from 5-50 p, 2) the transient nuclei mode in the
range from 0.01-0.04 p and 3) the accumulation mode in
the 0.1-1.0 p range. The accumulation mode is the long-
est lived, is the most respirable, accounts for about
95% of visibility reduction and thus the most significant
factor in air pollution. He has concluded that there is
practically no interaction in the aerosol state between
the fine particles and the mechanically generated course
particles larger than 2.0 p.
The association of silicates with trace elements
characteristic of airborne surface dust? supports the
reasonable assumption that silicates are primarily in
the coarse particle mode. Such silicates at the levels
found in the Bay Area would aid greatly in accounting for
the poor relationship between TSP and visibility reduc-
tion. H The recent (January, 1974) EPA "Guidelines for
Designation of Air Quality Maintenance Areas'1^ takes
cognizance of this non-anthropogenic factor in recommend-
ing that high TSP concentrations "due to uncontrollable
fugitive dust from natural causes" not be a basis for
projecting air quality.
As a first approximation to a more meaningful total
anthropogenic suspended particulate (TASP), the 5-year
averages for eight Bay Area stations were adjusted to ex-
clude the silicate and chloride components, as given in
Table 3. It is not reasonable or necessary to assume
thai all of the silicate or chloride is non-anthropogenic.
On the other hand, as determined by Lundgren.^3 a large
part of the coarse particles are organic materials of
non-anthropogenic origin, i.e., insect fragments, fibers
and pollens. A differentiation of such organics and those
of combustion origin cannot be made in our analytical
procedure. Thus, a conservative trade off of such or-
ganics would appear to justify the full exclusion of si-
licate and chlorides in this TASP first approximation.
The TASP values compared with TSP station values in
Table 3 give a measure of air quality that relates far
better to local experience. For example, San Francisco
had poorer TASP air quality than upwind suburban San
Rafael, a reversal of the TSP values. The low-density
inland towns of Pittsburg and Livermore in the open,
dusty hills have much lower TASP than TSP values. Fremont
and San Jose at the mouth of the sheltered Santa Clara
Valley, became more sharply the focus of high TASP values.
These higher values would suggest a greater particulate
level in Whitby's "accumulation mode" and would help ex-
plain the much greater visibility restriction reported
in this area.
Particulate. Gas and Emission Relationships
The relationship of gases to particulates for the
sulfur and nitrogen species is a complex one involving
slow transformations. It is further complicated by back-
3 - 12-4
-------
250 .2
1970
1971 1972
Finure 2. Annual 8-station moans for sulfite particulate and S02
!as (both in ug/m5) compared with District annual S0X emission
nventories (In tons/day).
ground levels (which for sulfates are close to observed
levels In most of the Bay Area) and by direct particu-
late emissions. With the mean annual particulate spe-
cies data now available for the first time, it was of
interest to compare them with the corresponding gas
data and emissions inventory data. Mean annual values
for the entire air basin were thus compared, as shown
in Figure 2 for sulfur species. The annual S0X emissions
decreased sharply in 1970 and 1971 in response to strin-
gent District regulations but rose in 1972 and 1973 in
response to curtailments and shortages of clean fuels.
Both the S02 gas measurement and the SO4 particulate
measurements tracked the falling emissions but the SO^
data reflected the increase considerably better.
The relations among nitrogen species are less
straightforward. The N0X emissions from 1969 to 1973
showed a slow, steady increase from automotive sources.
This increase was best reflected in the nitrate curve,
except for an anomalously sharp rise in 1972, followed
by a dip in 1973. An explanation lies in the December
1972 "interruptible" source fuel switch which produced
a monthly average of 11.3 (jg/m^. Most noteworthy in
the nitrate data was the close geographic relationship
to high photochemical oxidant, including year-to-year
parallel displacements of highest nitrate and oxidant
from one southern crescent station to another.
References
1. Levaggi, D.A., J.S. Sandberg, M. Feldstein, and S. -
Twiss, "Total anthropogenic suspended particulate as
derived from chemical analysis of chloride and silicate
on high-volume samples." J. Air Poll. Control Assoc.
24, in press, (1975).
2. Sandberg, J.S., D.A. Levaggi, R.E. DeMandel, and
W. Siu, "Sulfate and nitrate particulate as related to
SO2 and NOx gases and emissions." Presented at 68th
Ann. Mtg. of Air Poll. Control Assoc., (1975).
3. Finklea, J.T., D.B. Turner, G.G. Akland, R.I. Larsen,
B. Hesselblad, and S.D. Shearer, Briefing Notes: A Status
Report on Sulfur Oxides, EPA (1974).
4. Hemeon, W.C.L., "A critical review of regulations
for the control of particulate emissions." J. Air Poll.
Control Assoc. 23 (5) 376-387 (1974).
5. Robinson, E. and R.C. Robbins, Emissions, Concentra-
tions, and Fate of Particulate Atmospheric Pollutants,
Stanford Research Inst., Menlo Park (1971).
6. Novakov, T., P.K. Mueller, A.E. Alcorer and J.W.
Otvos, "Chemical composition of Pasadena aerosol by par-
ticulate size and time of day: III. Chemical states of
nitrogen and sulfur by photoelectron spectroscopy." J.
Coll. and Interface Scl. 39 225-234 (1972).
7. John, W., R. Kaifer, K. Rahn, and J.J. Wesolowski,
"Trace element concentrations in aerosols from the San
Francisco Bay Area," Atmos. Environ.. _7> 107-118 (1973).
8. McMullen, T.B., R.B. Faoro, and G.B. Morgan, "Pro-
file of pollutant fractions in nonurban suspended par-
ticulate matter," J. Air Poll.. Control Assoc., 20
369-372 (1970).
9. Cahill, T.A., Crocker Nuclear Laboratory, University
of Ca., Davis, CA. , Private Communication (1975).
Conclusions
1. By the periodic use of cellulose rather than glass
filters, the widely deployed high-volume particulate
sampling networks can be used to obtain a silicate
component In TSP values.
2. The use of a silicate component correction to TSP
values can provide a reasonable first approximation
Total Anthropogenic Suspended Particulate (TASP) values.
3. Over a five-year period, the annual mean sulfate
data track to SOx emissions data more closely than do
the annual SO2 gas data.
4. The nitrate salts (primarily ammonium nitrate) are
of greatest importance in areas with greatest visibili-
ty restriction; as major end-products of the photo-
chemical process, they may provide a good index of
photochemical activity.
5. Both sulfate and nitrate components of TASP appear
to offer better indices of air quality than do TSP data.
Thus future air quality standards should give appropriate
weight to these species.
10. Whitby, K.T., "On the multimodal nature of atmos-
pheric aerosol size distributions, "Presented at VIII
International Conference on Nucleation, Leningrad, USSR
(Sept. 1973).
11. Thuillier, R., J. Sandberg, W. Siu and M. Feldstein
"Suspended particulate and relative humidity as related
to visibility reduction," presented at 66th Ann. Mtg. of
Air Poll. Control Assoc. (1973).
12. Environmental Protection Agency, Guidelines for
Designation of Air Quality Maintenance Areas. OAQPS No.
1. 2-016 (Jan. 1974).
13. Lundgren, D.A., "Mass distribution of large atmos-
pheric particles," Ph.D. Thesis, Univ. of Minn. (1973).
- 4 -
12-4
-------
THE CONTINUOUS MONITORING OF SULFUR DIOXIDE AND
SUSPENDED PARTICULATES NEAR NON-FERROUS SMELTERS
R. D. Putnam, Ph.D., M. 0. Varner, K. W. Nelson
M. A. Yeager, R. B. Watson
Department of Environmental Sciences
ASARCO, Inc.
3422 South 700 West
Salt Lake City, Utah 84119
Summary
In 1937, American Smelting and Refining Company
initiated continuous monitoring for sulfur dioxide
near its smelters, utilizing the Thomas conductimetric
analyzer. Since that time, the analyzer has undergone
many developmental changes and is now called the
ASARCO Autometer S02 Monitor. It is considered to be
the most reliable, accurate, and maintenance-free SO2
device for use near isolated non-ferrous smelters.
The advantages and disadvantages of the ASARCO Auto-
meter are presented.
In 1967, ASARCO developed a unique low-volume air
filter system for continuous monitoring of suspended
particulate matter near its non-ferrous metal opera-
tions. The system basically consists of filter holder,
pump, dry test meter and pulsation dampener. Air at
the rate of one cubic foot per minute (cfm) is drawn
through a 4-inch Gelman metricel filter, Type GA-4,
with an 0.8y pore size. A "respirable" or particle
size selective unit has also been developed, which
operates at approximately 0.5 cfm. The filters are
normally changed once per week or once every two weeks.
Advantages and disadvantages of the system are also
shown, particularly in relating the average concentra-
tions obtained to 24-hour readings normally collected
with a high volume air sampler.
Trace metal determinations are discussed and cor-
relations between SO2 and SO4 data are presented.
In 1937, ASARCO initiated continuous monitoring
stations for sulfur dioxide near its smelters utiliz-
ing the Thomas Conductimetric Analyzer first invented
in 1928 by Dr. M. D. Thomas, an ASARCO scientist.1
The first monitoring networks were limited in number
and were placed around the smelters for the primary
purposes of studying the relationships of ambient SO2
concentrations to varying meteorological conditions
and to the occasional occurrences of observable
effects on SO2 sensitive species of vegetation. Since
this time the number of stations has been expanded to
include a total of 50 monitors located at a number of
ASARCO operations throughout the United States. By
the addition of telemetry equipment, meteorological
instruments and computers, the role of these monitors
has been expanded and they now play an important part
in the successful operation of supplemental control
systems.^
In 1967 ASARCO developed a unique low volume air
filter system for continuous monitoring of suspended
particulates, trace metals, and sulfates. A total of
22 low volume samplers have now been installed near
our non-ferrous metal operations.
SO2 Monitor
The SO2 monitors still employ the conductivity
method utilized by Di. Thomas in the original Thomas
autometer. However, over the past 37 years this
instrument has been subjected to many developmental
changes and is now called the ASARCO Autometer S02
Monitor. The most significant change was the addition
of a temperature compensator that automatically makes
adjustments for temperature fluctuations in the solu-
tion. The recycling mechanism that changes the solu-
tion in the reaction vessel has also been modified to
permit the operator to select sampling times of 15
minute intervals. Maintenance problems have been
greatly reduced by the replacement of the mechanical
valves with more durable teflon stopcock valves. The
original recorder has been replaced with a Leeds and
Northrup "Speedomax" which provides a more linear re-
cording of the S02 concentrations versus time.
The primary advantage of the ASARCO Autometer is
that it is extremely reliable, accurate, and rela-
tively maintenance free. To determine the reliability
of the SO2 monitors, the records for the entire year
of 1974 from three sampling stations were randomly
selected and reviewed to determine the percent of time
that the units were functioning properly. The relia-
bility ranged from a low of 92.2% to a high of 99.3%
with an average of 96.7%. The lower reading of 92.2%
was attributed to a failure in the pen assembly of the
unit which placed it out of service until a replace-
ment could be installed.
Because of the remote areas in which many of the
units are located it is impossible to check the units
on a daily basis. However, to insure the proper oper-
ation of the instruments we have found that a routine
visual check should be made every three days. Other
than the normal servicing of the instruments, such as
changing the solution and maintaining an adequate
supply of ink in the recorder, maintenance is normally
limited to a yearly inspection by representatives of
the Department of Environmental Sciences. At this
time all worn parts are replaced and a dynamic cali-
bration of the instrument is conducted. The dynamic
calibration procedure is to inject a known amount of
S02 into the inlet tube and compare this with the
amount shown on the recorder chart. Normally about
1.6 parts per million (ppm) is injected and if the
instrument fails to record that value within plus or
minus 0.05 ppm, readjustment of the electrode plates
is required followed by another dynamic calibration.
Recalibration is also done whenever a major unit such
as recorders, amplifiers, or electrodes are replaced
or when the unit is moved to a new site. Static cali-
brations are conducted on a routine basis by service
representatives in the field.
Apart from the low maintenance requirements this
instrument has a number of advantages over other com-
mercially available instruments. Because of the high
air volume on which this analyzer operates it has a
response time of approximately 10 seconds. By utiliz-
ing a batch-type operation the unit automatically
re-zeros at the beginning of each thirty minute sam-
pling period. The spent solution in the reaction
vessel can be changed in thirty seconds resulting in a
down time of only one minute per hour. The addition
of the temperature compensator allows the unit to be
operated at a temperature range of from 5° to 45° cen-
tigrade. Generally this eliminates the need to
condition the air within the building in which the unit
is housed.
The primary disadvantage of this unit is that it
employs the conductivity method which has been shown to
12-5
-------
be subject to a number of interferences. The interfer-
ents of primary concern are hydrochloric acid, ammonia,
and nitrogen dioxide. However, the concentration of
these substances in the ambient air is not high enough
to cause interferences unless there is a major source
within the immediate vicinity of a monitoring station.
The Environmental Protection Agency has recently
promulgated standards setting forth the requirements
that SO2 monitors must achieve in order to be classi-
fied as a reference or equivalent method.3 The ASARCO
Autometer is currently being tested by an independent
consulting firm to determine if this instrument will
meet these specifications and the results of these
tests should be completed at the end of this year.
Typical S02 concentrations found near a major
copper smelter are shown in Table 1. The 3-hour and
24-hour SO2 values represent maximum concentrations
that occurred at each station during 1974.
TABLE 1
Sulfur Dioxide Sampling Data for 1974
Concentration (ppm)
Maximum 24-hr.
Station Annual Average Average
1
2
3
Federal
Standard
0.007
0.005
0.012
0.020
0.090
0.051
0.061
0.140
Maximum 3-hr.
Average
0.57
0.31
0.15
0.50
Low Volume Particulate Sampler
bi-monthly basis. By doing so, the costs involved in
having plant personnel change filters at each station
on a twenty-four hour basis are reduced. It also
allows for continuous sampling over weekends and holi-
days without the problems normally encountered in
scheduling personnel on a seven day per week basis.
Although this system does provide an accurate
estimate of the average concentration of particulates
over a single week or two week period it does not per-
mit one to determine a maximum 24-hour concentration.
It does, however, provide an excellent means of deter-
mining the long-range impact on the environment in
remote areas. Another disadvantage of this system is
that if a mechanical malfunction should occur during
a sampling period a loss of data over a significantly
longer period of time may occur. To reduce this pos-
sibility we require that the low volume monitor be
checked on the same frequency as the SO2 monitor. If
a problem is encountered then the necessary mainte-
nance is performed and a new filter is installed
thereby reducing the loss of data to a period of two
or three days. Because of the durability and relative
simplicity of this operation we have not experienced
any loss of data which would exceed that normally en-
countered using the traditional methods of particulate
sampling such as the high volume sampler.
A comparison of the concentration of suspended
particulates collected using a high volume sampler
versus a low volume sampler is shown in Table 2. The
filters on the high volume sampler were changed daily
while the filter on the low volume sampler was uti-
lized for the entire two week period. The lower parti-
culate readings on the low volume filter can possibly
be attributed to errors in weighing or the chipping
of the filter on the outer edges. This is substanti-
ated in part by the fact that there were no significant
differences in the lead values reported.
The low volume sampling unit, as presently de-
signed, allows for the collection of both total and
respirable fractions of suspended particles. Both the
total and respirable fraction are collected using a
4-inch Gelman metricel filter. Type GA-4, with 0.8 y
pore size. This particular filter has an advantage
over others in that the background levels of trace
metals and sulfates are considerably lower than those
found in the traditional glass fiber filters. They
also dissolve completely when subjected to a nitric-
perchloric acid digestion which assures the analyst of
a complete recovery of the particulates collected on
the filter. The main disadvantage of these filters is
that they are relatively fragile and caution must be
taken during shipping and handling to prevent breakage.
The connection between the filters is made using
a high vacuum hose in which a glass or plastic
Y-connector has been inserted. The air flow is regu-
lated to each filter by means of a screw clamp. Air,
at the rate of one cubic foot per minute (cfm), is
drawn through the total particulate filter while a re-
duced flow of 0.5 cfm is drawn through the respirable
filter. Any standard vacuum pump capable of drawing
three cfm of air is suitable for use in this system.
One problem we have experienced in the past is
the flaking of particles from the filters due to ex-
cessive vibration caused by the pulsations from the
pump. This has been alleviated by placement of a pul-
sation dampener in the system. The simplest and least
expensive dampener consists of a sealed, one gallon
plastic container.
The air volume is measured using a standard dry
volume test meter. Experience has shown that the
model of dry test meter should be standardized to
reduce errors in recording the air volume.
The advantage of this system is that it is a low
maintenance method of collecting suspended particu-
late data. Because of the low air flow it is only
necessary to change the filters on a weekly or
TABLE 2
Average Concentrations of Particulates
Collected Over a Two Week Period Using
High and Low Volume Samplers
High Volume Sampler
Low Volume
Sampler
Particulate
Lead
Particulate
Lead
Station
(yg/m3)
(yg/m3)
(yg/m3)
(yg/m3)
1
36.3
0.47
26.8
0.40
2
38.6
0.50
19.7
0.51
3
36.0
0.50
42.3
0.62
Average
37.0
0.49
29.6
0.51
The annual average concentration of suspended
particles of lead and sulfate for the same three sta-
tions are shown in Table 3.
TABLE 3
Low Volume Sampling Data
(Yearly Average Concentration of Suspended
Particulate, Lead and Sulfate)
Concentration (yg/m3)
Station
Particulate
Lead
Sulfate
1
37.3
0.58
2.88
2
28.7
0.51
3.33
3
31.9
0.26
2.78
Average
32.6
0.45
3.00
2
12-5
-------
With the increasing concern over suspended sul-
fates in the atmosphere, studies have been initiated
to determine if there is a relationship between the
levels of SO2 and SO4 found near non-ferrous smelters.
The data from five sampling stations located within
a 45 mile radius of a copper smelter are shown in
Table 4. There is a positive correlation between SO2
and SO
-------
ON THE DEFINITION OF REQUIREMENTS FOR MONITORING OF THE PHYSICAL AND
CHEMICAL PROPERTIES OF TROPOSPHERIC SULFATE AEROSOLS
RELEVANT TO HEALTH EFFECTS
R. J. CHARLSON
A. P. WAGGONER
D. S. COVERT
N. C. AHLQUIST
Department of Civil Engineering
University of Washington
Seattle, Washington 98195
Abstract
Recent studies in Europe and the U.S. have shown that
SOg derived sulfate aerosols are present in the forms
(NH4)2S04, NH4HS04, H2S04 and possibly (NH4)3H(S04)r
Other work has shown that these aerosols are found in
the size range roughly between 0.1 and 1.0 pm, still
other work has demonstrated that the sulfate concen-
tration is highly correlated with light scattering.
Based on these results, knowledge of the hygroscopic
growth characteristics of sulfates and of pulmonary
deposition, this presentation will suggest a possible
set of measurements for monitoring sulfates vis a vis
the problem of health effects.
1. Introduction
As a consequence of burning of oil and coal, fossil
sulfur is released to the atmosphere, mainly as S02
and a small amount of some form of sulfates. The SOg
is subsequently recycled back to the ground via a
variety of pathways with varying residence times and
fluxes for each pathway. Figure 1 illustrates the
atmospheric portion of the sulfur cycle, with many
pathways and intermediate substances produced from
SOg on its way through the atmosphere.
Recent studies such as CHESS1 have implicated sulfates
as possibly being hazardous to man in even low concen-
trations (ca. 10 yg m ). Several different sulfates
are among the molecular forms naturally produced as
sulfur is cycled through the atmosphere. However,
sulfate per se is not a single molecular form, and
only a few sulfates are known to be produced from the
oxidation of SOg. Table 1 is a list of sulfate forms
known to exist in air, along with some of their more
outstanding chemical characteristics. Three features
should be noted:
a) There are at least these seven compounds, all of
which are sensed as "sulfate", but which have radical-
ly different chemical properties dictated by their
molecular nature.
b) Some sulfates are in the respirable, or optically
important submicrometer size classes, while others
are in larger sizes.
c) Sulfate as a whole comes from a wide variety of
sources, but much of it is emitted to the atmosphere
as S02.
This short note will consider the consequences of
these facts.
2. Molecular Forms of Sulfate Commonly Found in Air
Sulfates are either produced as such and directly in-
jected into the air, or they are formed in the atmos-
phere from gas-particle conversion. The first four
substances in Table 1 (HgS04 and its products of
netralization with NH^) seem to be formed from SOg
oxidation. The exact nature of the oxidation reactions
is currently being studied, but both homogeneous and
heterogeneous reactions are possible. The remainder
of the sulfates are directly injected by either indus-
trial or natural sources. The chemical nature of
these substances varies from the strong acid nature
to to
-O in
O to
O
E
P
<
u
•r- CO
B 0>
, ,
o
o
o
o
O f—
tO i- r—
§.
9
url
* P •!-
a.
O C E
r—~
i— to o
i
1
1
i
o
/sua.
A
A A3 A
r-»
C r—
n..~3 3
Q_
cuo a.
o
d
d
o
O w Q.
Q
O *—¦" o
p
CD
c
P
c
>»'r-
QJ
U
\
to
S aJ
C
J-
to
S- 05
c
'i—
U P
-a az
3 X
p crx
~. fO
>> >»
o. a>
a-ex
<0 •<- q;
CD >,
r-
o c u
* &
a r-
r- P r-
•i- a;
(j «r* -r-
•a u o
-o u
r-
0) **
C Oi
r— >
v> x
•r- «r—
ai to
»T5 O
3
o o
u a.
o a
¦o
tO O
O to P
3 <0
cri aj
o ra
a
U di > > c
o> 10
o> >
U r—
7 0 0)
s- <— -o
P >> CTl
¦p >>
U 0-.
a; O S- ro
c
jZ
o
P
•r- E
•r-
¦P O
3S
•P to
ra U
CO 4- -P
-a <+-
CM
CM
CM
3 O U
cy,
O
O
o
*0 3
c
X P
CO
I/)
oo
"O
o o a
o> 10
c t. o
CL
> r—
U U »r-
o o
O
O
•r-
o -p a
ro 3
•1- •!- im
•r-
a. u
r— (J
u a.
i- "O 3
C P
c
c
r— O
\o <0 &
a.
0) 4->
O T-
O c--1
O ro
3 U
4- 3
to im
JZ u
•r- -a
•r- X
¦r- X
Q. Q.
t? 3 CO
,
a>
p
f"
•r- +J
p
p
a;
(0
o
U T-
<0
a; f—¦
p
4-
•o t-
E
m
•r—
E >
3 a)
4-
r-*
to P
(C -r-
C
3 O
to P
r— 40
3
p—•
>
•r- *P
3 -
(/I
£ <0
a
c a»
E d
(/> i.
3 til
•r- 4-
£ a)
O -J
3 cn
E i
s- o
to P
r-
-
£ Si
3 3
to E
3
ra
C O
3 3
to
0) ©
<4- r—
"O 4-
m
•r- ffl
U CL
C 101
¦i— r—
•p-
S
o
X
o
CO
CSJ
o
O
x
*3*
00
o
O
-------
of H2S04 and NH4HS04 to the relatively inert salts.
The water solubility, hygroscopicity and/or deliques-
cence vary from extremely high (I^SO^) to quite low
(CaS04). The toxicity and type of physiological
effects expected due to inhaling these substances
also vary widely.
In maritime areas, the sulfate due to sea salt
(largely MgS04) usually dominates the supermicrometer
aerosol concentration, as shown by Junge and many
others. The ammonium compounds have been identified
3
by titration and x-ray diffraction by Brosset as the
dominant forms in southern Scandinavia. Cunningham^
finds the same result near Chicago by infrared absorp-
5
tion and x-ray diffraction methods. Charlson, et al.
found acid sulfates as well as (NH^gSC^ via optical
detection of deliquescence in nitu in and near St.
Louis with a humidograph.
The consistency of these qualitative molecular analy-
ses is noteworthy both because different schemes were
used for the analyses and because the same result was
achieved in widely differing locations. Perhaps the
most important result of all is that the acid sulfate
"family" (first four in Table 1) are often found in
air to have adequate purity to exhibit true molecular
nature. That is, the range of substances H^SC^ to
(NH4)2S04 are often sufficiently pure to have the same
hygroscopic/deliquescent response, infrared spectra
and x-ray diffraction patterns (ammonium compounds
only) as the pure substances. We do not suggest that
these compounds are pure, or that the traditional
Junge^ hypothesis of chemically mixed particles is
inappropriate; rather we find that at times the aero-
sol particles are sufficiently pure to show the
properties of the dominant sulfate species. Wc esti-
mate in the use of the humidograph, '^50% (mole) is
required to yield clear-cut deliquescence.
Hence it is clear that any monitoring or measurement
program must have as a basis some means to resolve the
molecular nature of the sulfates. None of us is a
health expert, but even cursory examination of hand-
books shows drastically different toxicities and
physioloqical effects of the different sulfates in
Table 1.
3. Particle Size Dependence
Of equal importance are the revelations by Whitby,®
that the particle size distribution by volume is
usually bimodal and that various sulfates appear in
different size classes, due largely to differences
in formation mechanism. Whitby,^ using data from
Patterson and Wagman (unpublished), finds the volume
distribution on sulfate to be nearly identical to the
log normal physical volume distribution, with a
geometric mean diameter by volume of 0.35 ym and a
geometric standard deviation of 1.9. This has been
verified by Dzubay and Stevens,7 showing that_the sul-
fates in and near St. Louis are likely to be in the
0.1 - 1.0 ym size class. Since these are probably
the first four molecular forms in Table 1, we thus
know that acid sulfates are in a respirable size
class and in the size class which most affects visi-
bility. We can subsequently conclude that sulfate
molecular composition should be size-resolved in some
manner, both to determine respirWiTTty~anTTo pro-
vide information on production mechanism.
4. Light Scattering as an Aid to
Monitoring Sulfates
Although there are promising techniques for sampling
and analysis of sulfate on time scales as short as
10-30 minutes (e.g. filters and wet chemistry as
described by Granat,° or flame emission as described
by Roberts®), it is useful to have a continuous record
or a means for interpolating between results from sam-
pling. For example, aircraft monitoring of sulfates
as performed in the OECO study of transport of sulfur
compounds required locating the aerosols prior to sam-
pling. In this study, Waggoner, et al JO showed that
in instances where sulfate concentrations are high,
they are often the dominant substance controlling
light scattering. Hence, there is a high correlation
between sulfate and nephelometric light scattering and
its instantaneous value may be used to infer local
values of sulfate. Figure 2 is a plot of sulfate and
light scattering data taken in central Sweden and in
the area of St. Louis. It is important to recognize
that the S02 derived sulfates are expected to be
present in the light scattering size range of 0.1 -
1.0 ym, so that the mechanically derived sulfates
would not be as influential or as well correlated with
light scattering.
5. Summary
Since the health effects of aerosols depend on
a) toxicity of the substance
b) particle size and location of deposition in the
pulmonary tract
c) particle growth with RH and its influence on depo-
sition
and since the toxicity and growth depend on molecular
composition as a function of particle size, as a
minimum, monitoring for health effects must include:
a) size resolution of some sort
b) molecular analysis.
The size resolution may be achieved with a size segre-
gating sampler such as was used by Dzubay and Stevens^,
by Lundgren'' or by Lee.^ The molecular analysis is
probably most easily achieved in the normal laboratory
by wet chemistry as described by Brosset,3 although
x-ray diffraction and/or infrared analysis are useful
if available. Humidography as described by Charlson,
et al.b is useful because it allows rapid (2-3 minutes)
resolution of the molecular form in situ, but this
apparatus is not readily available. Interpolation be-
tween sampling periods may be possible with the light
scattering coefficient from an integrating nephelome-
ter in those situations such as St. Louis or
Scandinavia where sulfate is correlated with and
dominated by S02 derived sulfate aerosols.
References
1. Shy, C. M. and Finklea, J. F., Environmental Sci.
and Techno!., 7, 204-208 (1973)^ "
2. Junge, C, E. Air Chemistry and Radioactivity,
Academic Press, New "York (1963"K
3. Brosset, C., Andreason, K. and Feron, M.,
Atmosphe ic Environment, 9, 631-642 (1975).
4. Cunningham, P. T., Johnson, S. A. and Yang, R. T.,
Environ. Sci. & Techno!., 8, 131-134.
2 12-6
-------
5. Charlson, R. J., Vanderpol, A.. H., Covert, D. S.,
Waggoner, A. P., and Ahlquist, N. C., Atmospheric
Environment, 8, 1257-1267 (1974).
6. Whitby, K. T., "Modeling of Multimodal Aerosol
Distributions," in Proceedings of the Geselschaft
fiir Aerosol Forsching, Bud Soden, FRG 17 October,
1974.
7. Dzubay, T. G. and Stevens, R. K., Environ. Sci. &
Techno!., 9, 664-668 (1975).
8. Granat, L. and Rodhe, H. Atmospheric Environment,
7, 781-792 (1973).
9. Roberts, P. Ph.D. Dissertation, California
Institute of Technology, Pasadena, CA (1974).
10. Waggoner, A. P..^Vanderpol, A. H., Charlson, R. J.,
Granat, L., Tragard, C and Laisen, S., Report AC-
33, Dept. of Meteorology, University of Stockholm,
Sweden; UDC 551.510.4:535.33 (1975).
11. Lundgren, D. A., Ph.D. Dissertation, University of
Minnesota, Minneapolis (1974).
12. Lee, R. E. Jr,, and Goranson, S., Environ. Sci. &
Techno!., 6, 1019 (1972).
Acknowledgment
The work in this paper was sponsored in part by grant
R0800665-11 from the EPA, grant GA27662 from the NSF
and by the International Meteorological Institute,
Stockholm University.
HgS,RSH,RSR
SOg, SO 3
tnyztn
H2S04, (NH4)zS04, NH4HSO4
(NH4)3H(S04)2, MgS04l CaS04
Na2S04
H+, NH4+, Na...
H , NH4
Na ...
HSO4 , S04
I
H* NH4+, Na+..
Hso' ,_A
SO3
H+, NH4+ A/
Na+...
SO*"
Figure 1, The tropospheric sulfur cycle. Rectangles
are recognizable entities in the atmosphere.
Triangles represent processes which have a
single direction of material flow, and
diamonds (two triangles) represent reversi-
ble processes, (a) Sources, (b) Sinks,
(c) Gas-to-particle conversion, (d) Sorp-
tion. (e) Deliquescence, (f) Efflores-
cence. (g) Raoult's equilibrium, (h) Reac-
tion in Concentrated Solution droplet,
i) Nucleation and condensation of water,
j) Evaporation, (k) Capture of aerosol by
cloud drops. (1) Reaction in dilute solu-
tion. (m) Rain, (n) Freezing of super-
cooled drop by ice nucleus. (0) Melting,
(p) Direct sublimation of ice on ice
nucleus, (q) Precipitation.
3
I2-6
-------
igm
100
-2
0.) gm
10
i
E
o>
O
in
Figure 2. Sulfate, [S04~] and light scattering (b )
data for Scandinavia (S) and St. Louis (T)
The two lines are for the given ratios of
[S04=] to bsp.
4
12-6
-------
THE NEED FOR INTERNATIONAL MODELS OF ENVIRONMENTAL STANDARDS
RAISAKU KIYOURA
Professor Emeritus, Tokyo Institute of Technology,
Research Institute o£ Environmental Science
5-Chome, 4 Kojimachi Chiyoda-ku, Tokyo, Japan
SUMMARY
The voluntary setting of environmental standards
among industrial nations has presented various
problems. The implementation plans for achieving
the goals and their time schedules have a significant
impact on economic activities. However, it is a fact
that better environmental standards can be introduced
only through the accumulation and re-evaluation of
past data of the world.
Above all else, agreement between the nations on
definitions and methods of measurement of pollution
is obviously desirable if the results of research work
in different countries are to be compared. For
instance, the Japanese NO2 Air Quality Standard was
established through the comparison and review of
domestic epidemiological and animal studies and of
87 documents among which 71 were from overseas.
Among various international problems concern-
ing air quality control, the Japanese NO2 Standard is
most disturbing: Whereas the standard in the U.S.A.
is 0.05 ppm/year, 0, 13 ppm/24 hrs (calculated
approximation); in Canada (Ontario), 0. 1 ppm/24 hrs
0.2 ppm/1 hr; in West Germany, 0.05 ppm for long-
term exposure, 0. 15 ppm for short-term exposure;
in Japan, it is 0.02 ppm/24 hrg--far more rigid than
in any other country.
Why these differences in standards despite the
fact that the human body is the same?
To make it possible to study health criteria and
control standards on a global basis, there is an
urgent need for international standardization of
monitoring methods, on which ISO is now working.
Activities of the Organization for Economic Cooper-
ation and Development (OECD) in this field--such as
the assessment of criteria in setting standards by the
Air Management Sector Group, the study of economic
implications of pollution control by the Economic
Policy Group, and the specification analysis of the
"Polluter Pays Principle" by the Environment
Committee-- should also be highly rated. The fruits
of these studies are certain to prove helpful to the
establishment of more discreet environmental quality
standards.
Available materials for this presentation were
so limited that the author hopes to accumulate more
through the assistance of other countries.
I. WHO CRITERIA AND GUIDES FOR AIR QUALITY
The World Health Organization (WHO), in a tech-
nical report^), sets forth the following levels of
atmospheric pollutants, which provide valuable
suggestions to all countries;
"Level I. Concentration and exposure time at
or below which, according to present knowledge,
neither direct effects (including alteration of reflex-
ed or of adaptive or protective reactions) have been
obs erved.
"Level II. Concentrations and exposure times
at and above which there is likely to be irritation of
the sensory organs, harmful effects on vegetation,
visibility reduction, or other adverse effects on the
environment.
"Level III. Concentrations and exposure times
at and above which there is likely to be impairment
of vital physiological functions or changes that may
lead to chronic diseases or shortening of life.
"Level IV. Concentrations and exposure times
at and above which there is likely to be acute illness
or death in susceptible groups of the population. "
Short- and long-term goals In another techni-
cal report (2), a WHO committee discussed the
philosophy for the interpretation of air quality criteria
for the purpose of developing standards and empha-
sized that standards, particularly short-term goals,
"may evolve differently in different countries depend-
ing on the exposure conditions, the socioeconomic
situation, and the importance of other health
problems. "
The situation is somewhat different for long-term
goals, the report said. "Without giving priority to
the adverse effects of air pollutants over other health
problems," the report noted, exposure to air pol-
lutants should be kept as low as possible, because the
subthreshold levels are not yet well defined and
"probably will not be defined with any great degree of
certainty for a long time to come. "
Since knowledge of the health effects of pollu-
tants decreases with their concentrations, the report
. 1 . 12-7
-------
said, any forecast of the possible effects of low levels
is bound to be speculative. By using the information
available, the Committee said, "it is possible to set
a level between these concentrations and the natural
background level that the Committee would like to see
adopted as an ultimate goal, with the hope that this
intermediate level would be unlikely to produce any
ill effects at all," The committee offered recom-
mendations as long-term goals intended to prevent
undersirable effects from air pollutants, emphasiz-
ing that these are tentative recommendations subject
to change as more data on dose-response relation-
ships within different populations become available.
II. ENVIRONMENTAL STANDARDS OF SOME
COUNTRIES
U.S. Guidelines
Air Quality Criteria According to the Guide-
lines for the Development of Air Quality Standards
and Implementation Plans published by the U.S.
Department of Health, Education and Welfare
(HEW),(3) air quality criteria documents summa-
rize the results of research on the adverse effects of
air pollutants. They are valuable in predicting
effects associated with exposure to various concen-
trations of air pollutants for various time periods,
HEW went on. They summarize what is known about
effects on health--with respect to both general and
specific populations--and effects on materials,
vegetation, visibility, and weather and climate. They
also summarize available data on economic losses
associated with air pollution. In addition, as far as
existing knowledge permits, the criteria reflect the
ways in which the effects of air pollutants may be
modified by interactions between pollutants and
between pollutants and physical factors in the environ-
ment. Thus, HEW states, air quality criteria are
"intended to provide a broad scientific basis for the
selection of air quality standards for the protection
of public health and welfare. "
Air Quality Standards In the same "Guidelines,"
HEW notes that air quality standards represent air
quality goals established with a view to protecting
public health and welfare. They provide a basis for
planning for the abatement and control of pollutant
emissions from existing sources and for preventing
urban and economic growth from adding to communi-
ty air pollution problems.
Later revisions are presented in "The Clean Air
Act" of the U.S. Environmental Protection Agency
(EPA). '4)
Meanwhile, in the Fourth Annual Report of the
Council on Environmental Quality,' ' Chairman R.E.
Train described his Council's belief that the U.S.
primary standard is the level of the particular pol-
lutant below which, based on current information,
human health is believed to be adequately protected.
Time Schedule The HEW "Guidelines"^3) states
that: "Establishment of air quality standards as
goals connotes the fixing of at least an approximate
time schedule for progress toward their attainment. "
Where this is difficult, HEW suggests, it maybe
desirable to adopt interim air quality objectives to
be achieved in relatively short time periods, such as
two or three years, adding that such interim objec-
tives provides for significant increments of progress
toward attainment of the standard through orderly
application of the best available techniques for the
prevention, control, and abatement of air pollution.
As for the time it will take to attain an air
quality standard or interim objective, HEW said it
would depend largely on the nature of the implemen-
tation plan--or the steps to be taken to abate and
control pollution.
In respect of the attainment of the goals, EPA
Administrator R. E. Train in a press conference^)
gave a progress report on the Clean Air Act, which
set forth health-based air quality standards to be
achieved by May 31, 1975, although extensions up to
mid-1977 were permitted for cases where technology
or other alternatives were not available earlier.
Salient points of his remarks were as follows:
The U.S. made significant headway in cleaning
up the air, but there was still a long way to go. If
all concerned worked hard together, health protection
goals established in the Clean Air Act could be attain-
ed. No one back in December 1970 imagined it would
be easy to achieve clean air. However, many under-
estimated the complexities involved, and few foresaw
the worldwide energy crisis and economic recession.
"Nitrogen oxides, a pollutant related about equal-
ly to mobile and stationary sources," Administrator
Train commented, "is a problem now only in certain
areas."
Canadian National Air Quality Objectives
A broad range of comments has been received
the Department of Environment, Canada, says, (7)
from the public, industry and environmental groups
since the air quality objectives were first proposed
in November, 1971. The views expressed were
carefully considered and prompted some revision of
the original proposals and the addition of standard
reference methods for measuring each pollutant.
Canada's Federal Clean Air Act was said to be
unique as it called for three levels of air quality
objectives--"desirable, " "acceptable," and "toler-
able"--of which only the first two were officially
introduced. Under the Clean Air Act, the National
Air Quality Objectives are designed to protect
public health and welfare by setting the following
limits on levels of air pollution:
The maximum acceptable level is intended to
provide adequate protection against effects on soil,
water, vegetation, materials, animals, visibility,
personal comfort and well-being. It represents the
realistic objective today for all parts of Canada.
When this level is exceeded, control action by a
regulatory agency is indicated.
The maximum desirable level defines the long-
12-7
-------
term goal for air quality and provides a basiB for an
anti-degradation policy for the un-polluted part of
the country and for the continuing development of
control technolcgy. At this level there would be no
detectable adverse biological effect on any receptor.
Maximum tolerable levels are intended to indi-
cate the onset of an "imminent danger" requiring
immediate abatement action. Air pollution episodes
which sometimes result when pollutants accumulate
during adverse weather conditions would fall within
this category.
West German Air Quality Standards
In West Germany, the Maximale Immissions-
werte (MIK) was revised by VDI (Verein Deutscher
Ingenieure) in September 1974. Then, on the basis
of this MIK, the Federal Government published TA-
LUFT (Technical Instruction for Maintaining Air
Purity)^) in the same month, pursuant to No. 48 of
the Federal Emissions Control Act of March 1974.
Japanese Air Pollution Control Measures
(9)
Establishment of Environmental Standards
As stipulated in Article 9 of the Basic Law for
Environmental Pollution Control established in 1971,
"environmental standards" are standards for environ-
mental conditions of air, water, soil and noise,
maintenance of which is desirable for the protection
of human health and preservation of the living
environment. They aim at effecting measures for
air pollution control such as individual emission
standards in a rational way. They attempt to improve
the environment where pollution from individual
sources has accumulated. Thus, environmental
standards should serve as the basis or objective of
pollution control measures such as formulation of
pollution control programs, statutory regulation
under the Air Pollution Control Law, Water Pol-
lution Control Law and Noise Regulation Law,
improvement of land use and upgrading of control
facilities.
Accordingly, environmental standards are a
"desirable level to be maintained," rather than the
maximum tolerable concentrations or acceptable
levels. As science and technology continue to
develop, care must be taken to expand the scope of
proper scientific judgment so that environmental
standards may always encompass new data on the
effects on the human health and living environment
and newly introduced pollutants. Accordingly,
standards must undergo scientific review and need
revision.
EI. JAPANESE NQ2 STANDARD
Air Quality Standards of the U.S. , West Germany
and Japan do not greatly differ from one another
except for those for N02 (See the Tables).
As for N02 standards, however, it can be found
that the Japanese value is significantly stringent:
five to seven times stricter than the equivalent values
of other countries. The Japanese N02 standard level
is exceeded by measured levels at 224 (98%) out of
the 228 domestic monitoring stations. On the other
hand, 98% of the Japanese stations will be able to
meet the equivalent standards of the U.S. and West
Germany.
Moreover, the Japanese standard, which provides
for an interval of 5 to 8 years for the attainment of
the standard, will be inconsistent with the OECD
GUIDING PRINCIPLES which will be referred to
later in this paper.
Domestic industries and scientists have severe-
ly criticized the standard, claiming that the establish-
ment of the standard has its basis on scientific
uncertainty. The epidemiological study of 6 com-.j.
munities described in an expert committee report '
was held to be erroneous. The present author feels
that the Japanese environment authorities responsi-
ble for the absurdly rigid NC^ standard refuse to
admit their error simply because they insist on
"saving face." By so doing, they should realize,
they are causing confusion in various degrees, at
home and abroad. In the 6 region housewife survey
cited in the report, pollutant concentrations were
determined during a three month period, in the
winter of 1970-71; measurements were obtained for
8-72 hours during each month, for a maximum of 9
days during the 90-day study period. The following
respiratory Symptom rates and pollutant concen-
trations were reported in the survey report (See
Table below and Fig. 1 on page 13):
The monitoring time and frequency are clearly
inadequate. Above all else, the number of the
sampling sites is unlikely to provide proper sta-
tistical evaluation. Also, the statistical processing
of the data is unconvincing.
The report gives the following formula and its
correlation coefficient in an attempt to support the
claim of high correlation between the chronic
bronchitis (persistent cough and phlegm) prevalence
rates of 6 communities and the concentration of NC>2:
(Y = Prevalence %)
Y = 54. 16x + 1.67
r = 0.7095 (p = 0.05)
But a proof test done by the critics indicates,
§ = n - 2 = 6 - 2 = 4,
r(4, 0.05) = 0.812, r(4, 0. 10) = 0.734
Therefore, no correlation between them can be
recognized at 5 and 10% significance levels.
In addition, the report introduced the Fisher's
Exact Test (p = 0.05) but it is clearly inappropriate
to apply such a method to this statistical analysis
because of the very few samples. Therefore, the
conclusion, based on statistical epidemiology, of
the report cannot be accepted as rational.
In the U.S. Senate Public Works Committee, R.
Nader's Group asked to have the Government
response to the Japanese N02 standard' upon
which Senator E.S. Muskie, Committee Chairman,
required the Environmental Protection Agency (EPA)
to comment on the standard.
- 3 -
12-7
-------
Location
Sakura,
Ichihara,
Tonda-
Fuse,
Fukuoka
Ohmuta,
Chiba
Chiba
bayashi,
Osaka
West,
Kyushu
Prefecture
Prefecture
Osaka
Prefecture
Kyushu
Prefecture
Cough and
phlegm (%)
Cough (%)
Phlegm (%)
NO 2 ppm
S02 ppm
TSP ug/m3
1.9
3. 5
4. 1
.015
.024
196
1.9
4. 8
5. 3
.013
.027
352
1.2
3.0
4.3
.017
.013
111
5. 6
7. 9
11.0
.077
. 050
350
4. I
5.8
12. 3
.042
.010
183
5. 3
9.4
11.0
.020
.042
498
FIG. 1. CORELATION BETWEEN PREVALENCE
OF PERSISTENT COUGH & PHLEGM AND N02
IN 6 COMMUNITIES (1970, JAPAN)
sS
pu
-a
o
+->
G
4)
M
•
V)
f-l
In respect of the medical study on which the
standard is based, Prof. C. M. Shy, M. D. , of
North Carolina University, former Director of EPA
Human Studies Laboratory, made the following
comments and presented a report"2' to the Senate
Committee:
"... The value of the Japanese studies lies in
their qualitative confirmation of experimental results
in animals, namely, that N02 exposure can produce
lung tissue changes similar to that found in cases of
human chronic respiratory disease. Heretofore, no
human evidence on chronic bronchitis supported the
experimental findings. ., As I interpret the Japanese
results, excess chronic bronchitis would be more
attributable to high concentrations of suspended
particulates, which ranged from 111 to 498 ^ig/m3
(U.S. primary standard is 75 jig/m ), than to levels
of N02 which ranged from .013 to .077 ppm (U.S.
primary standard is .053 ppm). Further, the
chronic disease excess should be attributed to long-
term averaging times of one year or more rather
than to a 24-hour maximum concentration. For
these reasons, I cannot accept the rationale for the
Japanese N02 standard of ,02 ppm 24-hour maxi-
mum. "
Many Japanese people expect the Government to
review the standard and modify it to a more reason-
able one.
IV- OECP RECOMMENDATIONS
On August 6, 1975, the U.S. and Japan signed
an agreement to strengthen cooperation in the field
of environmental protection. Similar agreements had
reportedly been concluded between the U.S. and
U.S.S.R. , and the U.S. and West Germany.
One of the items of the U.S. - Japan agreement
calls for: Making activities under the agreement
subject to budgetary appropriations and laws of the
two countries and having the "Guiding principles
concerning international economic aspects of environ-
mental policies" adopted in 1972 by the Council of the
Organization for Economic Cooperation and Develop-
ment serve as a basis of such activities.
The OECD "guiding principles"^^ referred to
above sets forth: "In many circumstances, in order
to ensure that the environment is in an acceptable
state, the reduction of pollution beyond a certain
level will not be practical or even necessary in view
of the costs involved. "
(14)
The OECD "Economic Implications" recom-
mends the following: "...Typically, additional gross
investment requirements may, on the average, come
to around 0. 5 per cent of GNP over the period con-
6
i
i
5
>< Ohmuta
A
' Fukuoka W
< 2^*
Fuse
2
1
a
U
_ £ */ Sakura
u 1 V =
_ ^ i
Tondabayashi r =
54. l6x + 1. 67
0.7095 ( p< 0.0
* i t
0.02 0.04
0.0 6
0.0 B
NOz (ppm)
0.019
3% ! 0
1
i ^
!
i 3
0
Fisher's exact test
(p = 0.05)
- 4 -
12-7
-------
sidered. The one important exception is the much
larger (yet still incomplete) Japanese figure of 2. 6
per cent. Other resource requirements--i. e. inter-
mediate inputs--may be rather less than 0. 5 per cent
in the first half of the period but could move up to-
wards something nearer to 1 per cent after 1975.
Annualised costs of these programmes by 1975 will
remain well below 1 per cent of GNP for all countries
except Japan, where they may be substantially larger.
By 1980 annualised costs may have risen to between
1 and 1. 5 per cent for several countries and to more
than 2 per cent for Japan. . . . The most important
item in total expenditure on pollution control from
the point of view of timing is investment. ... If
competitive demands for investment arise, short-run
inflationary pressures may build up which could
require policy offsets by government to ensure ade-
quate leeway for pollution control investments. . . . "
p. 462.
12. Ibid. p. 464
13. "The Polluter Pays Principle." Organization
for Economic Co-operation and Development,
Paris, 1975.
14. "Economic Implications of Pollution Control: A
General Assessment." OECD, Studies in Re-
source Allocation No. 1, 1974.
REFERENCES
1. "Atmospheric Pollutants." World Health
Organization, Geneva, Switzerland, Technical
Report Series No. 271, 1964.
2. "Air Quality Criteria and Guides for Urban
Air Pollutants." WHO, Technical Report
Series No. 506, 1972.
3. "Guidelines for the Development of Air Quality
Standards and Implementation Plans." U.S.
Department of Health, Education, and Welfare,
Washington, D.C., U.S.A., 1969.
4. "The Clean Air Act. " U.S. Environmental
Protection Agency, December 1970.
5. "Environmental Quality". 4th Annual Report
of the Council on Environmental Quality, 1973,
p. 271.
6. Train, Russell E. "Opening Statement of
Press Conference on Air Quality Progress."
U.S. Environmental Protection Agency, May
30, 1975.
7. "National Air Quality Objectives Announced."
Environment Canada, Ottawa, Canada, January
3, 1973.
8. "Technische Anleitung Zur Reinhaltung Der
Luit - TA LUFT - West Germany, 1974.
9. "Quality of the Environment in Japan."
Environment Agency, Japan, 1972, p. 63.
10. "Report of the Expert Committee on Air
Quality Criteria for Oxides of Nitrogen and
Photochemical Oxidants (Japan)." Central
Council for Control of Environmental Pol-
lution, Sub Council for Air Pollution Control,
June 20, 1972.
11. "Hearings before the Committee on Public
Works United States Senate, 93rd Congress,
Serial No. 93-H23, November 5 and 6, 1973,"
- 5 -
12-7
-------
SUMMARY OF NATIONAL AMBIENT AIR QUALIT Y STANDARDS (U.S. A) Curtesy of U.S. EPA
FEDERAL
AVERAGING PRIMARY SECONDARY REFERENCE
POLLUTANT TIME STANDARDS STANDARDS
METHOD (FRM)
PARTICULATE Annual 75 ug/m 60 ug/m Hi-Volume
MATTER (Geometric Mean) ? Sampler
(1) 24 - I (our* 260 ug/m 150 ug/m"
SULFUR Annual 80 ug/m
OXIDES (Arithmetic Mean)
24 - Hour-" 365 ug/m (0. 14ppm)
Pararosaniline
3 - Hour - 1300 ug/m ^(0. 5ppm)
CO
8 - Hour*
1 - Hour*
10
40
. 3
mg/m (9ppm)
mg/m (3 5ppm)
(Same as
Primary)
Non - Dispersive
Infrared
Spectrometry
N02 (2)
Annual
(Arithmetic Mean)
100
, 3
ug/m (0.0 5ppm)
(Same as
Primary)
Jacobs -
Hochheiser
(Rescinded)
PHOTOCHEMICAL
1 - Hour*
160
3
u£/m (0.08ppm)
(Same as
Primary)
Chemilumines-
OXIDANTS (3)
cence
HYDROCARBONS
3 - Hour*
160
3
ug/m (0, 24ppm)
(Same as
Primary)
Flame
(non-Methane) (4)
(6 to 9 a. m.)
Ionization
Not to be exceeded more than once per year.
NOTE: The air quality standards and a description of the reference methods were published on
April 30, 1971 in 42 CFR 410, recodified to 40 CFR 50 on November 25, 1972.
January 30, 1974 - JDC
Foot Note: COMMENTS (1) The Secondary annual standard (60 ug/m ) is a guide for assessing
SIPs to achieve the 24-hour secondary standard.
(2) The continuous Saltzman, Sodium Arsenite (Christie), TGS, and
Chemiluminescence have been proposed as replacements for the
J-H method. New FRM to be decided upon by Jan. 1975.
(3) The FRM measures 0^ (ozone)
(4) The HC standard is a guide to devising SIPs to achieve the Oxidant
standard.
The HC standard does not have to be met if the oxidant standard
is met.
- 6 -
12-7
-------
CANADA: NATIONAL AIR QUALITY OBJECTIVES
Maximum Acceptable Levels
Maximum Desirable Levels
Sulphur Dioxide
(0.02 ppm) annual arithmetic mean.
(0. 11 ppm) as a maximum 24 hour
concentration. (0,34 ppm) as a
maximum one hour.
(0.01 ppm) annual arithmetic mean.
(0.06 ppm) as a maximum 24 hour
concentration.
(0.17 ppm) as a maximum one hour
concentration.
Particulate matter
3
70 ug/m annual geometric mean.
120 ug/mS as a maximum 24 hour
concentration
60 ug/m^ annual geometric mean.
Carbon Monoxide
(13 ppm) as a maximum eight hour
concentration.
(30 ppm) as a maximum one hour
(5 ppm) as a maximum eight hour
concentration.
(13 ppm) as a maximum one hour
concentration.
Total Oxidants
(0.015 ppm) annual arithmetic mean.
(0.025 ppm) as a maximum 24 hour
concentration.
(0.08 ppm) as a maximum one hour
(0.015 ppm) as a maximum 24 hour
concentration.
(0.05 ppm) as a maximum one hour
concentration.
Nitrogen Dioxide (0.20 ppm) 1 hour
(Ontario) 1974 (0.10 ppm) 24 hour
WEST GERMANY:
AIR QUALITY STANDARDS-TA
LUFT-Sept., 1974
2.4 Immission Values
2.4.1 General
Immission values shall be the values for long-
term exposure (IW 1) short-term exposure
(IW 2) established in 2 . 4. 2, 2.4.3. Dusts with
a particle size less than 10 um and sulfur di-
oxide, whose simultaneous occurrence must be
taken into consideration, the above shall refer
to the lone action o£ the particular air pol-
lutants .
2.4.2 Immission Values for Dusts
2.4.2.1 Dust Precipitation
The following immission value is established
for nonhazardous dust precipitation: _
IW 1 0. 35 g/(m2d), IW 2 0.65 g/(md)
2.4.2.2 Dust Concentration in the Air
The following immission value is established
for the mass concentration of nonhazardous
dusts with a particle size of less than 10 um
in the air: ,
IW 1 O.lOmg/m , IW 2 0.20 mg/m
The following immission value shall apply-
when measurement methods are used with
which dusts having a particle size greater than
10 um are also included:
IW 1 0.20 mg/m"', IW 2 0.40 mg/m
2.4.3 Immission Values for Gases
The following immission values are establish-
ed for individual gaseous immissions in the air:
Type of Immission
Mass Concentration
3 3
mg/m , IW 1 mg/m ,
IW 2
Chlorine 0. 10
Hydrogen chloride 0. 10
-- given as inorganic
gaseous chlorine
compounds
Hydrogen fluoride 0.0020
-- given as inorganic
gaseous fluorine
compounds
Carbon monoxide 10.0
Sulfur dioxide 0. 140
Hydrogen sulfide 0.0050
Nitrogen dioxide 0. 10
Nitrogen monoxide 0.20
0. 30
0. 20
0.0040
30.0
0.40
0. 010
0. 30
0.60
3
- 7 -
12-7
-------
JAPAN: AIR QUALITY STANDARDS
POLLUTANT
ENVIRONMENTAL CONDITION
MEASURING METHOD
SULPHUR
DIOXIDE
0.04 ppm as a maximum 24 hour
concentration.
0. 1 ppm as a maximum 1 hour.
Electro- Conductimetry
CARBON
MONOXIDE
10 ppm as a maximum 24 hour
concentration.
20 ppm as 8 hour mean.
Non-Dispersive Infrared
Spectrometry
PARTICULATE
MATTER
0.10 mg/m as a maximum 24 hour
concentration.
0.20 mg/m3 as a maximum 1 hour.
NITROGEN 0.02 ppm as a maximum 24 hour
concentration. Saltzman reagent
PHOTOCHEMICAL 0.06 ppm as a maximum 24 hour
OXIDANTS concentration.
Neutral KI solution or
Chemilumines cense
NOz LEVELS OF NON URBAN COMMUNITY, JAPAN (1974 March) (24 hra Mean)
Community
Population
NOz (ppm)
1
Atuma (HOKKAIDO)
2,645
0. 005
2
Wakayanagi (MIYAGI)
17,617
0.011
3
Kasama (IBARAGI)
30,696
0.025
4
Matsushiro (NAGANO)
20,000
0.011
5
Asuke (AICHI)
11,422
0,013
6
Shinoyama (HYOGO)
14,146
0. 014
7
Shobara (HIROSHIMA)
6,622
0.014
8
Tsuda (KAGAWA)
9,940
0.011
9
Dazaifu (FUKUOKA)
31,800
0.021
Mean
16,099
0.013
(1973)
NOz LEVELS OF URBAN AREA
Tokyo Metropolitan Area 11,458,099
Tokyo City Area 8,796,293
0.035 (Annual
Mean)
0.038 (Annual
Mean)
- 8 -
12-7
-------
MACRO AND MICRO APPROACHES TO Tin: DETERMINATION
OF PESTICIDE RESIDUES IN HUMAN AND ANIMAL TISSUES
Robert. E. Moseman
Environmental Protection Agency
Environmental Toxicology Division
Research Triangle Park, N.C. 27711
Summary
Analytical approaches to the determination of
pesticides and metabolites in human and animal tissues
will take many forms. Several factors must be consid-
ered in choosing an analytical scheme if the results
are to be meaningful. Whenever possible the residue
chemist will use standardized methodology which has
been subjected to evaluation by collaborative studies.
In the absence of such methods, the analyst must be
aware of the capabilities and limitations of the vari-
ous steps in the analytical procedure of choice. Ex-
traction and cleanup efficiency must be considered in
view of the determinative method to be used. Highly
efficient gas chromatographic columns coupled with
sensitive and selective detector systems are powerful
analytical tools. One must be cautious, however, when
relying on gas chromatographic peak retention time for
compound identification. For a complete analysis,
confirmatory techniques should be used. The available
sample size may dictate the use of a micro method of
analysis and, at the same time, preclude certain con-
firmatory techniques. Advantages of micro methods
include, speed, smaller glassware and minimum solvent
volumes.
Introduction
The determination of pesticide residues in human
and animal tissues has been of concern to analytical
chemists for many years. A wide variety of factors
will influence the approach used in any particular
analysis. Some history of the sample can be quite
valuable. What compounds can be expected in the sam-
ple? Is it reasonable to look for other compounds?
Should one analyze for the parent compound or is there
a known metabolite that will be present instead? For
example, dieldrin and heptachlor epoxide will be found
in general population human adipose tissue as metabo-
lites of aldrin and heptachlor. What are the expected
levels of pesticides and/or metabolites? Is the se-
lected methodology capable of the desired lower limit
of detection with the size of sample available? Will
the compounds of interest be quantitatively recovered
with the analytical methodology used? What kind of
confirmation is required? Is the sensitivity of the
confirmatory technique commensurate with the sample
size and/or cleanup procedures used? These are some of
the questions that must be considered when approaching
the determination of pesticides or metabolites in
tissue samples.
The entire sample preparation procedure, be it
macro or micro, must be considered when dealing with
pesticide residue analysis. If losses of the compounds
of interest occur at any step, the results of the
analysis become questionable.
Extraction procedures for several grams of tissue
generally call for blending with an organic solvent or
grinding in a mortar with anhydrous sodium sulfate.
The tissue may be minced and ground with sodium sulfate
prior to Soxhlet extraction. Milligram quantities of
tissue can be conveniently macerated in a small glass
tissue grinder with an organic solvent. The resulting
homogenate can be centrifuged and the solvent decanted.
This process requires a minimum of glassware and sol-
vent transfer, thus reducing chances of loss and back-
ground contamination.
The cleanup required for sample extracts will vary
from minimal to most elaborate depending on the sample
coextractives, and the desired level of detectability,
sensitivity and selectivity of final detection systems.
Sample cleanup in some cases provides preliminary sepa-
ration of pesticides into groups thus enhancing the
reliability of correct identification. Solvent parti-
tioning is helpful in separating lipids and waxes from
non-polar pesticides but some pesticides may be par-
tially lost at this step.
Adsorption column cleanup is widely used in con-
junction with solvent partitioning. Adsorbents most
frequently used include Florisil, silica gel, alumina
and charcoal. An endless variety of solvents and
solvent mixtures are applicable for the particular
cleanup and/or separation desired. Exclusion chroma-
tography or gel permeation chromatography has recently
gained in popularity. Thin-layer or channel-layer
chromatography may be substituted for column chroma-
tography. This technique can provide good cleanup
and separation with an absolute minimum of solvent
interference at the determinative step. Acid or base
partitioning may be used for cleanup of a limited num-
ber of the more stable pesticides and polychlorinated
biphenyls (1'CB's).
Depending upon the determinative step to be used,
derivatization may be required to render the compounds
of interest suitable for analysis. For gas chromatog-
raphy, polar compounds may be converted to more vola-
tile esters, ethers or amides prior to analysis. De-
rivatization may also be used to introduce heteroatoms
into molecules thus making them responsive to element
specific detectors. Fluorescent derivatives can be
prepared which wi11 be detected in the presence of
large amounts of non-fluorescent background materials.
The final determination of pesticides extracted
from tissues may take many forms. The most useful
technique used today is probably gas chromatography
(GC) interfaced with highly sensitive and selective
detector systems such as electron capture, flame photo-
metric and electrolytic conductivity. Alkali flame,
microcoulometric and flame ionization detectors are
also used. Several widely used GC columns are avail-
able which have been evaluated extensively. The
literature abounds with relative retention data for
hundreds of pesticides on these columns under vigor-
ously standardized conditions. In many instances,
tentative identification of unknowns may be made on the
basis of relative retention alone. This does not
suggest that the analyst should rely on only relative
retention data, but that his choice of pesticide
standards to try can be greatly narrowed.
Thin-layer chromatography (TEC) finds use in resi-
due analysis for quantitation and confirmation of
results obtained by other means. Recent developments
in scanning techniques show promise for improving the
quantitative aspects of TLC.
1
13-1
-------
Colorimetric and paper chromatographic techniques
are still in use, but suffer from lack of sensitivity
and specificity.
Confirmatory techniques for suspected pesticides
or metabolites in tissues range from the quick and sim-
ple p-value to the most elaborate and sophisticated
instrumental analyses. In laboratories where expensive
equipment is not available, there are alternate means
for confirmation of residues that are of reasonable
reliability. One such means is chemical derivatization
followed by gas chromatography or other suitable quali-
tative techniques. Where sufficient residue is pre-
sent, TLC can be used effectively. P-values yield
limited information, but in the absence of other con-
firmatory means, some confidence can be gained. In
addition, p-values can be obtained on as little as 100
microliters of sample extract.
Confirmation by instrumental methods has proved to
be quite valuable. Infrared spectra matching provides
very good evidence where sufficient compound is pre-
sent. Mass spectra, especially when obtained in con-
junction with gas chromatography, is an excellent means
of confirmation.
Macro Methods
Historically, pesticide residues were determined
by colormetric techniques after extraction and occa-
sional cleanup. Paper and thin-layer chromatographic
methods were, and still are, used for multiple resi-
dues. These methods, though suitable for many applica-
tions, suffer from lack of sensitivity and specificity.
Fairly large sample sizes are usually required and
rigorous cleanup is needed. With the advent of gas
chromatography and the associated selective and sensi-
tive detector systems, a wider range of analyses and
lower limits of detection of pesticide residues became
possible. Many of the extraction and cleanup proce-
dures were directly applicable to the new GC methods.
The multiresidue procedure outlined in the FDA
Pesticide Analytical Manual''' is probably one of the
most widely used methods for pesticide residues. This
2
procedure is based on the original work by Mills , and
Mills et al^, and has evolved over the years to include
a sizeable number of pesticides and metabolites in a
multitude of sample types. The essentials of extrac-
tion with an organic solvent, partitioning between two
immiscible solvents and column chromatography on Floi'i-
sil remain an effective means of sample preparation.
Modifications of this methodology have been applied to
almost every conceivable substrate. In general, this
methodology has been used on a macro scale (greater
than one gram of sample) but micro modifications have
been reported. Alternate elution solvents have been
used to improve upon the lipid/pesticide separations,^
The Environmental Toxicology Division of EPA has
adapted the Mills method with minor modifications and
included it in their Manual of Analytical Methods^.
This modification is used in the Human Tissue Monitor-
ing Program for the determination of pesticide residues
in adipose tissue. The methodology has been subjected
to collaborative study, and in conjunction with stand-
ard GC columns and instrumented parameters, has pro-
duced very acceptable interlaboratory precision and
accuracy.
Also included in the EPA Manual is a micro modifi-
cation which can be used for tissue samples of a few
hundred milligrams.^ In many instances pesticide de-
terminations are required for biopsy samples or for
2
small laboratory animals where most of the sample has
been used for other determinations.
Some success has been reported with a sweep
codistillation procedure for the cleanup of oils and
fats. The sample extract in an organic solvent is in-
jected into a heated tube and the pesticides are
flushed through with additional solvent leaving the
coextracted interferences behind. A micro Florisil
column is required for the determination of chlorinated
hydrocarbons. It was reported that sample preparation
time could be reduced from 2-4 hours to 50-40 minutes.
Up to one gram of oil or fat could be cleaned up.
Gel permeation chromatography has been used as a
cleanup technique for the determination of chlorinated
O
hydrocarbon pesticides in fish tissue. This system
9 in
was subsequently automated and evaluated. Twenty-
three sample extracts can be loaded into the commer-
cially produced unit and the cleanup is subsequently
accomplished unattended. Due to some carryover of
lipid into the pesticide fraction O.lppm of chlori-
nated hydrocarbon in one gram of fats and oils is con-
sidered a practical lower limit without additional
cleanup. With this system the potential exists for
various column packings and eluting solvents. As
such, the cleanup would probably be considered to be
on the macro scale. Use of other adsorbents could
most likely allow handling of micro samples. Presently
the large column bed of gel (50gm) will handle up to
about one gram of lipid. Simply scaling down the size
of the cleanup column would allow as little as 50 to
lOOmg of sample to be cleaned up but the minimum prac-
tical level of determination would still be O.lppm.
Combined with other cleanup techniques, this system
has been applied to the separation of ultra trace
quantities of dioxin from lipids.11
PCB's were first recognized as environmental con-
taminants in the mid 1960's. Since that time metho-
dology has been developed to chromatographically
separate PCB's from pesticides thus allowing more
12 1 ^
confident identification. ' Variations of this
procedure have been used in conjunction with many
different multiresidue procedures. Confirmatory tech-
niques such as GC-MS and chemical derivatization have
proved to be valuable.
Micro Methods
When the residue chemist is confronted with very
small samples such as tissue biopsies the analytical
approach must be quite different than that used when
unlimited sample is available. In such instances the
entire sample may be required, thus precluding the
possibility of a "rerun". Reliable and accurate
methodology is a prerequisite. One must consider that
lOOmg of sample containing a few nanograms of pesticide
or metabolite can easily get lost on the surfaces of
large glassware and column adsorbents. Concentrating
several hundred ml of eluting solvent to a very small
volume so that the desired amount of pesticide can
be introduced into the analytical system can obviously
create interference problems.
Numerous micro methods have been developed for
application to the cleanup and analysis of tissue for
pesticide residues. Adsorbents used for columns range
the entire gamut as do elution solvents. In many cases
the investigator is attempting to clean up and separate
multiple pesticide residues in one step. Adsorption
column chromatography is a valuable asset for compounds
which are difficult to separate by GC. In addition,
13-1
-------
selective elution of pesticides adds confidence to the
identification of suspected pesticides.
ICadou^"''"^'''"^'''"^ devised separation schemes for
several chlorinated hydrocarbon and organophosphate
pesticides on silica gel. In this series of papers he
investigated various solvent systems in attempts to
clean up plant, animal, soil and water extracts. As
with most micro methods, more samples could be handled
in a shorter period of time and solvent volumes were
kept to a minimum.
18
Leoni used a silica gel system similar to that
reported by Kadoum, and demonstrated the cleanup and
separation of fifty pesticides into four groups. His
interest was primarily in the analysis of water.
19
Holden and Marsden used alumina and silica gel
columns in sequence to clean up animal tissues for
residue determinations. The alumina was more effective
in removing lipid from the extract While silica gel
was useful for separating the pesticides into groups.
One hundred milligrams of tissue equivalent was chro-
matographed, thus lending this technique to micro
analysis.
20
Law and Goerlitz evaluated micro columns of
alumina, silica gel and Florisil for the cleanup and
group separation of pesticides in water. The micro
Florisil column was prepared in a disposable pipet and
the eluting solvents used were 6% and 15% ethyl ether
in hexane, This system is essentially a scaled down
2
Mills Florisil column. These authors noted several
advantages of micro columns including speed, conven-
ience and recovery.
21
Johnson used a one gram column of silica gel,
deactivated with 1% water, to clean up water sample
extracts. Chlorinated pesticides were separated into
two fractions before quantitation by electron capture
GC.
22
Recently, Erney reported on the use of miniature
Florisil columns for the separation and cleanup of
chlorinated pesticides. The column and eluting sol-
vents were the same as those used in the official FDA
procedure but the size was reduced by about one fifth.
Separations and recoveries were found to be comparable
to that obtained with the larger column.
23
Erney also described a miniature silica gel
column that was used for the separation of some PCB's,
DDT and analogs. Five grams of silica gel was slurry
packed in a 1cm i.d. column and eluted with hexane
and methylene chloride.
A micro silica gel column separation scheme was
developed by Sherma and Shafik^ and used in conjunc-
tion with air sampling for pesticides. Group separa-
tion was obtained by collecting three fractions of
15ml or less from a lgm silica gel column. Chlori-
nated hydrocarbon, organophosphate and carbamate
pesticides were quantitated. The carbamates were
derivatized with pentafluoropropionic anhydride prior
to electron capture gas chromatography. Adaptation
of this methodology for the determination of residues
in other sample types should be possible.
Enos et al^ developed a micro method for the
determination of pesticide residues in several differ-
ent tissues. Samples of 500mg or less are extracted
with acetonitrile, diluted with aqueous salt solution
and partitioned against hexane. The hexane layer is
concentrated to a small volume and transfered to a
column of 1.6gm of Florisil. The column is eluted
successively with hexane followed by 1% methanol in
hexane. The two fractions collected are concentrated
to a suitable volume and analyzed by gas chromatog-
raphy on standard columns using an electron capture
or flame photometric detector. A modification of this
methodology for pesticides in human milk is included
in the EPA Analytical Manual^.
Applications
The recent interest in hexachlorobenzene (HCB]
prompted us to look for ways to improve the recovery
of this compound from adipose tissue. Both the
Mills procedure and the micro modification as outlined
in the EPA Analytical Manual yielded about 60% recov-
ery. With the Mills modification, losses of HCB oc-
curred at the acetonitrile/petroleum ether partitioning
step. The low recovery with the micro procedure was
traced to the acetonitrile extraction step.
The recovery of HCB in the micro method was im-
proved to 801 or better by siirply changing the extrac-
tion solvent to 20% acetone in acetonitrile. Recovery
of other commonly found chlorinated pesticides was not
altered.
Attempts to improve the recovery of HCB with the
Mills procedure were not successful. However, a simple
and quick method was worked out which yielded close to
100% recovery. Fat samples of 500mg or less were
extracted with hexane and transfered directly to a
standard 100 x 22mm Florisil column. Hexane (100ml)
was used as the eluting solvent and yielded quantita-
tive recovery at HCB levels as low as 5ppb with a
minimum of interference. Incomplete recovery of most
other chlorinated pesticides was noted, and as a re-
sult, this procedure could not be used as a multi-
residue method. This technique would, however, be
useful in conjunction with the Mills method by reserv-
ing 0,5gm equivalent of the original sample extract
and carrying the major portion (2.5-3gm) through the
standard multiresidue routine. If HCB were suspected
in the sample, the "direct elution" method could be
used for the remaining extract for more reliable
quantitation of HCB.
A rapid and simple technique for the confirmation
of HCB in adipose tissue was also developed. Mono-
and disubstituted ether derivatives of HCB were formed
by reaction with KOH in various alcohols. Quantita-
tive and qualitative data for HCB derivatives were
obtained on standard GC columns coupled with an elec-
tron capture detector.
During the course of a rat feeding study, the
need arose to determine residues of the fungicide
dichloran (2,6-dichloro-4-nitroaniline) in various
27
tissues and excreta. A micro procedure was required
because of limited sizes of some samples such as eye
tissues. Only slight modification of the micro pro-
cedure of Enos et al^ was required to obtain accep-
table recovery. Because of the polar nature of the
compound, it was necessary to reduce the volume of
the acetonitrile extracting solvent before partition-
ing the dichloran into hexane. It was also found that
a larger volume of 1% methanol in hexane was required
for the second fraction from the Florisil column.
Several standard GC columns were found to be acceptable
for quantitation with an electron capture detector.
The average overall recovery of the method for eight
tissues and excreta was greater than 90%. Liver and
brain were the most troublesome tissues and yield
recoveries of 82 and 84%, respectively.
3
13-1
-------
When the residue chemist is faced with the task
of quantitative determination of a particular compound
or group of compounds, he can many times use existing
methodology even though the procedures have not been
applied to his specific need.
Advantages and Disadvantages
Micro methods have several advantages which in
many cases are overlooked. Analysis time can be sub-
stantially shorter with a small cleanup column. Con-
sider the time required for elution of several hundred
ml of solvent through a large column and add to that
the time for concentration to a suitable volume for
the determinative step. Elution through a micro Flori-
sil column can be accomplished in about one fifth the
time. Large volumes of solvents can be awkward to
handle and expensive in the long run. Smaller pieces
of glassware are more easily handled and provide less
surface area on which to lose pesticides.
Micro cleanup columns can be time savers when one
is screening various adsorbents and solvents in search
of suitable elution patterns. Many columns can be run
in a short period of time with minimum solvent volumes.
Once an acceptable system has been found, the column
size can easily be scaled up if desired.
When only very small sauries are available the use
of a micro procedure is mandatory. Attempts to work up
lOOmg of a tissue with methodology designed for 10 to
lOOgm are certain to fail, if for no other reason than
the tremendous reagent blank that would surely arise.
If unlimited sample is available, but the chemist
chooses to use a micro analytical method, he must con-
cern himself with obtaining a representative subsample
that will reflect the true level of pesticide or metab-
olite in the total sample. Use of a large sample for
extraction and taking an aliquot from the total extract
for cleanup by a suitable micro method would produce
such a representative sample. Micro methods of analy-
sis limit the range of confirmatory techniques that can
be used. TLC, ultraviolet, infrared and some mass
spectrometric techniques are not sensitive enough to
detect suspected residues in extracts from small scale
sample preparations. Derivatization and p-value pro-
cedures can provide valid confirmatory information for
nanogram quantities of pesticides or metabolites.
References
14. Kadoum, A.M., Bull. Environ. Contam. Toxicol. 2,
264 (1967).
15. Kadoum, A.M., Bull. Environ. Contain. Toxicol. 3,
65 (1968).
16. Kadoum, A.M., Bull. Environ. Contam. Toxicol. 3,
354 (1968).
17. Kadoum, A.M., Bull. Environ. Contam. Toxicol, 4,
120 (1969).
18. Leoni, V., J. Chromatog. 62, (1971).
19. Holden, A.V. and MaTsden,~K., J. Chromatog. 44,
481 (1969).
20. Law, L.M. and Goerlitz, D.F., JAOAC 53, 1276 (1970)
21. Johnson, L.G., Bull, Environ. ContamT"Toxicol. 5,
542 (1970).22.
22. Emey, D.R., Bull. Environ. Contam. Toxicol. 12,
717 (1974).
23. Erney, D.R., Bull, Environ. Contam. Toxicol. 12_,
710 (1974).
24. Sherma, J., and Shafik, T.M., Arch. Environ.
Contam. Toxicol., 3, 55 (1975).
25. Crist, H.C., Moseniah, R.F, and Noneman, J.W.,
Bull. Environ. Contam. Toxicol., In Press,
(1975).
26. Crist, H.C. and Moseman, R.F., submitted J. Ch.ro-
matogr. (1975).
27. Moseman, R.F., 165th ACS Meeting, Dallas, Texas,
1 077
1. FDA Pesticide Analytical Manual.
2. Mills, P.A., JAOAC, 42, 734, (1959).
3. Mills, P.A., Onley, J7H., Gathier, R.A., JAOAC,
46, 186, (1963).
4. Wills, P.A., Bong, B.A., Kanps, L.R. and Burke,
J.A., JAOAC, 55, 39, (1972).
5. EPA Manual ofTjialytical Methods, J.F. Thompson,
editor, 1974.
6. Enos, H.F., Biros, F.J., Gardner, D.T., and Wood,
J.P., 154th ACS Meeting, Chicago, 1967.
7. Storherr, R.W., Murray, E,J., Klein, I., Rosenburg,
L.A., JAOAC, 50, 605, (1967).
8. Stalling, D.L., Tindle, R.C., and Johnson, J.L.,
JAOAC, 55, 32, (1972).
9. Tindle,T^.C. and Stalling, D.L., Anal, Chem., 44,
1768, (1972).
10. Griffitt, K.R. and'Craun, J.C., JAOAC, 57, 168,
(1974).
11. Stalling, D.L., 3rd International Conference on
Pesticide Chemistry, Helsinki, Finland, July, 1974.
12. Armor, J.A. and Burke, J.A., JAOAC, 53, 761,
(1970).
13. Masumoto, H.T., JAOAC, 5J5, 1092, (1972).
4
13-1
-------
CHOLINESTERASE ACTIVITY AS A BIOCHEMICAL INDICATOR
FOR MONITORING EXPOSURE TO CERTAIN PESTICIDES
F. Kaloyanova
Center of Hygiene
Sofia, Bulgaria
Review of the monitoring system of ChEA is per-
formed with agricultural workers exposed to phosphor-
organic pesticides. The observation period comprises
11 years (1964-1974). From several thousands of per-
sons observed during the first years the system en-
larged and comprised 30,000 persons during the last
year. Studies are carried out before, during and after
the working season. The per cent of workers with
reduced ChEA before the season varies in different
years between 2 to 12%. During the working season
the percent of workers with reduced ChEA is from
1.2 to 17.8% and after the working season from 2.9
to 16%. A rapid paper test is recommended for field
studies of ChEA in serum. Workers are removed from
further exposure when ChEA is inhibited at 30% of the
preexposure level.
Many organophosphate compounds (OPhC) have been
synthesized in the past 30 years and some of them have
been widely used as agricultural insecticides.
They have a general mechanism of effect expressed
as inhibition of cholinesterase activity (ChEA), which
leads to accumulation of acetylcholine. Carbamates
have a similar effect on ChEA. The phosphororganic
pesticides caused 74% of the systematic acute poison-
ings with pesticides in California and 83% of the same
in Bulgaria during the period 1965-1968 (Kaloyanova-
Simeonova, F. and E. Fournier3). The mass contact with
OPhC requires special attention in order to eliminate
the acute poisonings and prevent the chronic ones.
At present the application of pesticides is highly
dangerous for workers which apply them and for agricul-
tural workers engaged in other kinds of work in the
treated areas.
If we have to make gradiation of the risk for the
general population, we shall have to put in the first
place contamination of foodstuffs as a c&nsequence of
incorrect technology of pesticides application and con-
tact with treated areas as well as their home usage.
Laboratory studies may serve both as a measure of
exposure and as measure of a response or resulting
pathology. The distinction between exposure and mea-
sure of effect is not always clear. The choice of
parameters in the routine supervision of workers was
discussed during the first and second Workshop of the
Subcommittee of Pesticides of Ind. Assoc. Occup. Health
(Zielhuis, Z.12, Kaloyanova, F. and Z. Zielhuis11, 1974).
The Workshop recommended for the routine supervision of
workers the monitoring tests applied should be confined
to those for which the significance in terms of human
health has been established. Nevertheless, deviation
of test findings should always be compared with pre-
exposure levels or with levels in adequate controls.
Cholinesterase Activity
OPhC act on the body principally by inhibition of
the enzyme acetylcholinesterase and pseudocholinesterase
which is responsible for hydrolyzing acetylcholin at
synaptic sites. The enzyme inhibition is almost irre-
versible. OPhC are also inhibitors of other esterases.
Acetylcholinesterase - specific ChE is found in
the CNS, erythrocytes (EChE) and musculature. Pseudo-
cholinesterase (SChE) - in plasma and liver.
Clinical Importance
Reasonable correlation exists between ChEA and
the clinical signs of poisoning.
The plasma cholinesterase activity decreases and
is normalized more quickly than that of the cell, when
in contact with phosphororganic preparations.
After a severe intoxication the reduction of en-
zymes lasts up to 30 days in plasma, and to 80-100 days
in erythrocytes. In the examination of people exposed
to chronic action of small doses of phosphororganic
compounds, the reduction of cholinesterase activity was
observed more frequently in erythrocytes than in the
plasma.
Although the measurement of ChEA may give an indi-
cation of the dose of the inhibitor absorbed, it cannot
be used alone to predict the clinical sequence of events.
A sudden slight enzyme depression is often associated
with moderate illness, while a gradual severe depres-
sion may be compatible with good health because the body
has been adapted to the high levels of the accumulated
acetylcholine (SvetliJfic, B. and K. Wilhelm9, 1972).
The clinical progressive decrease of ChEA is tol-
erated easier than the abrupt decrease. After a con-
tinuous exposure disturbances could be manifested only
at inhibition with 75-85% of the level before the ex-
posure of the erythrocyte ChEA.
Individual Variations
Individual differences between cholinesterases
are great in human blood serum, depending on physiolog-
ical and genetic factors regulating the synthesis of
the enzyme in liver. So far, 5 genes responsible for
the production of cholinesterase in human serum have
been recognized. These genes function within the same
allele; the allele is the reason for the existence of
10 genotypes.
One can suppose that people are sensitive to the
action of phosphororganic compounds depending upon
their sets of genes synthetizing this enzyme.
1
13-2
-------
The method presented by Bonderman, P. et al, 1971
distinguishes between lowered cholinesterase due to
exposure to cholinesterase-inhibiting insecticides and
a normally low activity due to atypical esterases. The
numbers of persons with the genetic anomally are low
(about 3%).
It is possible on the basis of dibucaine numbers
using the pH-Stat method to separate accurately cases
of lowered esterase due to OPhC from those due to
genetic anomalies.
The cholinesterase activity in some people is
lower than the mean activity, which can be regarded as
a normal individual deviation. Dispersion of acetyl-
cholinesterase activity is ± 20 percent, and in plasma
100 percent in relation to an average activity of the
noriexposed group. That is why the application of
assessments of cholinesterase activity as an index of
exposure of people in occupational contact with phos-
phororganic compounds is a very difficult task in the
case where no determinations of cholinesterase activity
had been made before the beginning of work. Only when
the initial determinations have been made it is possi-
ble to state whether the decrease in cholinesterase
is significant or not.
Threshold Limits of Inhibition of ChEA
Literature data give different values on this
problem. As a result of studies performed in agricul-
ture (Kaloyanova-Simeonova, F.2, 1959) we accepted
inhibition with a 30% preexposure level as an index
for absorption of phosphororganic compounds. Those
working with inhibition of ChEA above this percent are
removed from work to avoid poisoning.
Cage, I.1 (196?) proposed as safe threshold the
same percent for ChEA in the plasma and in the erythro-
cytes. Strict control and safety measures should be
taken with values of inhibition between 20 and 257.
without the necessity for interrupting work. The de-
crease of inhibition at 30% requires interruption of
the exposure. The opinion of Teizinger I.10 (1969) is
analogical.
Michaux et al.^ (1970), Podolak M.® (1971) have
suggested that a decrease of 25Z of SChEA and of 20%
of EChEA should be a signal to remove the worker from
exposure.
Zavon M.R. (1965) reports that the normal fluc-
tuations of ChEA vary as much as 25% in separate indi-
viduals, therefore the intial level should be known.
He suggests that the determination of the erythrocyte
ChEA is more important, as the serum ChEA is effected
by other substances also, and recovers more quickly.
Decrease of ChEA with more than 40% has to be consider-
ed as alarming signal, and with more than 60% the work-
ers have to be removed from the contact with phosphor-
organic pesticides.
As a general guide, levels below 50 percent of
the mean values for general population are taken as
indicative of significant depression by Mastromateo E.5
(1971) and at levels below 25 percent, it is recommended
that workers should be removed from further exposure
and the test repeated. Greater attention, of course,
is paid to red cell activity.
Regulations for removing an aircraft pilot engaged
with spraying of phosphororganic preparations should
be more strict.
Investigating the cholinesterase activity is a
very important test in the preliminary examinations of
persons exposed to professional risk of pesticides,
because an inhibition of ChEA is ofcen established,
which could be explained with a chronic exposure of
these workers for a long period of time, or with liver
disturbances, which could be a contrary indication or
with individual peculiarity.
Increase of ChEA is established with diabetes,
hypertonia, arthritis, hyperthroidism.
A system of ChEA testing is known to prevent
manifested intoxication and to detect Incipient ones.
Monitoring System of ChEA
After a system proposed by us the workers in
Bulgaria engaged with pesticides are under dispensary
observation. It is performed by the district physicians
under the management and assistance of the Hygiene
Epidemiological Inspectorated and particularly the
toxicologic laboratories, which accomplish the labora-
tory tests, and the district hospitals.
Every year, before starting the plant protective
work, a preliminary examination of the workers is
carried out including tests on the cholinesterase
activity and some samples for liver function.
In the course of the working season ChEA tests
are carried out for measuring the exposure of workers
and light poisonings. Another examination is also con-
ducted at the end of the working season.
We conducted analysis on the results received
from investigated workers with full documentation from
the Hygiene Epidemiological Inspectorates for an 11 year
period for the whole country (Table 1).
For the period 1965-1968 the Pravditch Neminska
colorimetrlc method with modification by Kaloyanova, F.
(1959) was used for whole blood and for the period
1969-1974 the paper test (Mosheva N. et al., 1969) was
used for plasma.
From several thousand persons observed during the
first years the system was enlarged and comprised 30,000
persons during the last year. Studies were carried out
before, during, and after the working season. The
percent of workers with reduced ChEA before the season
varied in different years between 2 and 12%, during the
work season from 1.2 to 17.8% and after the working
season from 2.9 to 16%.
A high estimation was used for the individual and
group prophylactic .action during the discussion of the
system. In the course of the preliminary examinations
persons with lowered ChEA,'because of different reasons,
were registered, and not permitted to work with inhib-
itors of ChEA. The studies were carried out during
work for a period of more than 10 days. The results
are being used for the improvement of environmental
control, personal protection or personal hygiene.
The least useful data are from those workers who
register after the working season. It could be argued
that these results are superfluous and this problem is
under discussion. Their only importance is when a
light effect is considered to be cause for placing the
worker in a free of charge spa or sanatorium regime.
2
13-2
-------
From all different methods which have been used
during both periods 1965/1968 and 1969/1974, a pref-
erence could be given to the rapid paper test because
of its convenience as a field test and the compatible
results with other methods.
Conclusion
Monitoring of a level of ChEA (as a specific
parameter of response in OPhC exposure) will be made
available to all workers as well as to the exposed
population. Such monitoring shall be performed to
ensure that no person absorbs an unacceptable amount
of OPhC.
In acute poisonings there is a good correlation
between the inhibition of ChEA and the clinical mani-
festations. During continuous exposure to small con-
centrations OPhC a considerable suppression of ChEA
Is often present without symptoms, which points to a
certain adaptation of the organism.
It is imperative to study the cholinesterase
activity before starting work because of the individual
variations and the presence of genetically determined
deficiency of ChE.
Suppression of 30% from the initial level of ChEA
should be considered as a signal for considerable
absorption of OPhC and removal of the worker from
further contact until he recovers. As an exposure
test - inhibition of 20% with a group of workers points
out to sanitary failures.
Other factors, which decrease the activity of
ChEA, should be taken into consideration.
All of the numerous methods existing in literature
should be suitable, if correctly applied. The best
method is to study the serum and erythrocyte ChEA. The
paper tests are most convenient and comparatively
accurate.
3
13-2
-------
TABLE I
NUMBER OF PERSONS CHECKED FOR ChEA AND PERCENT WITH INHIBITED ChEA
Years of Before the Working Season During Spraying After the Working Season
Examination Total Number with Total Number with Total Number with
Number Inhibition % Number Inhibition % Number Inhibition %
1964
2021
37
1.8
2392
44
1.4
436
41
9.4
1965
3182
106
3.3
2322
181
7.8
542
34
6.12
1966
5589
147
2.6
1729
132
7.67
1672
176
10.5
1967
10301
1237
12.0
2585
154
5.95
2602
77
2.9
1968
13711
705
5.2
1161
14
1.2
4031
580
14.6
1969
19372
1188
6.13
4737
280
5.9
4528
360
7.95
1970
19288
1568
8.10
4594
816
17.8
5618
486
8.66
1971
28486
2183
7.70
9062
703
7.75
5928
619
10.40
1972
27461
1982
7.25
9135
1017
11.05
6385
473
7.41
1973
28747
1153
4.04
9558
886
9.26
8971
989
11.00
1974
31291
2099
6.72
8520
749
8.80
6703
332
5.00
References
1. Gage, J. C. The significance of blood cholin-
esterase activity measurements. Residue Reviews,
1967, 18, 159-173
2. Kaloyanova, F. Effects of certain organophos-
phorous insecticides upon cholinesterase activity
in persons in conditions of agricultural labour.
Sb. Trudove na NIOTPZ, 1959, 6, 105-114
3. Kaloyanova-Simeonova, F., E. Fournier. ^Les
pesticides et l'homme. Coll. de Med. legale et
Tox. medicale, Masson and Sie, 1971, Paris
4. Kaloyanova, F. and B. Zielhuis. Occupational
Health Criteria for evaluation of pesticides.
Int. Arch. Arbeitsmed., 33, 335-341, 1974
5. Mastromateo, E. Cholinesterase testing program in
persons exposed to organic phosphorous insecti-
cides. I Workshop Pest. Comm., Amsterdam, 1971
6. Michaux, P.,^H. L. Boiteau, F. Tolot. Valeur et
limites du depistage clinique et biologique en
pathologie professionnelle. Arch. Malad. Prof.
Med. Trav. Sec. Soc. 32, 1971, 1-2 Janv. - Fevr.,
1-124
7. Mosheva, N., Iv. Benchev, F. Kaloyanova. Reactive
papers for rapid determination of serum cholin-
esterase. Rationalization certificate, N-x,
1960/11 Dec. 1969
8. Podolak, M., Szuki, B. Human exposure to phos-
phororganic insecticides taking into account the
action of these compounds on particular phenitypes
of cholinesterase. I Workshop Com. Pest. Amsterdam,
1971
9. Svetlicic, B., K.Wilhelm. Methods of measuring
exposure to anticholinesterase insecticides. Arch,
hig. rada, 24, 1973, 357-365
10. Teisinger, J. Les tests biologiques d'exposition,
Cahiers de Notes Documentaires - Note N 766, 65-71,
d'apres Pr. lekarstvi, 1969, Vol. 21, No 9
11. Zavon, M. R. Blood cholinesterase levels in
organic phosphate intoxication. J, Amer. Med.
Assoc. 192(1) 137, 1965
12. Zielhuis, Z. Epidemiological toxicology of pesti-
cide exposure. Arch. Env. Health, 25, 1972, 393
4
13-Z
-------
AQUATIC SNAKES AS COMPOSITE SAMPLES FOR ORGANOCHLORINE PESTICIDE RESIDUES
H. Erie Janssen, Jr., Environmental Engineering,
Texas A&M University, College Station, Texas
Reid Dennis, Biology Department, Sam Houston
State University, Huntsville, Texas
James R. DeShaw, Biology Department, Sam Houston
State University, Huntsville, Texas
Summary
Introduction and Literature Review
The environmental monitoring of organochlorine
pesticides has largely been based on water or sediment
samples. These may or may not indicate the levels of
organochlorine pesticides in the watershed. When
pesticide residues are found, they are in the ppb (ug/
kg) range, which in itself presents a difficult analy-
tical task.
Persistent organochlorine residues have been
quantified in many forms of wildlife in the ppm (ug/g)
range. The level of residues provides an indication
of pesticide exposure based directly on the diet,
ranqe, and trophic level of the organism analyzed.
Most of these organisms have a wide range and varied
feeding habits, thus they provide little indication
of the pesticide levels in the area. An exception to
this is the aquatic snakes. They feed predominantly
on water-related organisms, and are limited by the
aquatic environment.
A survey of the pesticide residues in aquatic
snakes from several Texas watersheds has provided data
which relate to the general land use and thus the
pesticide use in the watershed. DDE residues in
aquatic snakes range from 0.2 ppm in a non-agricultural
area to over 1000 ppm in an area used for extensive
agriculture. The presence of fat soluble persistent
organochlorine compounds in a watershed could be more
accurately assessed by the use of aquatic snakes as
a primary survey tool with sediment samples providing
a supportive method for tracing a particular pesticide
source.
Table 1. Selected organochlorine compounds reported for water
from Brazos River, Richmond, Texas, 1968-1973.
The river basin monitoring of organochlorine
pesticide residues has been based upon surveys that
determine the concentration of these substances in
water and sediments. The United States Geological
Survey and various state departments of agriculture
currently sample water and sediments from established
stations on major rivers throughout the United States.
The majority of the water analyses report concentra-
tions of the major organochlorine insecticides to be
below their minimum detection limits of 0.01 ppb.7/10-!5
The analysis of sediments for these organochlorine
compounds produces a higher percentage of positive
samples. The values are in the ppb range and seem to
provide a measure of the insecticide residues in a
watershed. But these samples show wide monthly
variation.9 f13-15,16
The use of water and sediments as a method of
monitoring organochlorine residues has some disadvan-
tages. Frequently, these surveys yield data that
suggest zero to very low quantities of these sub-
stances in an aquatic environment associated with an
area where pesticides are known to be used. An
example of such an area is the Brazos River flood
plain in Texas. Table 1 indicates the occurrence of
detectable quantities of selected organochlorine
compounds in water from the Brazos River at Richmond,
Texas. The most common insecticide residue was DDE
which was detected in 38 percent of the samples. The
highest concentration of DDE reported was 0.06 ppb.lCH-^
YEAR
DDE
DDD
DDT
OBSERVATIONS
(ppb)
(PPb)
(ppb)
(ppb)
(PPb)
REFERENCES
1968
o.oi i o.oi*
(80%)
O.OO ± 0.00 0.01 t 0.01
(30%) (60*>
o.oo i o.oo
(10% >
Not:
Reported
0.01 t 0.02
(20%)
0.00 t 0.00
(20%)
0.01 t 0.02
(20%)
( 0%)
Not
Reported
0.01 t 0.02
(30%)
0.01 t 0.01
(30%)
°*02 t 0.03
(30%)
0.00 t 0.01
(10%)
0.5
(10%)
0.01 ± 0.02
(17%)
( 0%)
0.07 ± 0.16
(17%)
< 0%) 0.1 t 0.2
(33%)
( 0%)
( 0%)
( 0%)
0.1
(100%)
0.00 t 0.01
(33%)
( o%)
( 0%)
( 0%)
( 0%)
1968- 40 0.01 t 0.01 0.00 t 0.01 0.02 t 0.06 0.00 + 0.00 (10%)
1973 (38%) (20%) (30%) ( 3%)
*The mean and standard deviation is indicated for each compound.
NOTE: Values in parentheses indicate percent of samples positive for residue.
13-3
-------
Table 2. Selected organochlorine compounds reported for sediments
from the Brazos River, Texas, 1910-1973.
YEAR
NUMBER OF
DDE
DDD
DDT
Dieldrin
PCB
OBSERVATIONS
(ppb)
(ppb)
(ppb)
(ppb)
(ppb)
Richmond
70-71
3
<0.2 - 15.0*
< 0.2
-
5.1
< 0.2
1.3
< 0.2
Not
13
6.6 t 7.6
2.5
2.5
0.5
+
0.7
Reported
70-71
9
0.0 - 34.0
0.0
-
59.0
0,0
_
24.0
0,0
0.0
9
11.6 t 12.2
12.0
+
19.3
5.0
+•
8.5
72
2
0.0
0.0
-
4.0
0.0
5.0
0.0
0.0
9
2.0
2.B
2.5
i
3.5
73
3
0.0 - 2.7
0,0
_
0.6
0.0
_
2.2
0.0
0.0 - 14
15
1.1 t 1.4
0.2
+
0.4
0.7
t
1,3
4.7 + 8.1
70-73
17
0.0 - 34.0
0.0
59.Q
0.0
-
24.0
0.0
0.0 - 14
7.5 ± 10.3
6,4
±
14.0
3.2
t
6.4
Washington
70-71
0.0 - 68.0
18.8 t 23.6
0.0 - 27.0 0.0 - 67,0
7.4 1 9.1 11.0 I 25,0
0.0
0.0
Bty&si
70-71
0.0 - 29.0
15.6 + 10.5
0.0 - 16.0
6.1
S.£
0.0 -
2.0 1
9.0
3.8
0.0
0.0
All Three Sites 70-73
12.0 I 14.6
<72»>
6.5 t 11.2
(69%)
5.3 I 12.9
(31%)
0.0
0.7 ± 3.0
(12%)
•The range, mean and standard deviation are indicated for each compound.
I Values in parentheses Indicate percent of samples positive for residue.
NOTE
The analyses of sediments from the same station
(Table 2) produced more positive results than did the
water. DDE was detected in 59 percent of the samples
with a mean concentration of 7.5 ppb. The maximum
level of DDE reported was 34.0 ppb.9>13-15 Numerous
data have been published2'5 that indicate that these
compounds are rapidly concentrated into the sediments,
plants, and organisms when introduced into the aguatic
environment. These compounds tend to be more concen-
trated in the predatory organisms, particularly those
who obtain their food via the aquatic community.2'5
Fleet, et al.3 and Stafford8 have presented data
that indicate that snakes might provide a measure of
the organochlorine insecticides use in a watershed.
Fleet, et al_., reported DDE levels in snakes from the
Brazos River flood plain, an area used for extensive
cotton production, to be from 32 to 434 times greater
than the DDE levels from the nearby Navasota River
flood plain, a partially wooded watershed used pri-
marily for grazing land.^
Stafford® presented data on DDE and PCB residues
in snakes from three areas near Bryan, Texas for 1972-
1973. The study area included the two areas previously
studied by Fleet, et al.,3 with an additional area near
a major highway.
Methods
During the summer of 1974, a research project was
undertaken by a student group at Sam Houston State
University. Among the environmental considerations of
the project was the possibility of large amounts of
organochlorine pesticide residues being concentrated
into the alligator via the food chain. In order to
explore this possibility, it was decided that aquatic
snakes would be explored as a measure of the residues
in a watershed*
Aquatic snakes were captured live from the water
or nearby shore at each sample site during the period
from June through August, 1974. The snakes were
transported alive to the laboratory and frozen until
processed for analyses. The snakes were thawed,
measured, and dissected. The snakes' fat bodies were
removed, weighed, and retained for analyses.
The analyses of the snake fat were performed
according to the methods given in the Pesticide
Analytical Manual.'' The procedure involved fat
extraction, clean-up solvent partitioning with
acetonitrile, and Florisil chromatography. The
residue determinations were made on a Micro-Tek 220
Gas Chromatograph with a Ni63 detector and fitted
with a 4 mm x 6 ft. glass U-column packed with 4 per-
cent SE-30 and 6 percent QF-1 on Chromosorb W-HF, A
second 4 mm x 4 ft. glass U-column packed with 3 per-
cent SE-30 on Varaport 30 was used to verify the
results of several samples. The results are reported
on the basis of petroleum ether extracted fat
obtained from the snakes' fat bodies.
Results and Discussion
A total of eleven snakes were captured and
analyzed from five locations across the State of Texas
(Figure 1).
The fat extracted from the fat bodies of the snakes
accounted for 0.5 to 2.0 percent of the snakes' total
body weight. Organochlorine residues were detected
in all samples in the ppm (ug/g) range. The major
compound detected in all samples was DDE with DDD,
DDT and Dieldrin also found in the majority of the
samples. The DDE levels ranged from 0.2 to 32. ppm,
with the lowest values recorded from the Nueces River
site and the highest recorded for the Trinity River
site (see Table 3).
2
13-3
-------
Navaaot
Bry
Pashiagt
HWY 21
docket
¦Madieoflville
Trlflit/ River
-Bedias Creek
Bueeea Blver
Saither's Lake
'Matagorda Salt Mareh
Figure 1. Index map of study areas.
The results indicate that aquatic snakes can be
used to estimate the organochlorine pesticide levels
in a watershed. Unlike many predatory animals, the
aquatic snakes tend to have a somewhat limited range
and obtain a large portion of their diet from the
aquatic community.1 From the data reported in this
survey (Table 3), it can be noted that residues in
aquatic snakes show a wide variation when compared
overall, 0.3 to over 1000 ppm, but a low variation
when compared for snakes analyzed from the same
sample site: Matagorda, 1.3 and 1.7 ppm; smither's
Lake, 14.8 and 15.5 ppm; and from Fleet's, et al.,3
1971 data, 8.7 to 16.2 ppm for Navasota Basin and 370
to 1000 ppm for the Brazos Basin. When the general
land use of the upstream watershed is compared with
the level of pesticide residues, a general correla-
tion of pesticide residue level to upstream land use
becomes apparent. Fleet, et al.,^ found residues in
aquatic snakes from 370 to 1000 ppm from an area used
for extensive agriculture as compared to 8.7 to 16.2
ppm for a nearby non-agricultural area.
Table 3. Organochlorine compounds reported for aquatic snakes
from various watersheds in Texas.
YEAR
r—
NUHBER op
DBE
DOD
DVT
Dieldrin
PCB
SKfcKES
(ppm)
{ppm)
(ppm)
(ppm)
(ppm)
W11
10
283.6 - 1009.4*
509.0 t 214.9
+ - 7.3
1.1 ± 2.2
4.7 - 37,9
14.6 ± 12.5
1.3 - 9.7
4.4 i 2.6
Not
Reported
3
Br&gos
1972-
19*73
12
31.0 - 1161.2
528.1 ± 449.7
Not
Reported
Not
Reported
Wot
Reported
Not
R^portfcd
6
1974
12
73.9 -
202.2 ±
388.0
108.8
Not
Reported
0.6 - 8.2
3.6 ± 2.9
wot
Reported
Not
Reported
6
Navasota
1971
1972-
1973
10
17
0.4 -
5.1 ±
0.6 -
2*3 t
14.6
5.4
6.8
1.9
+ - 0.6
0.2 £ 0.2
Not
Reported
+ - 1.2
0.4 ± 0,4
l*ot
Reported
+ - 0.2
0.1 * 0.1
Not
Reported '
Not
Reported
Not
Reported
3
8
Highway 21
1972-
1973
14
2.3 -
16.2 +
68.2
16.4
Not
Reported
Not
Reported
Not
Reported
11.7 - 123.3
45.2 i 33.2
8
Trinity River
1974
1
31.6
3.9
4.7
0.9
Not
Reported
This
Report
Bedias Creek
1974
1
1.9
0.1
0.1
0.02
Not
Reported
This
Report
Smither's Lake
1974
2
12.8 -
13.3
13.7
0.6 - 0.7
0.7
0.3 - 0.7
0.5
0v4 - 0.6
0.5
Not
Reported
This
Report.
Nueces River
1974
1**
0.2
0.03
0,08
Not
Reported
Not
Reported
This
Report
Matagorda
Salt Marsh
1974
2**
0.5 -
0.6
0.6
0.12 - 0.21
0.19
0.07 - 0.73
0,15
O.S - 0.7
0.6
flat
Reported
Thie
Report
»Tha range, rce.an, and standard deviation are indicated for each compound.
"Sample represents composite of mora than one snalce.
NOTE; All residues are on basis of extracted fat obtained from the snakes' fat bodies.
3
13-3
-------
The data reported in this paper include analyses
of snakes from several different land use areas. The
Nueces River site was a remote arid region with little
agricultural use and showed the lowest residues
recorded in this study, 0.3 ppm. The Matagorda site
was a salt marsh with tidal influence and showed low
overall residues, 1.7 and 1.3 ppm. The Bedias Creek
site was an area predominantly used for grazing land
and contained a DDE residue of 1.9 ppm. The Smither's
Lake, Highway 21, and Trinity River sites contained
residues of DDE in the same general range, means of
13.3, 16.2 and 31.6, respectively. The Navasota site
showed low levels of DDE residues averaging 2.3 ppm.
The Brazos site yielded snakes with high DDE residues.
A decline in residue was noted for 1974. The DDE
means were 1971, 509 ppm; 1972-73, 528,1 ppm; and
1974, 202.2 ppm (see Table 3). The DDE residues in
snakes from the Brazos River were significantly
higher than those from the Navasota River. Likewise,
the DDE residue in snakes from the Brazos River were
significantly higher than those in snakes from the
Highway 21 location.
Table 4 contains the results of several pesti-
cide monitoring programs for two sites on the Trinity
River.
These sediment data show wide variation but fairly
low overall residues, DDE 2.0 t 8.0 ppb; DDD 2.2 ±
7.7 ppb; DDT 0.1 ± 0.2 ppb; Dieldrin 0.4 ± 1.8 ppb;
and PCB 14.3 ± 38.7 ppb. Table 2 contains the results
of two pesticide monitoring programs on the Brazos
River. These data show residues of DDE 12.0 ± 14.6
ppb; DDD 6.5 i 11.2 ppb; DDT 5.3 ± 12.9 ppb; and PCB
0.7 - 3.0 ppb. From these tables, it can be noted
that the Brazos River contains (based on sediment)
about six times the residues of the Trinity River
except for PCB. The snakes from the Brazos area con-
tain from seven to 16 times the residues of the snakes
from the Trinity River.
Based on data obtained by several investigators,
ar)<3 work of the authors, there appears to be
no reason to separate water snakes into species cate-
gories for analytical purposes. Generally speaking,
based on the level of DDE in snaXe fat (Table 5), the
species of water snakes studied have residues in the
same general range.
Conclusions
1. Even though the data on this topic are
limited, it appears that residue levels in aquatic
snakes would provide more consistent data on persis-
tent fat-stored compounds in a watershed than would
sediment or water analyses.
2. The analyses of aquatic snakes for persis-
tent organochlorine compounds provide the advantages
of a composite sample as compared to the grab sample
used in sediment and water analyses.
3. Aquatic snakes contain residues on the order
of 1000 times more concentrated than the levela
commonly found in sediments. This reduces the
problem of contamination during sample collection and
processing.
4. Based on the evidence shown in this paper,
the presence of persistent organochlorine compounds
in a watershed could be more accurately assessed by
the use of aquatic snakes as the primary survey tool
with sediment samples providing a supportive method
for tracing a particular source within an area.
Table 4. Selected organochlorine compounds reported for sediment
from the Trinity River, Texas, 1970-1973.
SITE
NUMBER OF
DDE
DDD
DDT
Dieldrin
PCB
OBSERVATIONS
-------
Table 5. DDE residues reported for aquatic snakes from the
Brazos River near Bryan, Texas.
NUMBER OF
DOII (ppra)
YEAR
SPECIES
SNAKES
Range
| Moan
Standard |
Deviation
1971
N, sipedon
3
445.2 - 673.0
590.1
130.0
3
N. erythrogaster
3
283.6 - 380.4
349.3
56.1
3
A. piscivorus
4
363.3 1009.4
568.5
297.8
3
1971
All
10
2B3.6 - 1009.4
509.0
214.§
3
1972-
1973
N. sipedon
N. erythrogaster
2
4
617.1 - 724.7
67.2 - 1161.2
670.9
787.1
76.1
507.13
8
8
A. piscivorus
6
31.0 - 1102.3
309.0
411.8
8
1972-
1973
All
12
31.0 - 1161.2
528.1
449.7
a
1974
N. sipedon
4
148.1 - 364.0
211.6
102.1
6
N. erythrogaster
4
73.9 - 361.6
178.9
136.35
6
A. piscivorus
4
155.5 - 388.8
216.2
115.12
6
1974
All
12
73.9 - 388.6
202.2
108.76
6
Acknowledgements
This study was supported by the National Science
Foundation Student Originated Studies Grant, No. GY-
11523, EPA Training Grant, No. T-900129-3, Dallas
Water Utilities, Dallas, Texas and the Civil
Engineering Department, Texas ASM Unviersity, College
Station, Texas.
The authors achnowledge Mr. w. H. Vance, for
his guidance and assistance in the analytical tech-
nique.
Bibliography
1.
Ditmars, R. L. 1960. Reptiles of the World. The
MacMillian Company. New York. 3120 pp.
2.
Edwards, Clive A. and Russel S. Adams. 1970. Per-
sistent Pesticides in the Environment—In CRC
Critical Reviews in Environmental Control, 1:7-
67.
3.
Fleet, Robert R., Donald R. Clark, Jr., and Frederick
W. Plapp, Jr. 1972. Residues of DDT and
Dieldrin in Snakes from Two Texas Agro-Systems,
Bio Science, 22:664-665.
4.
Food and Drug Administration. 1968. Pesticide
Analytical Manual, U. S. Department of Health,
Education, and Welfare, Washington, D. C.,
Volume 1.
5.
Meeks, Robert L. 1968. The Accumulation of 36 CI
Ring-Labeled DDT in a Freshwater Marsh, J.
Wildlife Management, 32:376-398.
Plapp, Frederick W. and Robert R. Fleet (Texas ASM
University),1975, Personal communication.
7.
Schulze, J. A., Douglas B. Manigold and Freeman L.
Andrews. 1973. Pesticides in Selected Western
Streams—1968-1971. Pesticides Monitoring
Journal 7(1):73-84.
Stafford, Duane, P. 1973. Relationships between
Detoxifying Enzymes in Several Snake Species
and the Occurrence of These Species in Clean
and Pesticides-Contaminated Ecosystems. M. S.
Thesis. Texas ASM University, College Station,
Texas.
9.
Tidswell, Brooke, III, and William E. McCasland. 1972.
An Evaluation of Pesticide Residues on Silt and
Sediment in Texas Waterways, Texas Department of
Agriculture, Austin, Texas.
10.
United States Geological Survey. 1968. Water
Resources Data for Texas. U. S. Department of
Interior, Washington, D. C., Volume 2.
11.
United States Geological Survey. 1969. Water
Resources Data for Texas. U. S. Department of
Interior, Washington, D. C., Volume 2.
12.
United States Geological Survey. 1970. Water
Resources Data for Texas. U. S. Department of
Interior, Washington, D. C., Volume 2.
13.
United States Geological Survey. 1971. Water
Resources Data for Texas. U. S. Department of
Interior, Washington, D. C., Volume 2.
14.
United States Geological Survey. 1972. Water
Resources Data for Texas. U. S. Department of
Interior, Washington, D. C., Volume 2.
15.
United States Geological Survey. 1973. Water
Resources Data for Texas. U. S. Department of
Interior, Washington, D. C., Volume 2.
16.
Whitman, Richard L. 1973. Survey of Pesticides in
the Upper Trinity, Texas. M. S. Thesis.
Stephen F. Austin State University, Nacogdoches,
Texas.
5
13-3
-------
A NATIONAL PESTICIDE MONITORING PROGRAM OVERVIEW
by
Thomas C. Carver, Jr.*
National Marine Fisheries Service
National Oceanic and Atmospheric Administration
Washington, D.C. 20235
Introduction
The National Pesticide Monitoring Program is sponsored
by the Monitoring Panel of the Federal Working Group on
Pest Management. Membership in the Monitoring Panel is
drawn from components of the Departments of Agriculture,
Defense; Commerce; Interior; Health, Education and Wel-
fare; Environmental Protection Agency; Tennessee Valley
Authority and the National Science Foundation.
In addition to sponsoring the National Pesticide
Monitoring Program, the Panel encourages the develop-
ment and use of uniform sampling and analytical method-
ologies. The Panel has recently published guidelines
for both. The Panel also sponsors The Pesticides
Monitoring Journal, now in its eighth year, which
serves as the primary outlet for reports resulting
from the National Pesticide Monitoring Program (NPMP).
The Monitoring Panel has defined monitoring as the
repeated sampling and analysis of components of the
environment to produce reliable estimates of the pesti-
cide levels in the components and the change of these
levels in time.
The National Program was described in Vol. 1, No. 1
of the Pesticides Monitoring Journal in 1967. The
program was initially designed on the basis of the
minimum monitoring needed to meet program objectives of
establishing baselines, detecting trends in time and
identifying potential problems.
Several component parts of the NPMP have been funded by
direct appropriation, while others have been accommoda-
ted within existing agency programs with no specific
appropriations. Irrespective of the specific fiscal
support, substantial accomplishments have resulted from
modifications of existing programs or new programs
having objectives other than monitoring as a primary
goal.
The National Program was updated in Vol. 5, No. 1 of
the Pesticides Monitoring Journal. Changes resulting
from advances in analytical and sampling techniques,
changing use patterns of pesticides, introduction of
new compounds and from variations in Departmental
missions, are a continuing process which will necessi-
tate periodic revision.
There are other aspects common to all elements of the
National Program. Problems certainly must be included
as well as objectives and meaningful analytical sensi-
tivity. We have recently addressed the analytical
sensitivity issue and have concluded that some
semblance of order Is desirable. Practical limitations
have a tendency to outweigh the biologist's need for
lower limits of biological significance. We believe
that an adequate compromise should and can be reached.
Problems were mentioned earlier as being common to all
elements of the National Program. In order to identify
and, hopefully, resolve the problems associated with
the fragmentation of agency priorities, the
Environmental Protection Agency funded a contract with
a private group for an assessment of the National
Pesticide Monitoring Program. The final report was
recently made available to members of the Panel and its
recommendations are currently being studied.
The National Human Monitoring Program functions to
determine the levels and incidences of pesticide resi-
dues and metabolites in the general population of the
United States. Adipose tissue is collected and
analyzed for chlorinated insecticides and PCB's.
Collection sites are selected according to a strati-
fied probability sample of population centers; data are
expected to be representative of the general population
of the United States.
All laboratories participating in the program are
required to maintain acceptable standards in an extern-
ally-moderated quality assurance network. Data
including age, sex, race, geographic location,
diagnostic information and residue profile, are
managed electronically.
In addition to the ongoing ambient human program, the
National Human Monitoring Program engages in several
short activities directly related to the regulatory
responsibilities of the Environmental Protection
Agency.
National Water Pesticides Monitoring Program
The first National Monitoring Program for the continu-
ous surveillance of pesticide residues in surface
waters was designed in 1965. The initial design
called for 53 sampling sites located near the mouths of
major rivers and intermediate points on several of the
largest rivers. The original plan was partially
implemented and remained in operation until 1973 when
a revised program was established between the
Environmental Protection Agency and the Geological
Survey.
The current design is based on a random selection of
sampling sites representing drainage from the major
river basins of the United States and Puerto Rico.
Sampling frequency was based on earlier experience and
now includes semi-annual bottom sediment sample collec-
tion as well as quarterly water sample collection at
153 stations. The residues that are monitored include
the common chlorinated hydrocarbons, PCB's and certain
herbicides, and organophosphate insecticides, and PCN's.
Results of the analyses are entered into the data
storage and retrieval systems of both agencies.
The National Soils Monitoring Program -was one of the
earliest operational components of the National
Pesticide Monitoring Program and was designed and
initiated by the U.S. Department of Agriculture. The
program was transferred to the Environmental Protection
Agency In December 1970.
The system is designed to determine ambient levels of
pesticides and other major pollutants in soil and in
raw agricultural crops and, through periodic sampling,
to determine changes in these levels.
*Designated Representative: Pesticide Monitoring Panel, Federal Working Group on Pest Management
-------
In the original design, two land-use categories, crop-
land and non-cropland, were recognized. Cropland was
to be sampled at the rate of one four-hectare block for
every 16,194 hectares, while non-cropland was to
sampled at the rate of one four-hectare block for
every 161,943 hectares, or one-tenth the late for crop-
land. One-quarter of the allocated sites in each state
were to be sampled each year.
At each cropland site, a composite soil sample and a
crop sample are taken. Sample collection is generally
timed to coincide with harvest of the crop. In addi-
tion, the following information is collected for each
site: crops grown on site, irrigation, pesticides used,
crop each pesticide is applied to, pounds active
ingredient applied, formulation and method of applica-
tion.
All soil samples are analyzed for arsenic, chlorinated
hydrocarbons, organophosphates, and triazines.
Samples are also analyzed for phenoxy herbicides when
use records indicate application. Crop samples are
routinely analyzed for chlorinated hydrocarbons and
organophosphates, while triazines and phenoxy herbi-
cides are analyzed for if applied.
Monitoring of urban soils has been conducted each year
since 1969 in selected cities and their surrounding
suburban areas. The selection of sites for the Urban
Program is based on one site per 2.59 within the
city limits and one site per 51.8km^ outside the city
limits. Urban samples are analyzed for chlorinated
hydrocarbons, organophosphates and arsenic. Since 1972,
the samples have also been analyzed for mercury,
cadmium and lead.
Results of the cropland and urban programs have
indicated that agricultural soils are contaminated with
low levels of pesticide residues and that urban soils
carry a higher pesticide burden than agricultural soils.
The primary problem which the National Soils Monitoring
Program has encountered is the lack of multi-residue
chemical methodologies for many of the widely-used,
newer generation pesticides. While these newer pesti-
cides can be investigated by short-term special studies
with a small number of samples, chemical methodologies
amenable to large numbers of samples are necessary if
routine analysis on all samples is desired.
EPA NATIONAL ESTUARINE MONITORING PROGRAM
Objectives ,
The Environmental Protection Agency determines the
extent to which man-made substances are directly toxic
to the biota, or indirectly harmful as a result of
environmental degradation.
Secondly, EPA must determine whether its regulatory
measures and other legal restrictions are effective in
controlling the introduction of toxic substances to the
environment at acceptable levels. Persistent residues
must be not only tolerable from the point of view of
man's health but also conducive to the welfare of eco-
systems in general.
Initial attempts to assess the pesticide pollution in
the marine environment were based on the concept that
estuaries act as sinks for water-borne persistent sub-
stances and on the knowledge that molluscs serve as
efficient collectors of materials present in only trace
amounts in their environment.
In 1965, the former Bureau of Commercial Fisheries
established a pesticide monitoring program using
estuarlne molluscs. Samples of oysters, clams, or
mussels were collected monthly at about 180 fixed sta-
tions in most of the coastal states. By 1972, more
than 8,000 samples had been screened for the most
common organochlorlne pesticides. In the final years
of this program, the decline in number and magnitude
of detectable residues made obvious the need for other
indicator organisms that could be sampled less
frequently and thus reduce the considerable physical
and financial burden.
The mollusc program, under the aegis of EPA since 1970,
undertook an entirely different approach in July, 1972.
Although fish were known to accumulate pesticides,
they posed a problem in that it was difficult to
determine when or where the residues were acquired, and
whether they represented the total accumulation. For
it was known that pesticide residues are mobilized and
lost to some extent in the production of spawn.
These difficulties have been largely overcome by
sampling juvenile fish on a 6-month basis. At least
two species of fish with different food habits are
sampled in each estuary to broaden our knowledge of
pesticide kinetics in different food webs. Each
sample consists of two groups of 25 pooled fish, which
are analyzed in replicate to minimize the effects of
individual variations.
In addition to the organochlorlne pesticides, samples
are now screened for organophosphate and carbamate
pesticides, phenoxy herbicides, and inorganic cadmium,
lead, and mercury.
In the first 30 months of the estuarine fish program,
data from 1,052 samples confirmed the general decline
in sDDT residues observed in molluscs. Localized
estuaries with high residues Identified in the mollus-
can program continue to present a picture of pollution
problems. There is evidence in some areas that re-
strictions on the use of DDT have led to increased
frequency and magnitude of some other pesticide
residues.
Mercury residues occur over a broad geographic area.
About 65% of the analyses were positive, but residues
were usually small as is to be expected in juvenile
specimens. There are still too few data for cadmium
and lead residues to draw any conclusions.
A recent review of the data indicates that samples
of yearling fish collected prior to their first spawn-
ing provide the most consistent picture. Future col-
lections will be made on this basis. The annual
sample collection dates will vary with the geographical
region.
It is proposed to reactivate the molluscan monitoring
program in the spring of 1976. One collection will be
made at each of the original stations to determine
whether significant changes in residues have' taken
place since 1972.
EPA-NOAA OCEAN FISH MONITORING PROGRAM
The identification of unacceptable residues in some
commercially important marine fish indicated the need
for a "problem seeking" effort rather than a randomized
monitoring program. The potential costs of sample col-
lection made it obligatory that such a project be
combined with some on-going program already funded for
the operation of vessels and collection of fish on the
high seas. In 1971, an Agreement of Understanding was
2
13-4
-------
reached between NOAA and EPA which recognized their
mutual interests in identifying pollution problems in
commercially important marine fish and their food
supply.
The Fishery Centers of NOAA's National Marine Fisheries
Service engage in annual surveys of bottom fish
resources on the continental shelf. In 1973, arrange-
ments were completed to freeze surplus fish from these
trawling operations and use them in the sampling
program.
Fish were frozen when caught and stored at the Fishery
Centers. Livers were removed from the fish and aliquots
of 20 pooled samples were dry-processed and shipped to
the EPA Pesticides Monitoring Laboratory in Mississippi
for analysis.
During 1973 and 1974, about 3,500 fish were distributed
to 175 samples and analyzed in replicate. Station
locations ranged from Georges Bank, 300 miles off the
New England coast, to South Carolina; the north-central
Gulf of Mexico; and from the Bering Sea in Alaska south
to Los Angeles, California. Bottom depths ranged down
to about 1,000 feet.
For the most part, residues were confined to DDT and
PCB and were in the low ppb range. Notable exceptions
were the area west of Los Angeles, the mouth of the
Columbia River area, and inexplicably, Georges Bank.
In some samples from these areas, sDDT and PCB residues
were in the 1-10 ppm range. Interpretation of these
large residues is complicated by the absence of in-
formation on the age of the fish. Repeat collections
are now underway (1975) in both inshore and offshore
areas; only juvenile specimens will be analyzed. These
data should indicate whether residues are being accumu-
lated currently and whether they represent new inputs
from terrestrial sources or a recycling of persistent
chemicals in the deeper offshore waters.
The magnitude of sDDT and PCB residues found in these
surveys demonstrate a need for continued surveillance
of commercial fish populations and the organisms they
prey on in order to safeguard this important segment of
man's food supply.
SUMMARY OF THE FOOD AND DRUG ADMINISTRATION'S
PESTICIDE SURVEILLANCE ACTIVITIES
The pesticide surveillance activities of the Food and
Drug Administration are geared towards FDA's statutory
responsibility of enforcing pesticide tolerances for
food established by the Environmental Protection Agency,
This is accomplished by sampling and testing individual
shipments of food for pesticide residues. In addition,
FDA gathers information on levels of pesticide residues
and other contaminants in the U.S. diet in order to
detect emerging problems and trends and to obtain a
national overview of pesticide residue levels in food.
The pesticide surveillance activities of FDA have been
in effect since the early 1960's. These activities
have been refined over the years by expanding coverage
of chemical contaminants, improved statistical design
of the sampling, and improved analytical techniques and
sensitivity.
National Food and Feed Surveillance Program: This
activity determines pesticide residue levels of indivi-
duals food commodities on a geographical basis using a
statistical sampling plan for food at point of origin.
Each FDA field office samples and examines major food
items to determine levels of pesticide residues. All
samples are examined by FDA's multiresidue methodology
which can simultaneously determine over 100 chlorinated
and organophosphate pesticide residues. Additionally,
selected samples are examined for other chemical
residues.
Total Diet Study: This activity involves examination
of foods as they are prepared and ready for consumption.
A total of 117 items are included In each total diet
sample. Samples are collected from 4 geographic
regions of the United States and sampling sites are
chosen from different cities representing Standard
Metropolitan Statistical Areas. Thirty total diet
samples are collected annually. The food items col-
lected are composited into various food classes, e.g.,
meat, fish and poultry; dairy products; leafy vege-
tables, etc. All total diet composites are examined
for chlorinated organophosphate and carbamate pesti-
cides, PCB's and various heavy metals.
General Findings
Over the years, chlorinated compounds have been the
most frequently found residues. DDT, dieldrin and
PCB's represent the most frequently found compounds
in those foods susceptible to environmental contamina-
tion, e.g., fish, foods produced by animals. Except
for dieldrin, all compounds detected have been shown
to be well below the WHO/FAO acceptable daily intake.
More recent data indicated some decline in the level
and instance of DDT compounds and PCB's.
Assessment of the Program with Reference to Problems
of Technology or Interpretation
The main problem with FDA pesticide surveillance
activities is lack of resources and lack of multi-
residue methods for chemical classes other than
chlorinated and organophosphate pesticides.
Identified Future Trends or Modifications
FDA will try to gear its programs towards environmental
chemical contaminants that may present future food
contamination problems and, where necessary, expand
its routine sampling activities to cover new
contaminants.
THE NATIONAL MEAT AND POULTRY MONITORING PROGRAM
Managed by the U.S. Department of Agriculture and is
largely regulatory in nature and has as a major
objective the removal of products containing violative
residues form the Nation's markets. This program
involves about 7,200 Federally inspected slaughtering
and processing plants and in 1972 about 25,000 tissue
samples were taken for chemical analysis.
THE NATIONAL FISH AND WILDLIFE MONITORING PROGRAM
Freshwater Fish Program
The original program of sampling fish from 50 stations
in the continental United States was started in 1967.
In 1970 this was expanded to 100 stations. Certain
rivers and areas have been identified as "hot spots"
with unusually high residues or the occurrence of
specific pesticides, metals or other organic contamin-
ants such as phthalate esters, PCB's, etc. These
findings have promoted research studies or investiga-
tive monitoring of concern to fishery resource manage-
ment and also stimulated laboratory research on
toxicological significance of specific pesticides and
development of better analytical methodology. Some
trends in residue levels are indicated but the degree
and significance is under evaluation.
3
13-4
-------
Mallard and Black Duck Program
Duck wings from all parts of the United States are
available for monitoring purposes as a byproduct of a
nationwide survey of waterfowl productivity in which
cooperating hunters mail thousands of wings to central
points for biological examination. Of the array of
species whose wings are available for monitoring,
those of the mallard and black duck were selected be-
cause their combined range covers the United States
and they are relatively abundant.
In a special study made for the purpose of determining
the relationship of pesticide residues in wings to
other parts of the body, highly significant correla-
tions were shown between DDT residues in wings and
those in skin, muscle, kidney, pancreas, brain,
gonads and liver. On the strength of these findings
and a pilot study, duck wings were considered
appropriate for monitoring and a full-scale effort was
started in the winter of 1965-66.
A sample for anlaysis consists of a pool of 25 wings
collected in a given State and the number of pools is
roughly proportional to the State's harvest. The
flight feathers are discarded and the 25 wings are
chopped and blended. In the monitoring program to
date, DDE proved to be the predominant residue
throughout the survey; dieldrin was detected in wings
from most states; heptachlor epoxide was reported in
trace levels from one-third of the States; a decline
in residues has been observed in DDE, dieldrin and
PCB's in certain flyways in recent samples. This
study is repeated every third year.
Starling Program
Starlings are collected especially for use in the
monitoring program. Their abundance and general dis-
tribution make it possible to sample them according
to a specially designed plan. Basically the sampling
design consists of 5 degree latitude/longitude blocks
covering the contiguous 48 States with up to four
randomly selected collecting sites within each block.
In actual practice this has resulted in a pool of ten
starlings collected in the fall of even years at
approximately 125 to 130 sites. Birds are trapped or
shot and immediately frozen. Beaks, feet, wings and
skin are removed and the remaining ten carcasses are
ground together. Allquots are taken for analysis of
chlorinated hydrocarbon pesticides and PCB's. In odd
years about half the stations are resampled for
metal analysis.
Stratification of sediment sampling is related to the
bodies of water, i.e., streams traversing the
installations, streams originating on the installation,
and impounded bodies of water.
Fish sampling is restricted to two categories broadly
designated as either bottom feeders or top feeders.
These samples are collected from the same bodies of
water designated for sediment sampling.
Generally the starling is designated as the terrestrial
vertebrate indicator of pesticides in the environment.
Provisions are made for special environmental sampling
where pesticide involvement is suspected or other
circumstances warrant such action.
Preliminary evaluation of limited data demonstrate the
high probability that the stratified sampling design
and the acquisition of all samples from one location
will produce a realistic pesticide profile of that
location.
THE MIREX MONITORING PROGRAM
Is also managed by the U.S. Department of Agriculture
and was designed to evaluate environmental residues
resulting from aerial application of Mirex to control
the imported fire ant. Environmental components from
the monitoring sites include soil, vegetation, water
sediment, fish, crayfish, birds and small mammals.
The program is limited to the 9 southeastern States
in which control efforts are concentrated.
DEPARTMENT OF THE ARMY MONITORING PROGRAM
Limited environmental sampling was initiated in 1972
and 1973. A preliminary analysis of the early data
demonstrated the need for a stratified sampling pro-
gram which was initiated in 1975 when soils, sediments,
fish and birds were designated for scheduled sampling
from 34 installations.
The soil environment has been stratified into three
broad subsets of highly specialized uses. The first
group is further subdivided into four further subsets
designated as sewage treatment, sanitary landfills,
pesticide shops and pesticide storage areas. The
second group involves rather high intensity human
occupancy areas, such as shops, offices, classrooms,
etc.; recreational areas, and golf courses. Golf
courses have been specifically separated from other
recreational areas in view of the specialized mainten-
ance operations requlreing both intensive and exten-
sive pesticide use. The third land use area is
generally further subdivided into military range and
training areas; agricultural areas; and grazing areas.
4 13-4
-------
THE DEVELOPMENT OF A NUMERICAL MODEL TO PREDICT POLLUTANT CONCENTRATIONS
DURING FUMIGATION CONDITIONS WITH AN ONSHORE FLOW
Harvey S. Rosenblum
Thomas F. Lavery
Bruce A. Egan
Environmental Research § Technology, Inc.
Concord, Massachusetts
Introduction
New industrial and electric power generating fa-
cilities are now constructing stacks typically taller
than ISO m for the discharge of air contaminants into
the atmosphere. These and taller stacks have general-
ly eliminated high ground-level ambient pollutant con-
centrations expected during "normal" meteorological
conditions. However, these tall stacks have not al-
ways eliminated high pollutant concentrations which
may occur during plume fumigation conditions. Hence,
in assessing the potential environmental effects of
new sources, it is necessary to accurately estimate
ground-level concentrations over short averaging times
during the fumigation process.
The prediction of hourly pollutant concentrations
due to point source emissions has usually been per-
formed using a version or some modification of the
Gaussian plume equation. This model of atmospheric
diffusion successfully predicts concentrations during
relatively steady meteorological events. However, it
does not consider the spatial variations of wind speed
and direction and turbulent mixing rates encountered
in fumigation phenomena and does not realistically
simulate the time-dependent nature of most fumigation
episodes. For these reasons and because of the gen-
eral need for better prediction of air pollution
concentrations, a numerical advection-diffusion model,
capable of including time and spatially varying velo-
cities and dispersion parameters, has been developed
to simulate atmospheric dispersion during fumigation
conditions associated with onshore flow.
This paper describes the numerical diffusion
model as applied to fumigation associated with onshore
flow and the validation of the model using observed
sulfur dioxide (SO2) concentrations and concurrent
meteorological conditions. In particular, the numer-
ical algorithm, the meteorological input data and
their assumptions, and the validation of the model
using observed meteorological and air pollution data
are described in the following paragraphs.
Summary
An adaptation of a numerical advection-diffusion
model1 is used to predict ground-level pollutant con-
centration experienced during plume fumigation, re-
sulting from onshore flow2.
The numerical solution is generated using mete-
orological and emissions data as input. The model
simulates the actual time and space-dependent charac-
teristics of fumigation conditions and produces a time
series of ground-level cross-wind integrated (CWI)
concentrations from which time-averaged as well as
peak concentrations can be calculated. The CWI con-
centrations are then related to ambient concentrations
by the calculation of appropriate plume widths which
are given as functions of the vertical potential
temperature gradient (3e/3z) in the stable air above
the inversion, vertical directional wind shear, stable
air plume rise and special source configurations such
as multiple stacks.
The vertical diffusivity, Kz, is calculated as a
function of height by a method proposed by Egan and
Mahoney3. The horizontal wind varies with height and
may be input externally as a time-varying quantity or
be generated internally. This last feature allows the
option of simulating actual cases or searching for
expected "worst case" conditions. A thermal internal
boundary layer (TIBL) profile is specified as a func-
tion of the square root of inland distance4, yielding
a distribution of Kz which varies with distance from
the shoreline. Steady-state concentrations can be
calculated since cases of local onshore flow can
persist for up to eight hours.
The lake breeze calculations have been validated
using a set of meteorological and air quality data
gathered during a joint ERT-University of Wisconsin
field program near Lake Michigan. Model performance
is good for cases during which representative data
were collected.
Description of the Fumigation Model
This section describes the theoretical basis of
the model and the data inputs required. These inputs
consist of the specification of meteorological param-
eters which simulate the fumigation episode and emis-
sions parameters which describe the physical charac-
teristics of the pollutant source.
The Numerical Solution
The fumigation model, called FUMIG, is an adap-
tation of a two-dimensional advection diffusion model
developed by Egan and Mahoney1. The model numerically
solves the tracer equation, which describes the change
of concentrations resulting from horizontal advection,
vertical diffusion, and source emissions:
¦5^- B _I| + JL_ (V IX.-) + 0 fi \
9t u 3x 3z ^Z3zJ 4 tlJ
where
X is the pollutant concentration
U is the horizontal wind
is the turbulent diffusivity
Q is the source emission
x,z are horizontal and vertical Cartesian
coordinates
t is time
To simulate pollutant dispersion across and downwind
of a source, a vertical cross-sectional region en-
closing the source and region of interest is divided
into a number of grid elements.
1
14-1
-------
The model simulates the advection and diffusion
of emissions from an elevated source with a forward
time step, finite difference algorithm. A major fea-
ture of the procedure is the suppression of "pseudo-
diffusive" errors associated with conventional finite
difference approximations to advective transport.
This provides an important improvement in the accuracy
of simulating pollutant transport numerically where
spatial and time variations of winds and diffusivities
are of major importance.
The material-conserving computation procedure in-
volves iterations of the zeroth, first and second mo-
ments of the concentration distribution within each
grid element with time. The scheme fundamentally con-
serves pollutant material and descriptive statistics
of the material distribution. The theoretical basis
for the model is described in detail by Egan and
Mahoney1.
Procedurally, the model first develops the wind
and diffusivity fields based on the meteorological in-
put. The model then applies these fields to the nu-
merical simulation procedure defined by the finite-
differenced tracer equation and the two-dimensional
grid system.
Description of the Fumigation Phenomena
The term fumigation was first used during the
1940's to signify the downward mixing of pollutants
which had accumulated aloft in stably stratified air.
It was observed during the gradual burning-off of a
nocturnal radiation inversion. In the 1960's, Lyons
and other investigators began to study similar effects
that occurred with onshore flow along Lake Michigan.
The inversion dissipation process produced high sur-
face pollutant concentrations for perhaps 30 minutes
or an hour. The fumigation which occurs with onshore
flow is a dynamic process which could occur for sev-
eral hours.
During the spring and summer months, the mean
water (lake or ocean) temperature lags behind the
simultaneous land temperature. The relative coldness
of the lake* surface temperature tends to stabilize
the air mass over the lake. With onshore flow during
the warm season, an air mass moving from over the lake
is rapidly modified as it moves inland on the lee
shore. A thermal internal boundary layer (TIBL) rises
from the heated land surface and the stable cold air
flows over the unstable boundary layer. Figure 1
depicts a temperature cross section measured during
onshore flow from Lake Michigan. It is during this
type of meteorological regime that continuous or
dynamic fumigation of a plume from an elevated source
near the shoreline could occur.
During this fumigation, a plume is emitted di-
rectly into the stable layer associated with the on-
shore flow (when the top of the TIBL is below the
stack height) or the plume has enough buoyancy to
penetrate into the stable air mass. Once in the
stable air, the plume is transported downwind and only
slowly spreads (the diffusion is usually described by
Pasquill-Gifford Stability Class F) in the horizontal
and vertical. At some point downwind the plume inter-
sects the rising TIBL and quickly fumigates to the
ground, with resultant high pollutant concentrations.
*The onshore flow fumigation phenomenon will be
described as occurring with a lake breeze through-
out the remainder of the paper.
Figure 2 is a schematic of the continuous or
dynamic fumigation process (Lyons and Cole, 1973)5
that occurs during onshore flow during the warm
months. The dissipation of a nocturnal inversion only
occurs for an hour or so, while onshore flow fumiga-
tion may continue for as many hours as the plume is
intersected by a TIBL.
Ground-level pollutant concentrations that occur
during dynamic fumigation processes vary considerably
with meteorological variables, the location of the
stack and the stack parameters themselves. The plume
azimuth from the stack and the downwind impact point
are a complex function of the vertical distribution of
wind speed and direction, TIBL slope and height,
diffusion characteristics of the stable air above the
TIBL, and plume rise. The wind direction generally
rotates throughout the daytime causing the "fumigation
spot" to veer with time.
Often the TIBL height will not reach the plume
height, and fumigation is thus precluded. On the
other hand, the depth of the boundary layer may be
sufficient to trap the plume completely, and the sub-
sequent plume dispersion could be simulated by a trap-
ping model. Continuous fumigation will occur more
readily with gradient, onshore flow than with a lake-
breeze situation. In fact, only in the deepest and
best developed lake breezes will the fumigation from a
tall stack's plume result in surface concentrations as
high as those found during simple gradient onshore
flow.
Carson® has developed a model that describes the
important meteorological properties of the lower tro-
posphere when an elevated inversion is present. The
model is illustrated in Figure 3. He describes four
layers of differing stabilities and mixing character-
istics. A deep non-turbulent stable layer exists
above the inversion in which 36/3z > 0. Separating
the turbulent layers below the inversion and the
stable air above is the interfacial entrainment layer.
For our purposes, this region will be considered as a
step discontinuity having little impact on the dis-
persion of elevated plumes. Below the entrainment
layer is a turbulent free-convection layer extending
upward from the top of the surface boundary layer.
The mixing in this layer is buoyancy dominated by
turbulent convective elements of large vertical length
scale. 36/3z = 0 here. The layer closest to the
ground is the surface layer characterized by a forced
convection regime in which 36/3z < 0. It extends
throughout the surface boundary layer or about 30 to
50 meters above the ground.
The heat flux of the lower boundary, given by
(9'w') where 6' and w' are the local departures of
potential temperature and vertical velocity, is im-
portant in determining the intensity of convection in
the "free" layer and the rise of the TIBL base.
Further theoretical studies on the nature of the
growth and maintenance of unstable layers capped by
inversions may yield new information regarding the re-
lationships between quantities. However, it is felt
that Carson's model contains the necessary features
appropriate to the present problem.
Model Input Requirements
Since FUMIG requires information on the horizon-
tal wind, U, and vertical diffusivity, K2, fields to
carry out the numerical advection-diffusion calcula-
tion, it is important to describe these quantities
consistently with the observed meteorology.
2
14-1
-------
Horizontal Wind Speeds
When upper-air wind speed data is available, a
profile can be directly used as input. When not
available, it is assumed the wind speed increases to a
maximum value at a height of 50 to 100 m above the
ground and then remains constant through the rest of
the model grid.
Vertical Diffusivity
The diffusivity, Kz, is a measure of the degree
to which small-scale eddies promote vertical mixing.
In actuality, the vertical component of Kz is composed
of two quantities that measure the effectiveness of
vertical momentum and heat transfer, Km and Kjj. They
are defined as follows:
3u/3z
*h
U*8*
W5T
where
local friction velocity = (ts/p) ^ ,
is the surface shear stress and p is
density of air
H
characteristic temperature = -u~, ,
Cp = the specific heat capacity of air
at constant pressure, H = turbulent
heat flux
We assume that the diffusion of pollutants is
better described by the thermal diffusivity Kft than
by the momentum diffusivity Kra. For neutral and
stable conditions it is generally assumed that
^h " Km-
In the stable air above the inversion base Kz
will be near zero since plumes in stable air undergo
minimal dispersion. In fact, typical values of the
vertical plume spread, az, indicate that assigning the
plume depth to a single vertical grid of SO m is a
conservative estimate of the real vertical dispersion.
Below the inversion base in the unstable air, the
method of Egan and Mahoney^ is followed:
Kz(z)
L1
C2)
where Kn is the profile of Kz in a neutral atmosphere.
The factors fj, £2 and are, respectively,
f (1 - - °-) ^ in forced convection regime
1 w
for z < 50 b
l/3
» 4.0|z/L| in free convection regime for
50 < z < TIBL
1 - T in unstable air for
U
z < TIBL
0 for z TIBL
11/3
* (*)
TIBL
0 for z > TIBL
The Kh profile is obtained from Blackadar' as given by
Egan and MahoneyS. The quantity L is the Monin-
Obukhov length, from Monin and Obukhov®. It is de-
fined in the surface layer as
0U
where
0 0 for stable con-
ditions and L = « for neutral atmosphere. Thus, L is
defined for a given time as a function of the surface
heat flux and friction velocity, (6'w') is input
externally, and u0* is estimated as
u * » 0.05 u
0 g
where is the geostrophic wind.
The Estimation of the TIBL Height
The TIBL height is governed primarily by the
distance away from the shoreline. However, Van der
Hoven^ has observed that the land-lake temperature
difference, A6, and wind speed, U, also play an im-
portant role. Bierly^ and Van der Hoven have noted
that for given meteorological conditions, the rela-
tionship between TIBL height and inland distance can
be approximated by
TIBL = Ax
1/2
(5)
where
A « constant which is a function of A6, u
x = distance from shoreline
Measurements confirming the adequacy of Equation (5)
are few; however, those due to Lyons2 are encouraging.
On Table 1 a comparison of observed TIBL heights with
those approximated by a square root law shows that for
distances beyond 1 to 2 km, the differences between
the two are less than 10 to 15%.
Weisman and Hirt4 have also measured the shape
and height of TIBL's on the northern shore of Lake
Erie. They have shown that the boundary top is semi-
parabolic with vertex at the shoreline, following a
power law of the form
y « Axn,
¦where n averaged 0.50 +_ 0.03. Hence, in the fumiga-
tion model the TIBL was assumed to vary as the square
root of inland distance. This permits the description
of the TIBL in terras of A only.
The TIBL height is an important factor in worst-
case estimates of pollutant concentrations. The value
of A in Equation (51 which yields the worst-case can
be determined by repeating the model calculations for
a range of TIBL heights and then calculating the re-
sultant peak ground-level concentrations. In this
manner, the worst-case TIBL profile can be readily
identified.
3
14-1
-------
TABLE 1
where,
COMPARISON OF TIBL PROFILES 6/28/74, WAUKEGAN, ILLINOIS
OBSERVED AND APPROXIMATED BY SQUARE ROOT LAW
yf '
and H
Downwind
Distance
X (km)
1
2
3
4
5
6
7
8
9
10
IX
12
13
14
15
16
17
Observed
TIBL W
30
30
225
300
325
350
375
425
500
500
500
500
500
500
525
550
625
Approximated
By H = 4.75 X*/2
W
141.4
212.4
260.2
300.4
335.9
367.9
397.4
424.9
450.6
475.0
498.2
520.3
541.6
562.0
581.8
600.8
619.3
yo
= (Ah)
+ R 0 + H tana
(7)
Difference
Cm)
111.4
182.4
35'. 2
.4
10.9
17.9
22.4
- .1
-49.4
-25.0
- 1.8
20.3
41.6
62.0
56.8
50.8
- 5.7
yo
H tana
initial plume width due to separation of
sources, i.e., multiple stacks
predicted stable plume width, i.e.,
= ax^ (see Table 2)
factor which relates ayf to observed a^'s in
the presence of significant vertical direc-
tional shear, R = MA + b where A is maximum
change in wind direction in the layer from
stack height to final plume rise, in radi-
ans. A and b are determined from field
data. Note that R = 1 for average shear
conditions.
"spreading" of plume during fumigation;
average TVA^2 value of a = 25.3° is used
for calculation.
TABLE 2
COMPARISON OF TIBL PROFILES 8/8/74, OAK CREEK, WISCONSIN
TVA
12
DISPERSION PARAMETERS USED IN CALCULATING
HORIZONTAL PLUME DISPERSION
Observed
H = 4.47 X 11
1
20
141.4
121.4
Stability Class
I2-, °K (100 in)"1
0 Z
a
2
200
199.9
- .1
3
250
244.8
- 5.2
Neutral
0.00
72.76
4
300
282.7
-17.3
Slightly Stable
0.27
67.69
5
300
316.1
-16.1
Stable
0.64
62.28
6
300
346.2
-46.2
Isothermal
1.00
60.73
7
325
374.0
-49.0
Moderate Inversion
1.36
59.15
8
375
399.8
-24.8
Strong Inversion
1,73
58. SO
9
425
424.1
.9
10
525
447.0
78.0
Based on the findings of Lague13 and Lyons2,
.758
.690
.633
.560
.503
.457
following scheme was developed for use in the lake
breeze model:
Plume Rise
All plume rise calculations are based on Briggsl1.
These calculations are appropriate only to situations
where the atmosphere is homogeneous in stability
through the depth between stack height and final plume
rise.
In the simulation of lake-induced fumigation
sometimes a plume is emitted below the TIBL but may
have sufficient buoyancy and momentum to penetrate the
TIBL and disperse in the stable air. This is assumed
to occur when
1)
For X < X
2) At X = Xo
3)
For X > X
TIBL < h +
(Ah).
where (Ah)s is the stable plume rise and hs is the
stack height. Otherwise the plume is trapped below
the TIBL.
Determination of Plume Width, Qy
Since the numerical model predicts cross-wind
integrated concentrations, it is necessary to relate
the model predictions to actual concentrations. The
relationship between concentration, x» and CWI is
diffusion in stable air is assumed
0^ = axb where a and b are deter-
mined from Table 2 as a function
3 0
of r— in the stable air below.
az
X is downwind distance, Xe is the
value of X where the emitted
plume first intersects the TIBL.
is described above.
Class "C" stability assumed such
that a virtual point source would
have Oy = ayf at X = Xe. Thus,
the "C'1
distance X - Xv where Xv is the
x =
CWI
/2rr* a
(6)
yf
y — "-'yf ai /\ — Ag •
ay used as appropriate for
X -
downwind distance of the virtual
point source. The relationship
used is oy = .139 (X-Xv)-903.
Application to Lake Michigan Cases
The lake breeze model was validated by comparing
observed SO2 concentrations to the FUMIG model predic-
tions. The observed concentrations were obtained by
Lyons2 during a series of comprehensive field experi-
ments at two sites near the Lake Michigan shoreline.
Data on ground-level SO2 concentrations resulting from
power plant emissions were obtained from both fixed
and mobile monitors. A mobile van was utilized to
4
14-1
-------
seek out and measure the location of maximum concen-
tration, the so-called fumigation spot. The spot
moves from place to place in response to changing
wind and turbulence conditions during the day. As a
result of the transitory nature of the lake breeze,
high SO2 concentrations are usually only briefly ob-
served at any one location.
Meteorological input data was obtained from
wiresonde, pilot balloon and instrumented helicopter
observations. The comparison of calculated and ob-
served ground-level concentrations is made in Table 3.
The results indicate that the fumigation model predic-
tions of hourly SO2 concentrations are consistent with
the observed ten-minute average concentrations. The
inclusion of wind shear (as observed on both days] im-
proves the model calculations.
TABLE 3
COMPARISON OF PREDICTED AND OBSERVED GROUND-LEVEL
S02 CONCENTRATIONS (ppm)
Date
6/28
8/8
Distance
(km)
13.8
11.6
Observed
Peak
10-Minute
Average
0.15
0.56
Predicted Predicted
Hourly Hourly
Average
(No Wind
Shear)
0.16
0.75
Average
(20° Wind
Shear)
0.11
0.51
3. Egan, B. A. and J. R. Mahoney, 1972b: Applica-
tions of a Numerical Air Pollution Transport
Model to Dispersion in the Atmospheric
Boundary Layer, J. Appl. Meteor., 11,
1023-1039.
4. Weisman, B. and M. S. Hirt, 1975: Dispersion
Governed by the Thermal Internal Boundary
Layer, 68th Meeting APCA.
5. Lyons, W. A. and H. S. Cole, 1973: Fumigation
and Plume Trapping on the Shores of Lake
Michigan During Stable Onshore Flow, J.
Appl. Meteor., 12, 494 pp.
6. Carson, D. J., 1973: The Development of a Dry
Inversion-Capped Convectively Unstable
Boundary Layer, Quart. J. R. Met. Soc., 99,
450-467.
7. Blackadar, A. K., 1962: The Vertical Distribution
of Wind and Turbulent Exchange in a Neutral
Atmosphere, J. Geophys. Res. 67, 3095-3112.
8. Monin, A. S. and A. M. Obukhov, 1954: Basic Laws
of Turbulent Mixing in the Ground Layer of
the Atmosphere, Akad. Nauk SSSR, 151,
163-187.
9. Van der Hoven, I., 1967: Atmospheric Transport
and Diffusion at Coastal Sites, Nuclear
Safety, 8, 490-498.
10. Bierly, E. W., 1968: An Investigation of Atmo-
spheric Discontinuities Induced by a Lake
Breeze, Ph.D. Dissertation, Univ. of
Michigan, Dept. Meteor, and Oceanogr.,
150 pp.
References
Egan, B. A. and J. R. Mahoney, 1972a: Numerical
Modeling of Advection and Diffusion of Urban
Area Source Pollutants, J. Appl. Meteor.,
11, 312-322.
Lyons, W. A., J. C. Dooley, C. S. Keen, M. A.
Schuh, K. R. Rizzo, 1975: Detailed Field
Measurements and Numerical Models of SO2
from Power Plants in the Lake Michigan
Shoreline Environment, Appendix E to ERT
Doc. P-1020-PP.
11. Briggs, G. A., 1969: Plume Rise, Critical Re-
view Series (TID-25075), Atomic Energy
Commission, Division of Technical Informa-
tion, Oak Ridge, Tennessee.
12. Tennessee Valley Authority, 1970: Report on
Full-Scale Study of Inversion Breakup at
Large Power Plants, Research and Develop-
ment, Muscle Shoal, Alabama, March 1970.
13. Lague, J. S., 1973: Observed Entrainment in a
Power Plant Plume, Master's Thesis, Mass-
achusetts Institute of Technology, Depart-
ment of Meteorology, May 1973.
POTENTIAL TEMPERATURE, °K 5 AUGUST 1971
Figure 1 Potential Temperature Cross-section Following an East-West
Transverse of Lake Michigan Shoreline at Oak Creek,
5 August 1971
6000
L 5000
5 4000
0
5
uj 3000
< 2000
it
£ '000
1
SURFACI
5
14-1
-------
, CMCULAHD to, fkoruc AT GROUND IIVII
Figure 2 Schematic of Classic Continuous Shoreline Fumigation
by Lyons and Cole (1973)
Stable Layer
Inversion Base
hi ( t)
Interfacial Entrapment Layer
z
Turbulent Free-Convection Layer
Surface Layer
( Forced Convection)
<0
9
Figure 3 Schematic of Inversion-Capped Convectively Unstable Boundary Layer
(after Carson, 1973)
6
14-1
-------
MULTIPLE-TRACER HIGHWAY DISPERSION STUDY
Walter F. Dabberdt
Atmospheric Sciences Laboratory
Stanford Research Institute
Menlo Park, California 94025
Summary
A series of multiple-tracer dispersion experiments
was conducted on the six-lane Bayshore Freeway (U.S.
101), approximately 60 km southeast of San Francisco.
These experiments are part of a larger study that is
directed toward the investigation of the interactive
role played by traffic, meteorology, and roadway config-
uration on air flow and air quality on and near road-
ways. In the current study, dispersion has been inves-
tigated at a simple highway site.
The experimental site was a straight, level, at-
grade section of U.S. Highway 101 with six lanes and an
ADT in excess of 100,000. Aerometric instrumentation
was supported by five 15-m towers: two were located
30 m to either side of the highway edges, two at 10 m,
and one in the median. Meteorological instrumentation
included 11 three-component anemometers, seven propeller
vanes, two four-level temperature systems, and a pyrano-
meter. These data were sampled continuously at a fre-
quency of 20 per second; 15-min summaries were computed.
Traffic measurements included 15-min frequency distri-
butions of vehicle speed, volume and axle number on a
lane-by-lane basis. Average hourly pollutant concen-
tration measurements were made at 35 locations out to
95 m from the roadway edges and up to 13 m above the
roadway, and were of two types: (1) carbon monoxide,
total hydrocarbons and methane, and (2) two trace
gases—SFfe and F13B1. Both of the trace gases were
released by each of two control vehicles driven con-
tinuously in a loop past the test site. Sulphurhexa-
fluoride was released exclusively in the westbound
lanes and freon in the eastbound direction.
Cross-roadway temperature gradients ranged up to
2.5°C at the 2-m level, 1.5°C at 3.8 m, and 0.75®C at
7.5 m. The cross-roadway wind speed component showed
significant positive correlation (r) with the hori-
zontal temperature gradient, ranging from 0.51 with
near-parallel winds to 0.82 for winds normal to the
roadway. Variations in the cross-roadway gradient of
turbulence intensity were investigated in view of
corresponding meteorological and traffic parameters.
In these tests, turbulence differences between upwind
and downwind sides of the road were "well" correlated
with only the cross-roadway temperature gradient
(r = 0.53) . Turbulence differences between upwind and
median strip locations also correlated well with only
one parameter—the cross-road wind speed component
(r = 0.52)
The ratio of the diffusion parameter for the trace
gas released from the upwind lanes to that released in
the downwind lanes for each test ranged over an order
of magnitude—from about 0.3 to 3.0. Variations in the
ratio were examined against corresponding changes in
ambient meteorological conditions and traffic para-
meters (e.g., volume, speed, occupancy). The highest
correlation (r » -0.35) was found with wind speed
alone; correlations with traffic parameters were gen-
erally less than 0.2. Individual diffusion values
(i.e., upwind or downwind road segments alone) corre-
lated well (r = 0.84 and 0.80) with the ratio of the
standard deviation of vertical wind fluctuations to wind
speed; inclusion of traffic volume in the ratio did not
change the correlation.
The tracer release and concentration data were used
with the carbon monoxide (CO) concentration data to com-
pute the CO emission rate of the mix of vehicles on the
highway. Computed values ranged up to 0.03 g nT^s'l
for each traffic stream. Corresponding emission values
were computed from a cruise-mode emission model. The
ratio of "tracer" to model values was 1.015, while the
average normalized difference between the two methods
was 35 percent.
Experimental Design
The experimental study was conducted January-
February 1975 (see Table 1) on a stretch of U.S. High-
way 101, in Santa Clara, California (about 65 km south-
east of San Francisco). The road is a major intrastate
freeway with three lanes of traffic in each direction
and an annual average daily traffic in excess of
100,000.
This particular site was chosen for several reasons
related to the overall objective of trying to study the
effect of traffic flow on near-roadway pollutant disper-
sion. First, the site is relatively simple and quite
homogeneous; within an 0.75-km radius of the sampling
location, the land is quite flat and consists mainly of
level fields with a low growth of grasses. The site
Table 1
SCHEDULE OF HIGHWAY TRACER TESTS
Number
of
Date
Hours
One-Hour
1975
PST
Tests
Figure Symbols
17 January
1200-2000
8
X
21 January
0500-1300
8
+
24 January
0500-1300
8
0
28 January
0500-1300
8
~
30 January
1200-2000
8
6
5 February
1200-2000
8
V
48
has other advantages: traffic flow is heavy and varies
markedly throughout the day both in speed and volume by
direction; also, the median strip is sufficiently wide
to permit installation of a tower for meteorological
and air sampling purposes.
To meet the experimental objectives, a comprehen-
sive microscale sampling network was established to
monitor wind, temperature, air quality, and traffic.
Figures 1 and 2 illustrate the location and orientation
of the meteorological Instrumentation. All of the 50
meteorological data inputs were sampled, digitized, and
recorded on magnetic tape every 2.5 s. Fifteen-minute
and hourly summaries of both primary and derived
14-2
-------
meteorological parameters were prepared; the format is
illustrated in Table 2. Comprehensive traffic informa-
tion* was recorded through the study and consisted of
speed, axle-number, and volume data, segregated on a
lane-by-lane basis and summarized over 15-min periods.
FIGURE 1 AEROMETRIC INSTRUMENTATION AND SRI MOBILE
ENVIRONMENTAL MONITORING LABORATORY AT
HIGHWAY TEST SITE
Programmable multiple-bag air quality samplers
(AQS) were used at 35 locations (Figure 2) during each
test to monitor hourly concentrations of carboh mono-
xide (CO), total hydrocarbons (THC), methane (CH^), and
two trace gases: sulfur hexafluoride (SFg) and fluoro-
trlbromomethane (F-13B1). Both trace gases were
released from each of two vans: SF was released
6
exclusively in the middle westbound lane and F-13B1 was
released in the middle eastbound lane. Each van typi-
cally made five to seven round trips past the site each
hour. The trace gases were analyzed by a dual-channel
gas chromatograph (GC) with electron capture detectors
coupled to a dual-channel peak integrator;2 CO, THC,
and CH, concentrations were measured using a GC with
4
flame ionization detector. All GC systems were rou-
tinely calibrated on each experimental test day, using
standard calibration gases.
Data Analysis
General
Analysis of the data was concentrated on two major
aspects:
• Impact of traffic parameters on near-roadway
pollutant dispersion, and
• Determination of vehicle CO emission rate.
The first aspect was approached in three ways. First,
the nature of near-roadway temperature gradients and
SRI MOBM.f
LASOftATOP>
SA-2761-46R
uvw 3-component anemometer
pv Propeller vane
T Absolute temperature
AT Temperature differential
O Air sampler
Note: additional air samplers located at ground level on both sides of road
beyond towers at 15.2m intervals to 91.5m from roadway edges.
TOWER 1
pv nO
pv
pv
O
O
TOWER 2
AT, uvw nO
At,uvw
At,uvw
T ,uvw
O
o
TOWER 3
pv nO
uvw
uvw
o
o
TOWER 4
AT,uvw po
At,uvw
At,uvw
T ,uvw
o
o
TOWER 5
pv rjO 14.2m
pv
pv
O 7.5m
O 3.8m
O 2.0m
¦19.8m
-10.7n>-»t
3 Eastbound Median 3 Westbound
Lanes Strip Lanes
-Edge of Active Roadway (36,
,6m)—^— 10.
7m-
19.8m-
FIGURE 2 METEOROLOGICAL INSTRUMENTATION LAYOUT
Physical and logistical support in the Installation
of the cable sensors and construction of barriers to
shield the towers was provided by the California
Department of Transportation, District 04, San Fran-
cisco.
1 Perkin-Elmer Model 3920
2 Spectra-Physics Autolab System IV
3 Beckman Model B6800
2
14-2
-------
Table 2
SAMPLE METEOROLOGICAL DATA SUMMARY
' I HE PER 1001 1345-1400 i
PST
DATE 1
17 JAN
HEIGHT ¦ 14,8 "ETEPS
TOWER 2
TOWER 3
TOWER 4
WIND SPEED (M/S)
3,10
3.02
2.89
WIND 01ft, TRUE
320*
330,
320.
SIGMA UR(M/S)
,92
1.05
1*05
SIGMA VR(M/S)
.61
,59
• 59
SIGMA U (m/s)
.63
,63
• 63
SIGMA V (M/S)
.92
1 ,05
1.03
SIGMA y (M/S)
.37
.36
SIGMA THf | A (0EG)
16.32
20.99
19.36
SIGMA PhJ OfcG)
6,07
7.85
TURH INT, HOR12(M/S)
1.11
1.05
1.20
TUR8 INT « TOTAUM/S)
1.17
1 .26
TEMPEKAIuKE
15.22
15.14
HEIGHT • 7.5 METERS
2.63
WIND SPEED (M/S)
2.18
2.58
WIND UIK, TRUE
323,
331.
320.
SIGMA UR(m/S)
1 .06
1 .09
1.06
SIGMA VR(M/S)
,68
.71
.65
SIGMA U (M/S)
• 59
.89
.74
SIGMA V (M/S)
1.12
.9#
1 .00
SIGMA « (M/S)
,36
.16
.33
SIGMA TH£1 A (DEG)
22.23
26.64
22.78
SIGMA PHI (PEG)
7.72
6.94
6.57
TURB INI, M0RIZ(M/S)
1.26
1.30
1.24
TURB INT, TOTAL (M/S)
1.32
1,31
1.29
TEMPERATURE (C)
15.64
15.36
HEIGHT - 3.A METERS
WIND SPEED (M/S)
2.58
1.68
2.34
WIND U]R, TRUE
318,
338,
317.
SIGMA U«(M/S)
.93
1.05
.99
SIGMA VR(«/S)
.78
.99
.65
SIGMA U (M/S)
.66
1.01
.80
SIGMA V (M/S)
1.02
1 .03
• A7
SIGMA w (M/S)
.35
.45
.26
SIGMA THETA (DEG)
22.01
40.73
23.32
SIGMA PHI coEG)
7.57
18. 1 1
8.88
TURB INT, H0RIZ(M/S)
1.21
1 .45
1,18
TURB INT, TOTAUM/S)
1 .26
1.51
1.21
TEMPERATURE (C)
16.15
15.56
INSOLATION (LY/MIN)
0.000
HEIGHT • 2 MtTFRS
WIND SPEED (M/S)
2.56
1.23
2.00
WIND 01R,TRUE
311,
350.
318.
SIGMA UR(M/s)
.76
.87
1*06
SIGMA VR(M/S)
.72
,97
• 69
SIGMA U (M/S)
• 64
,81
• 87
SIGMA V (M/S)
.83
1,02
• 91
SIGMA W (M/K)
.26
,48
• 22
SIGMA THETA (DEG)
19.95
44,59
35.94
SIGMA PHI (f)EG)
6,06
21.12
10.73
TURB INT, HORIZ(M/S)
1,05
1.31
1*26
TURB INT, TOTAKM/S)
1.08
1.39
1.28
TEMPERATURE (C)
16. *6
15.80
GRADIENT RI (2,0 - 3.8 M)
•34,85
••16
(3,6 - 7,5 M)
-27,29
• •40
(7.5 • 14,? M)
-.53
-•64
BULK RI NUMBER
•48.93
••71
* Similar summaries have been prepared for
Towers 1 and 5.
the relationship to traffic were examined. Second, the
structure of turbulence near the roadway and its varia-
tion with traffic and meteorological parameters were
investigated. And third, the dispersion of the two
trace gases was analyzed. Also, in-situ CO emission
rates were computed from the tracer release rates and
tracer and CO ambient concentrations; these were then
compared with CO emission rates determined from a
cruise-mode emission model.
Temperature Structure
Vehicle influences on near-roadway dispersion can
theoretically arise from one of three physical pro-
cesses: (1) buoyant mixing from atmospheric insta-
bilities created by vehicle thermal exhaust; (2) mechan-
ical mixing from wake turbulence; and (3) transport
from induced drag flow. To examine the first effect,
variations in the cross-roadway temperature gradient
are analyzed as they relate to wind direction and wind
speed, ambient turbulence intensity, vehicle volume and
speed, and height above ground. The cross-roadway
temperature gradient ^Tjlor^z^ is obtained by taking
the temperature difference at each level between the
north (number 4) and south (number 2) towers, after
first normalizing by the 14.2-m values (thus assuming
no perceptible ^j,oriz at t^at level—in fact, this
difference was only on the order of a few hundreths of
a degree). Thus
AT (ref.) = T^ (14.2 m) - (14.2 m)
and
AT
horiz
(i) = T^ (i) - T2 (i) + AT (ref.)
(1)
(2)
where i is the level (2.0, 3.8, or 7.5 m) and the sub-
scripts refer to tower location. Figure 3 shows 15-min
averages of AT, . at 2 m as a function of the aver-
horiz
age cross-roadway wind speed component (uroa(j) ^or
each of the three levels; winds from the south of the
roadway are positive. In these and following data
plots, different symbols are used to identify data from
each of the six test days.
-2
• i ¦ i—'—
I " ! —1 1
V
*
% » *
» x°
A otO V
a
*
A
4 n* '
v* *
x " V » • %
4 •* »*
• • .» *#e 1
^4.
" ,
.«* *° » * X*
" *
X o
~
X
—1 L , 1 ,
HEIGHT * 2.0 m.
. i.i.i.i
-2 -I O I 2 S
CR03S-R0A0WAY WIND SPEED — m/»
FIGURE 3 15-MIN VALUES OF CROSS-ROAD TEMPERATURE
GRADIENT VERSUS CROSS-ROAD WIND SPEED
COMPONENT
The range of cross-roadway temperature gradients
was quite large: at 2 m, maximum values of -1,5 and
2.5°C were measured with northerly and southerly winds,
respectively, across the 57-m tower separation of the
two towers; at 3.8 m the difference ranged from -0.75°C
to 1.5; while at 7.5 it was still moderately large
(from -0.4 to 0.75°C). The difference levels off at
the higher cross-road wind speeds. Figure 4 shows
^horiz 3t ^"m ^eve^ 38 a functi°n °£ the cross-
roadway angle (0__ , where
road
0 . = 110.6°
road
(3)
3
14-2
-------
9 is the vector average wind direction and 110.6 is
the orientation of the road (i.e., "eastbound").
1 ' 1 '
A
A a +
> i * i • i i , i
V
\ »~.- A
V ~ ~ ~
JesL a , . •
a
A A ~
~ ~ +
* ~ © ~ a
~ ~ ~ f
a c
X
— *
~
i i i i
x * °4
Ht\
* X
X
HEIGHT * 2.0 m. K
i 1 i I i 1 ,i -J 1
In spite of the moderating influence of vertical
mixing, some other phenomenon apparently controls the
sign of the cross-road temperature gradient. According-
ly, dependence of At
horiz
on various atmospheric and
vehicle parameters was investigated.
The data in Figure 4 suggest that
AT.
is
horiz
strongly dependent on wind direction relative to the
roadway. Therefore, the data were disaggregated into
six wind direction Categories, as follows:
Cate-
s°ry
1
2
3
4
5
6
Wind Direction (8 ,) Range
road
000.0-014.9 165.0-179.9
345.0-360.0° 180.0-194.9°
015.0-029.9
030.0-044.9°
045.0-059.9°
060.0-074.9°
075.0-089.9°
150.0-164.9 330.0-344.9 195.0-209.9
135.0-149.9° 315.0-329.9° 210.0-224.9°
120.0-134.9° 300.0-314.9° 225.0-239.9°
105.0-119.9
090.0-104.9°
285.0-299.9
270.0-284.9°
240.0-254.9
225.0-269.9°
-too -SO 0 60 100 ISO 200 2SO
CROSS-ROADWAY WIND A MOLE — daqrMs
FIGURE 4 15-MIN VALUES OF CROSS-ROADWAY TEMPERATURE
GRADIENT VERSUS WIND DIRECTION RELATIVE
TO ROADWAY
The cross-roadway temperature structure is of it-
self only an indirect indicator of potential mixing
near the roadway; however, further examination of its
cause may have significance for another reason. Such
examination may reveal whether cross-road temperature
gradients are a result of: (1) the waste heat emis-
sions of the vehicles, (2) differences in the thermal
characteristics of the toadway and adjacent soil, or
(3) mixing of the atmospheric surface layer by roadway
vehicles and subsequent changes in the downwind (verti-
cal) temperature profile.
If the third process was dominant, then under
lapse conditions (i.e., temperature decrease with
height), the effect of vehicle-induced atmospheric
mixing would be to lower near-ground temperatures down-
wind. Downwind temperatures would thus be lower than
their upwind counterparts. With Inversion conditions,
the reverse would apply and downwind temperatures would
be higher than those upwind. However, examination of
the data shows that there is no blmodal distribution by
stability. (Note that all six days were characterized
by both lapse and inversion conditions.) Rather, down-
wind temperatures were virtually always higher than up-
wind temperatures at each of the three heights. We
therefore conclude that this phenomenon is not the con-
trolling factor in the cross-roadway temperature
structure. However, it may explain why southerly winds
(positive cross-road component) have maximum cross-road
temperature gradients that are significantly larger
(1 C greater at 2 m) than northerly winds. In the
study area the local wind flow is controlled by a land-
bay breeze circulation; southerly winds blow at night
with stable or inversion conditions over land, while
the northerly bay breeze blows by day with lapse condi-
tions over land.
For each category, the 15-min average cross-road tem-
perature gradient at 2 m was correlated with each of the
following six independent variables:
1. TTI (upwind)—the total turbulence intensity (TTI)
at 2 m on the tower farthest upwind of the roadway.
It is defined by
TTI s I(u') +
(v1)2 +
(«')]
1/2
W
where u1, v', and w' are the longitudinal, lateral,
and vertical fluctuating wind components, respect-
ively;
2.
road
TTI x
—the cross-roadway wind speed component;
road
4. total vehicle volume;
6
5. £ (volume x speed)—the product of vehicle volume
i=l
and speed summed over all of the six lanes of the
roadway; and
S (volume x speed)
TTI x u
—scaling factor for the dis-
road
persion of heat from roadway vehicles.
The results of the various correlations are summar-
ized in Table 3. Of the six independent variables, the
cross-road wind-speed component has the highest and most
consistent correlation with AT, , . The ambient TTI
horiz
is virtually uncorrelated, and the traffic variables
alone are poorly and inconsistently correlated. In
fact, in no case do any of the five other independent
variables show a higher correlation than u alone.
road
Theoretically, differences in the thermal prop-
erties between roadway and the upwind soil could also
contribute to the variation of the cross-road tempera-
ture gradient. However, if this were the case, then
there would be a diurnal variation in the sign of
^Thoriz ^Ue to t^e different thermal properties of the
roadway and soil. But the data indicate no apparent
diurnal dependence.
4
14-2
-------
Table 3
MATRIX OF THE LINEAR CORRELATION COEFFICIENT BETWEEN THE
DEPENDENT VARIABLE AT, , AT 2 M AND SIX INDEPENDENT
horlz
VARIABLES FOR EACH OF SIX WIND DIRECTION (6 ,) CATEGORIES
road
Wind direction category
1
2
3
4
5
6
Number of data points
20
27
42
35
31
12
Independent Variables:
Correlation
Coefficients
TTI (upwind)
-.10
.06
.06
-.17
0
1
-.07
Cross-road wind speed
.51
.70
.73
.66
.81
.82
^"road^
TTI x u
road
.40
.61
.60
.52
.67
.49
Total vehicle volume
.19
-.57
.05
-.02
.27
.78
6
E (volume x speed)
.22
-.52
.02
-.12
.40
.57
i-1
^(volume x speed)
TTI x u
road
.46
.47
.71
.65
.48
.93
Also, the strong positive correlation of AT
with u
road
horiz
is contrary to intuition. That is, unless
ATTI^
oriz
(i)
TTI4 (i) - TTI2<2)
(5a)
1.0
0.5
-0.5
-1.0 -
-| . 1-
X ~
* o
~
~~~
:
oAomoA o
°A°
-JL
$
T+-
%
* „*
if,
"J. lAO
i
"« f*
a" V
HEIGHT » 2.0 m.
-100
the vertical mixing induced by thermal instabilities
from vehicle heat emissions and vehicle-induced mech-
anical mixing dominates under light wind conditions
thereby effectively dispersing vehicle thermal emissionr
more than under higher wind conditions when more "con-
ventional" dispersion dominates. This could possibly
also explain the apparent leveling off of at
the larger values of uroa(j (see Figure 3) .
This examination of thermal structure revealed
some interesting observations on the magnitude and
nature of cross-roadway temperature gradients. It
pointed out some of the complexities of near-roadway
dispersion, but has not provided a definitive under-
standing of the effects of vehicle motions and thermal
emissions. These will be addressed further in the
following sections in examinations of the variation of
near-roadway turbulence characteristics and trace-gas
dispersion in relation to meteorological and traffic
variables.
Turbulence Structure
Cross-roadway variations in the total turbulence
intensity were investigated to obtain a better under-
standing of the effect of roadway traffic on the local
dispersion of traffic-generated pollutants. The pre-
vious discussion indicated that both thermal and mech-
anical effects of the traffic stream may be important,
and that they apparently combine to produce a direct
relationship between wind speed and local dispersion
(contrary to the indirect or inverse relationship that
exists elsewhere).
Possible vehicle effects on dispersion near the
roadway were examined by looking at the difference in
the TTI between the upwind tower and (1) the median
tower, and (2) the downwind tower. Figure 5 illustrates
the variation of 15-min values of the TTI difference
(ATTI ) between upwind and downwind towers at the
horiz
2-m level as a function of the cross-roadway wind angle
(0 ,) where
road
-50 0 50 100 ISO 200
CROSS-ROADWAY WIND ANOLE — daq>«M
200
FIGURE 5 CROSS-ROADWAY DIFFERENCE IN THE TURBULENCE
INTENSITY AS A FUNCTION OF THE CROSS-ROAD
WIND ANGLE
and i is the level (height) and the subscripts denote
tower location.
Figure 6 shows the TTI difference between the up-
wind and median (number 3) towers as a function of
6 , , where
road
ATTIhoriz (1) ^ mu (1) " 1X13 (1) (5b)
and the subscript u denotes the upwind tower location
(either 2 or 4). There is considerable scatter in the
0.2
-0.2
-0.4
-o.e
-0.8
-1.0 -
-1.2
a
*A
o ~ a
D
1 r"
T
» »»
• w
r. ~
d + + _ * o *
» • '
V \ w
~ a °
a
*" -
X «o' J
~ A X *1
* A
m *
O HEIGHT » 2.0 m.
_» I . I . I L
-SO 0 50 100
CROSS-ROMMMY WIND ANOLE -
ISO
200
280
FIGURE 6
DIFFERENCE IN TURBULENCE INTENSITY BETWEEN
THE UPWIND AND MEDIAN TOWERS AS A FUNCTION
OF THE CROSS-ROAD WIND ANGLE
5
14-2
-------
data at all levels, with a slight inference of a depen-
dence of ATTI on the wind/roadway angle. The large
scatter does suggest the need to consider the depen-
dence on other independent parameters, e.g., vehicle
speed and volume, wind/vehicle orientation, and sta-
bility. The gradient is somewhat larger at the lower
levels, though not significantly. Values range from
+0.5 to -0.5 m s"1 with the largest scatter at the
lower levels.
The TTX difference between the upwind and median
sensors (Figure 6) shows that the two lower levels have
virtually the same distribution: the median tower al-
ways has the greater turbulence intensity, but has large
scatter from 0 to -1.3 m s"l. However, the situation
changes at the 7.5-m level. The horizontal gradient of
turbulence intensity drops by about a factor of two.
This suggests the need to consider vehicle effects on
variations in the TTI gradient. Data from the median
tower do suggest, though, that a uniformly well-mixed
layer is present on the road up to a height of at least
4 m, and then damps out significantly at 7.5 m.
The absolute value of the cross-roadway gradient
of turbulence intensity is possibly affected to a signi-
ficant degree by the magnitude of the TTI of the ambient
flow. In examining vehicle effects on pollutant dis-
persion, it is desirable to examine the cross-roadway
TTI gradient against the ambient level. Figures 7 and
8 present the variation of the normalized horizontal
gradient of turbulence intensity (ANTTI^ against
the cross-roadway wind angle where,
^horiz (i) = *TTIhoriz (i)/TTIu (1>
(6)
, , t —»
' 1 ' I ' J 1 '
~
A
-
~ a
o a
° A
- ~ * *
* ~
W „ .. • X
A A. ~
x ~ . ^
. O *
~ ~ ~
. ¦ -"Jffi
- ~ ~ «
a
~
1 » A/-
X
~
X
X -
i - ¦ 1 ¦
HEIGHT ¦ 2.0 m.
i 1 i 1 i 1 i,l i
-100 -SO 0 BO 100 ISO 200 250
CROSS-ROADWAY WIND ANOLE — d*qrMl
FIGURE 7 WIND DIRECTIONAL VARIATION OF THE RATIO OF THE
CROSS-ROAD TURBULENCE INTENSITY DIFFERENCE
TO THE UPWIND TURBULENCE INTENSITY
Figure 7 shows normalized differences between the up-
wind and downwind sensors at 2 m and Figure 8 shows
them for the upwind and median sensors.
-2
c -J
-8
~i—¦—r
O ~ AO
* 4 A ~
" 7s ° '~ ~
1—'—i—1—r
4 J&C
A
*
-------
• ATTI (1) -- upwind/median
horiz
• ANTTI, (1) -- upwind/downwind
horiz
• ANTTI (1) -- upwind/median
horiz
Independent Variables
• TTI (1) -- upwind turbulence intensity at
u
2 m
• u -- upwind reference wind speed
ref
• u -- cross-roadway wind speed component
road
• AT (1) -- cross-roadway temperature
horiz
gradient at 2 m
• 0CC(E) + 0CC(W) -- sum of eastbound and
westbound (by lanes) vehicle occupancy
• V0L(E) x SP(E) + V0L(W) x SP(W) -- sum of
the eastbound and westbound (by lanes)
products of vehicle speed and vehicle
volume
• V0L(Ii+W) -- sum of eastbound and westbound
traffic volumes
• 0CC(u) -- occupancy of the upwind traffic
stream
• V0L(u) x SP(u) -- product of volume and
speed (by lanes) for upwind traffic stream
• u -- vector sum of ambient wind flow and
net
vehicle-induced drag flow of the upwind
traffic stream
• V0L(u) -- volume of the upwind traffic
stream
• DRAG(u) -- vehicle drag flow of the upwind
traffic stream
• v -- vector sum of roadway-parallel
net
ambient wind component and vehicle-induced
drag flow of the upwind traffic stream.
Table 4 is a summary of correlation coefficients (r)
with the data stratified according to the six wind-
direction categories.
The results summarized in Table 4 show that the up/
downwind gradient of turbulence intensity at 2 m is con-
sistently well correlated with only cross-road tempera-
ture gradient; the average r is about 0.53. This
further suggests that thermal vehicle emissions are the
cause of the large cross-road temperature gradients
observed. Interestingly, the upwind/median gradient of
turbulence intensity is best correlated with the refer-
ence wind speed; the average r is about 0.52. None
of the other independent parameters show any consis-
tently significant correlation. (Note, however, that
this is not unexpected in the case of ATh as it
is defined as the difference between two geographically
fixed locations, whereas the upwind/median TTI gradient
is always taken as upwind minus median values.)
While these analyses have provided some Insight
into the structure and dependence of turbulent
7
fluctuations near the roadway, their impact on pollu-
tant dispersion can be more directly examined through
analysis of the tracer data.
Table 4
MATRIX OF THE LINEAR CORRELATION COEFFICIENT BETWEEN CROSS-ROACWAt
TURBULENCE GRADIENTS (DEPENDENT VARIABLE) AND VARIOUS METEOROLOGICAL
AND TRAFFIC PARAMETERS FOR EACH OF SIX WLND-DIRV-CTLON CATEGORIES
Wind direction category 1
Number of data points 20
2
27
3
42
4
35
5
31
6
12
Dependent variable
- ANTOb
L(1>
" upwind/downwind
TTI (I)
-.05
-.22
.15
.11
-.02
.09
Uref
.06
.36
•23
-.32
-.12
-.17
"road
.20
.40
.24
-.15
.21
.54
AT. , (1)
horiz
.37
.27
.10
.70
.59
.83
0CC (total)
.09
.12
.07
-.23
.17
.60
V0L x SP (total)
-.01
-.00
.00
-.14
.02
.66
V0L (total)
.05
.06
.05
-.21
.16
.70
Dependent variable
" OTlhorU(l> *
- upvlnd/dovnvind
mu(i)
.02
-.28
-.38
-.36 -
.06
-.01
Uref
-.22
.21
-.11
-.32
.19
-.20
"road
.14
.34
.18
-.19
.56
.34
ML .
horiz
.38
.41
.73
.31
.76
.72
0CC (total)
.08
.12
.03
.29
.37
.59
V0L * SP (total)
- .06
.13
-.22
-.11
.05
.64
V0L (total)
.01
.15
-.07
.18
.27
.68
Dependent variable
- upwind/median
-.51
-.22
.21
.25
.24
.78
uref
.67
.57
.57
.58
.42
.56
u
road
.48
.49
.47
.61
.31
.08
6Tt J (1)
hotlz '
-.10
-.02
-.26
.05 -
.08
-.11
tCC(v)
.25
-.05
.02
-.12
.19
-.06
V0L(u) x SP(u)
.26
.14
-.04
.24 -
.17
-.13
"net
-.41
-.20
.17
.14 -
03
.19
V0L(u)
.26
.05
-.02
.03
08
-.14
DRAG(u)
.26
.05
-.02
.03
08
-.14
Vaet
.47
.23
-.09
-.05
16
-.09
Dependent variable
• ATTI
horiz
(I) -
upwind/median
TTIu
-.57
-.34
.05
.48
.55
Uref
.59
.53
.65
.41
.52
"road
.50
.47
.68
.47
.20
-.07
.09
.30
.07
-.04
0CC(u)
-.01
-.03
-.02
-.16
-.04
V0L(u) x SP(u)
-.07
.OS
-.25
.18
-.09
u
net
-.19
-.17
.22
.27
.13
vei(u)
-.04
.01
-.16
-.05
-.10
DRAC(u)
-.04
.01
-.16
-.05
-.10
V
net
.32
.25
-.11
-.20
-.04
Trace Gas Dispersion
The objective of the dispersion analysis has been
to compare diffusion rates between the upwind and down-
wind tracers as a method for examining possible effects
of the traffic flow. Differences between the two re-
flect the influence of the intervening traffic stream
on pollutant dispersion. The general layout of the
near-roadway samplers was shown earlier jin Figure 2.
Using 8 and u , , together with the tracer
road
release rates (Q, g m"l a"*) , the vertical Gaussian
diffusion coefficient a (m) near ground level
z
(i.e., 2 m) was computed from the line source equation,
where
14-2
-------
= ^2/tt Q
U JX
road
(7)
„-3v
and x is the tracer concentration (g m ) .
The use of the Gaussian line source formulation is not
intended to imply that the two-dimensional pollutant
distribution near the roadway is adequately described
by Gaussian concepts. Rather the surface level diffu-
sion coefficient so derived is used as a scaling para-
meter of atmospheric mixing.
For each test,
a values were computed for each
z
of the downwind surface samplers when 6 was not within
20° of the roadway orientation, and the wind speed was
greater than 1 m s ^. Then, the diffusion data for each
tracer were analyzed for their functional dependence on
fetch (x) from the release lane according to the follow-
ing relationship:
(8)
(4) [ct (60-UP)-ct (40-DN)]/a (40-DN)
z z z
(5) a (40-UP)/a (40-DN)
z z
(6) a (60-UP)/a (40-DN)
z z
Eighteen independent parameters were defined, including
normalized volume, normalized occupancy, volume-wind
speed ratio, occupancy-wind speed ratio, turbulence in-
tensity, vertical wind variance, and wind speed. Wind
speed correlated "best" with the six dispersion ratios:
r ranged from -0.23 to -0.35. All other parameters
correlated far more poorly.
Carbon Monoxide Emissions
One additional advantage derived from the use of
gas tracers is that it permits the determination of
vehicle pollutant emission rates when both the pollutant
and trace gas ambient concentrations are measured, in
addition to the trace gas emission rate. When two tra-
cers are used (each emitted on only one side of the road-
way) then vehicle pollutant emissions can be determined
for both traffic streams.
The term a may be thought of as scaling the effect
z-o
of the initial mixing of the tracer gas on the near-
ground concentration at the release point. The coeffi-
cients a and b describe the distance-dependence of
the near-ground diffusion parameter. In determining
a , a and b , a standard nonlinear regression
z-o
technique in SRI's computer library was used. Then
a -values were computed for both tracers for each hour
z
using Eq. (8) at x » 40 m for the downwind tracer,
and x = 40 m and x = 60 m for the upwind tracer.
(Note that the center lane separation of the roadway is
20 m.) The ratio of a (40-UP) to a (40-DN) ranged
z z
from about 0.3 to 3.0 over the 21 hours that satisfied
the meteorological criteria. The two diffusion coeffi-
cients were correlated with each of 10 combinations of:
vehicle volume, vehicle occupancy, wind speed, and the
standard deviation of the vertical wind fluctuations
(er ) . The significant correlations are shown below in
w
Table 5.
Table 5
CORRELATION COEFFICIENTS
Independent Variables
Dependent Variables
a (40-UP) a (40-DN)
z z
road
VOL/u
Uroad
VOL
VOL x CT /u
w road
ct /u
w road
0.77
-0.59
-0.43
0.84
0.84
0.83
-0.57
-0.42
0.90
0.87
Differences in the upwind and downwind roadway
dispersion rates were also examined against meteoro-
logical and traffic variations. Six dependent vari-
ables were defined:
(1) [a (40-UP)-a (40-DN)]/ct (40-UP)
z z z
(2) [a (40-UP)-ct (40-DN)]/a (40-DN)
z z z
(3) [ct (60-UP)-a (40-DN)]/ct (60-UP)
z z z
For an inert vehicle pollutant and inert gas tra-
cers, the following relationships between emissions
-1 -1 -3
(Q, g m s ) and concentrations (\, g m ) hold,
provided the pollutant and tracer are released at the
same location and measured at common points.
V (CO-W) _ X (SF6)
Q (CO-W)
V (CO-E)
Q (CO-E)
Q (sf6)
X (F-13B1)
Q (F-13B1)
and
(9)
(10)
The letters E and W refer to the eastbound and west-
bound traffic lanes, respectively. Background concen-
trations of all gases are assumed to have been sub-
tracted. Of the eight parameters in Eqs. (9) and (10),
only the four terms on the right-hand side are known.
Furthermore, for those sampler locations downwind of the
roadway:
X (CO) = x (CO-E) + X (CO-W)
(U)
(12)
As a model assumption, the relative speed aependence of
emissions in the two traffic streams is represented by
the following:
r -i P
9 - (1(00..) Hg- [f$]
(13)
CO emission rates were computed in the above manner
using the experimental data, with the following excep-
tions:
14-2
-------
• January 17 -- malfunction of the CO analyzer
• Those hours with average wind directions within
20 degrees of the roadway orientation
• Those hours where the average 2- and 3.8-m wind
speed was less than 1 m s"*.
For comparison, directional CO emissions were also com-
puted using the cruise-mode data given for California
autos in EPA report APTD-1497 (1973); these data are
tabulated below:
Speed
CO Emission
(mph)
fgm/veh-mi)
15
69.1
30
29.5
45
24.6
60
25.5
1 _ X1 " X2
Ave. diff. ¦ - E —
N X,
(14)
where X^ is the tracer-derived value and X^ is from
the emissions model. For the eastbound data, the aver-
age difference is 0.35 and the average westbound differ-
ence is 0.36. Standard linear correlations were also
computed for each data set; the correlation coefficient
for the eastbound data was 0.81, and the westbound value
was 0.36
The results given here suggest that the emission
model tested provided a good estimate of actual CO
emissions as determined by the tracer method. Differ-
ences between the two may, in part, be explainable by
a temperature dependence of emissions not considered
here (see EPA report AP-42, Supplement No. 5, 1975).
KEY:
• Eastbound Traffic
A Westbound Traffic
Linear interpolation was used between these values.
Also, the data were updated (from 1971 vehicle mix to a
1974-75 mix) using CALTRANS factors and assuming a five
percent heavy-duty mix; the final factor thus applied
was 0.726.
Figure 9 is a comparison of the CO emissions com-
puted by the two methods. Note that each point on the
figure represents an average of all the downwind sam-
ples obtained for each hour. For the eastbound direct-
ion, the average ratio of all "tracer" CO emission com-
putations to those predicted by the emissions model is
1.00; in the westbound direction the ratio is 1.03.
There is however considerable variance in the individual
comparisons. An average normalized difference between
the two methods was defined by:
' ~
• •
_1_
_L
0 .01 .02 .03
CO EMISSION RATE (g nT's"1)
.04
tracer
FIGURE 9 COMPARISON OF MODEL PREDICTION OF CO EMISSION
RATE WITH COMPUTATION BASED ON TRACER AND
AMBIENT CO MEASUREMENTS
ACKNOWLEDGMENTS
The work reported in this paper was performed as
part of research contract number DOT-FH-11-8125 for
the Federal Highway Administration. Dr. Howard Jongedyk,
Office of Research, is the project manager. The untir-
ing and competent assistance of the following SRI per-
sonnel is gratefully acknowledged: E. Shelar, L. Salas,
A. Smith, C. Flohr, R. Ruff, and H. Shigeishl. The
Meteorology Laboratory of the U.S. Environmental Pro-
tection Agency furnished part of the meteorological
instrumentation used in this study.
9
14-2
-------
A MODEL STUDY OF THE IMPACT OF EMISSION CONTROL STRATEGIES ON LOS ANGELES AIR QUALITY
S. Hameed, S. A. Lebedeff*, R. W. Stewart
Institute for Space Studies
Goddard Space Flight Center, NASA
New York, N. Y. 10025
Summary. A generalized cell model is developed
for the calculation of city-wide averages of photoche-
mical smog components in Los Angeles. This model takes
into account the effects of variations with time, and
within the city, of the source strengths, the wind
field and the mixing depth. The effect of the influx
of background pollution from outside the modeled
volume is also included. Several control strategies
for reducing automobile emissions are then introduced
into the model and their impact on predicted pollutant
Levels particularly those of 03, are Investigated.
I. Introduction
The deteriorating quality of urban air has stimu-
lated widespread attempts to understand the fundamental
physical and chemical processes involved, to describe
the production and dispersion of atmospheric contami-
nants by means of mathematical models, and to devise
control strategies to reduce the quantity of pollutants
in the atmosphere and to ameliorate their effects.
The development of accurate and reliable mathematical
models is central to the problem of improving air
quality. The cost of attempting experimental determi-
nation of the effects of alternate control strategies
would be prohibitive*, thus atmospheric scientists must
have recourse to dispersion models which can predict
the pollutant distributions associated with specified
emissions for a wide range of meteorological conditions.
Models which incorporate widely varying degrees
of detail have been developed for the problem of des-
cribing tho growth *and dispersion of photochemical
smog in Los Angeles. These range from relatively
simple box models in which all input variables are
held constant in space and time^ and background con-
centrations are neglected1 to detailed, three dimen-
sional finite difference models which attempt to fully
account for spatial and temporal resolution of the
emissions source and meteorological variables as well
as the effects of background concentrations.2"" We
have constructed a cell model which is free of some
of the conventional restrictions of cell models. None
of the input variables are assumed to be homogeneously
distributed in the volume of the cell and expressions
are derived for mean pollutant concentrations, averaged
over this volume, which take account of the spatial
and temporal variations in input quantities.6
We have found that neglect of spatial variation
in meteorological variables within the modeling region
is not justified and does not lead to accurate predic-
tion of mean pollutant concentrations. The same is
true of background concentrations outside the modeling
region. On the other hand, we have found that mean
pollutant concentrations are well represented by the
generalized cell model (GCM) if care is taken to in-
clude the necessary resolution in input variables.
Our objectives in this paper are first to briefly
review the results of the validation and sensitivity
studies carried out with the generalized c6ll model
and, second, to study the impacts on Los Angeles air
quality of four control strategies. These strategies
have been previously tested by Martinez et al.7 using
*GTE Information Systems, New York, N. Y. 10025
the General Research Corp. (GRC) model.
II. Model Description
The starting point in the formulation of the model
is the set of N coupled continuity equations
dc
m + V . (u c ) - V . (K • Vc ) = R (1)
-g-— ~ m •• - m m
where c (x, y, z, t) is the concentration of the
constituent, K is the eddy diffusivity tensor, u(x, y,
z, t.) is the wind velocity, and Rm(c^...c^) is the net
rate of production of cm due to chemical- reactions.
Let h(x, y) be the topographic height and H(x, y,
t) the inversion height. The boundary conditions on
(1) are:
At z=h nh • (5 • V)cm - -Qm(x, y, t) (2)
At z=H nH • (V-K • V)cm = nH ¦ Vgffl for (3a)
-nH • (K • V)cm = 0 for n ' V >0. Ob)
is a unit vector normal to the base of the modeled
region, and equation (2) states that the mass flux of
specie m through the lower boundary is given by the
specified function Qm(X. y. t). Hjj is a unit normal
vector directed out of the modele3 region at the
height of the inversion base and V = u i+v} +
(w - ) k is the advective velocity of chemical spe-
cies relative to the moving inversion height. Boundary
conditions (3a) and (3b) state that the rate at which
pollutants are transported into the air-shed from above
is proportional to an assumed concentration, g , above
the inversion height and that the turbulent flux at
this height is negligible.
The horizontal boundary conditions are:
i ¦ (u - K • V) c = I ¦ u g' , I ¦ a < 0, (4a)
- £ • (K • V ) c =0 , £ ¦ u > 0, (4b)
m ~
where £ is a horizontal unit vector directed out the
air-shed and perpendicular to its boundaries. Equations
(4a) and (4b) state that the horizontal flux into the
air-shed is proportional to an assumed pollutant con-
centration, g1 outside the region and that the turbu-
lent flux across the horizontal edges of the air-shed
is negligible. We also require the initial conditions:
c (x, y, z, t ) = c° (x, y, z) (!>)
m o m
o
where the cm(x, y, z) are specified concentrations at
the start of the integration period.
The generalized cell model describes the temporal
variation of mean concentrations averaged over the
volume of the air-shed which we define by
1
14-4
-------
c (t:)
TTt
= r dxdy £ / c
AJ A \ A
dz cm(x, y, z, t) (6)
where = 1
<;>=
+ fm + f"'
top side
top side
c (t) + R (c )
m m n
'Qm (x, y, t) dxdy
v tj ¦ Ho (X/ yt t)
ni
, A 3hA
H at 7
repre-
sents the effect of the changing volume due to change
in inversion height on the mean concentration, and the
terms f and r are defined by:
top
m
side
top
-i. / dxdy n
A J , % !
' A'
H Y
1 k
ds x
B» -
/,
if
Vh
Ul?r»
dxdy 1 n ¦ V
H. ~H ~
* = "i k
side ^ ~
. f ds x 1 /" dz u
Vh
(8a)
(8b)
(8c)
(8d)
f. and fg'de represent the flux of background pollu-
tion from tfteetop and side boundaries respectively,
while rtop and r ., are the removal rates at the top
and side boundaries? A' is that part of the top layer
with incoming flux and A" that with outgoing flux.
For comparison, we also consider a conventional
type of cell model in which the cell is rectangular
with constant height H and length L; a horizontal velo-
city averaged over the cell U is used and inflow of
background pollution is ignored,
such a model is:
The equation for
dcm(t)
dt
Qm't)
H
Uav(t)
°m + Vcn> O)
We will refer to equation (9! as the Standard Cell
Model (SCM).
XXI. Results
A. Model Validation and Sensitivity
To validate the generalized cell model we have used
as data base the valiles of the input parameters used
by the Systems Applications Inc. (SAX) modeli!~5 over a
grid'of 2x2 miles throughout the Los Angeles basin.
This data base constitutes a "model city" and for the
purpose of validating our model the detailed SAI calcu-
lations are defined as the exact solution of the smog
problem for this model city. This procedure insures
that validation of the generalized cell model is riot
influenced by arbitrary choices of parameters, with
the same end in view we have used the same chemical
kinetic scheme and the same boundary conditions as the
SAX model and carried out the validation for the day
for which the most complete data were published, Sept.
29, 1969.5 We then compare the predicted concentra-
tions of the GCM with the averages of the predictions
on the 2x2 mile grid of the SAI model. It is impor-
tant that comparison should be made with these "model
city" averages rather than with the averages over the
Los Angeles observation stations since the latter may
not correctly reflect the average behavior for the
entire city.
The results of the validation study for ozone and
nitrogen dioxide are shown in Figs. 1 and 2.
7 8 9 10 II 12 13 14
TIME
Fig. 1. Comparison of predicted spatially-averaged
concentrations for the Systems Applications Inc. (SAI),
generalized cell iftbdel (GCM), and standard cell model
(SCM).
NSCM
TIME
Fig. 2. Same as in Fig. 1 for NO2.
14-4
-------
The peaks of these species occur as a result of their
production in photochemical processes and their tem-
poral behavior is very sensitive to approximations made
in the model formulation. These are hourly averaged
concentrations, the value for 7 a.m. being the time-
average of those obtained between 7 and 8 a.m. The
SAI and GCM curves are compared in these figures with
the result of the standard cell model or SCM, Eq. 9.
The GCM gives a good representation of the averages of
the SAI model, but the SCM predictions are relatively
poor especially for ozone where no peak Is obtained
and the concentration becomes much too large.
Fig. 3 shows the sensitivity of the predicted 03
concentration to assumed levels of background pollu-
tion. Curve A is the standard result (also shown as
the GCM curve in Fig. 1). The concentration changes
to curve B if the assumed background concentration is
halved and to curve C if it is taken to be zero. Curve
D results from including the flow of background
pollution across lateral boundaries while taking
pollutant concentrations to be zero above the inversion
layer and curve E included the pollution flow from
aloft and excludes it across lateral boundaries.
C
o
o
£
8
c
o
Q
•\P
14
13
12
7
8
10
II
9
TIME
Fig. 3. Sensitivity of predicted 03 concentrations
to assumed background pollutant concentrations. A -
GCM baseline case from Fig. 1. B - background con-
centration reduced by h. C - no background concentra-
tion. D - no background above inversion layer. E -
no background outside lateral boundaries.
Despite the non-linearity of the photochemistry and
the relatively large size of the modeled region the
mean concentrations are predicted almost as well by
the GCM as by the more comprehensive SAI model. As
shown in Fig. 3 it is essential to legislate values
for the background pollutant concentrations to obtain
accurate values fori mean concentrations and we may
conclude that the latter can be accurately predicted
because the non-linearities in photochemistry have a
small impact within the modeled region relative to the
effect of exchange of pollutants with the background.
B. Impact of Emission Control Strategies
An interesting application of pollution dispersion
models is in devising control schemes to improve air
quality. We can not expect accurate assessments of
the impact of a given emission control strategy from
the present generation of urban dispersion models.
The exploration of the relative effects of alternate
control strategies with such models may, however, pro-
vide useful qualitative insights and is certainly
warranted by the relevance of the problem. Such calcu-
lations also provide insight into the modeling process
itself and should aid in the development of future
models.
We have applied our model to the four control
strategies studied by Martinez et al.7 using the
General Research Corp. (GRC) model. This model is
quite different from our generalized cell model since
it utilizes a Lagrangian formulation in which pollu-
tant concentrations are computed in selected air par-
cels as they are transported by the wind along spe-
cified trajectories. Since our model computes spatially
averaged concentrations for the entire city we cannot
expect completely analogous results though we would
hope to find some similarity in the relative impacts.
For each strategy considered we calculate the concen-
trations of smog components from 7 A.M. to 3 P.M.
Starting and ending times for the two Sept. 29 trajec-
tories of Martinez et al. are 0 530 hrs. to 1230 hrs.
and 0230 hrs. to 1230 hrs.
The four strategies chosen for study are:
1) 30% reduction in carbon monoxide (CO), reactive
hydrocarbon (HC), and nitric oxide (NO) emission
from mobile sources.
2) A central park strategy in which vehicle emissions
are reduced by 90% over a 64 sq. mile area nearly
centered in downtown Los Angeles. This approximates
the second strategy considered by the GRC model.
3) The EQL strategy which consists of: a) 62% reduction
in CO emissions together with a 50% cut in initial
CO load, b) 79% reduction in HC emissions from mo-
bile sources together with a 63% cut in initial
load, and c) 73% reduction in nitric oxide emissions
from mobile sources with a 58% cut in initial N0X
load.
4) Zero emissions strategy - the total elimination of
all emissions without changing initial conditions.
In strategies 1 and 2 the initial conditions are
assumed to be the same as in the baseline case. The
results of the GRC study indicate, however, that the
predicted impact on ozone levels is very sensitive to
the initial conditions. We therefore also consider
strategies 1 and 2 in combination with reductions in
initial concentrations of pollutants by the same frac-
tions as the proposed reductions in their emissions.
Thus we have the additional cases:
1A) Strategy 1 with modified initial conditions. Since
automobile emissions account for 99% of CO emis-
sions, 71% of reactive hydrocarbon emissions and
58% of NO emissions the initial concentrations of
these pollutants are 0.70, 0.79 and 0.83 of base-
line values respectively.
2A) Strategy 2 with modified initial conditions. Total
HC and NO emissions are reduced to 0.92 of their
baseline Values in this case and the initial con-
ditions are assumed to be lowered by the same fac-
tor.
These should give a somewhat more realistic assess-
ment of the impact of the 30% and central park strate-
gies since the implementation of emission controls on
a long term basis is likely to alter morning concentra-
tion levels, though not necessarily in the linear
manner assumed in 1A and 2A.
3
14-4
-------
Table Comparison of effects of various emission strategies in the GCM and GRC models.
Numbers shown are the ratios of the guantity after applying controls to the
quantity before controls are applied.
Strategy
(HC/NOx)
t=0
(HC+NO)
x t=0
Oj peak
NO2 peak
GRC Model O, peak
average of trajectories
1 and 2
1.
1A.
2.
2A.
3.
4.
30% Strategy
30% Strategy
with modified
initial con-
ditions
Central Park
Central Park
with modified
initial con-
ditions
EQL Strategy
Zero Emission
.95
.88
.80
. 92
.39
.96
.65
.99
.86
.30
.78
.95
.71
.99
. 88
.29
.76
.96
1.01
.84
.78
The results of these calculations are shown in
Table 1 and in Figs. 4-8. The first two columns of
this table give the HC/NO ratio and the initial load
level (HC+NO ) relative to the baseline case for the
GCM. Ozone fnd nitrogen dioxide peaks relative to
the baseline are shown in columns 3 and 4. The fifth
column gives the ratio of the O.o peak, averaged over
trajectories 1 and 2, relative to the baseline case
for the model of Martinez et al.
Baseline and
Strategy 2
ll 12
TIME
Fig. 4. sensitivity of 30% and central park strategies
for Oa to initial conditions. 1 - 30% strategy with
baseline initial conditions. 1A - 30% strategy with
proportionate reduction in initial concentrations. 2 -
central park strategy with baseline initial conditions.
2A - central park strategy with proportionate reduction
in initial concentrations.
In Figure 4 the ozone predictions of strategies
1, 1a, 2 and 2A are compared. It can be seen that
the impacts of the strategies are much more significant
when the initial conditions are modified. It should be
pointed out that in our model (as in the SAI model) the
specification of initial conditions prescribes the
levels of background pollution outside the modeling
region. Thus, for background concentration gffl above
the inversion layer it is assumed that
g (t)
m
- q°
in 'm
"2
(10)
750
; < 750 feet
750 feet.
where c
is the initial concentration, is the average
mixing 'depth and Hj? its initial value. g° is the
assumed 'clean air' value of background concentration
above a height of 750 feet. Similar formulas are adopt-
ed for prescribing background levels across the lateral
boundaries. Since ozone levels in our model are very
sensitive to assumed background concentrations (Fig. 3)
it is not surprising that they are also sensitive to
the initial values, as seen in Figure 4.
In Figures 5-8 strategies 1A, 2A, 3 and 4 are com-
pared with the baseline results. We find that there is
a significant reduction in ozone peak in each case.
Also, the shape of the ozone curve is nearly preserved
in all cases; this means that ozone 'dosage', or the
total amount of ozone produced with each strategy, is
nearly given by the ratio of the peak to the baseline
peak.
From Table 1 we see that the reductions in ozone
peaks for strategies 1, 2 and 4 obtained by our model
are close to the average of the reductions found for
trajectories 1 and 2 (both for Sept. 29, 1969) in the
GRC model. The predicted changes in ozone levels are
however quite different in the two models for the EQL
strategy.
14-4
-------
Baseline
2A
g
c
8
§
7
8
9
10
II
12
13
14
15
TIME
Fig. 5. Impact of control strategies on predicted O3
concentrations. 1A - 30% strategy with modified ini-
tial conditions. 2A - central park strategy with mo-
dified initial conditions. 3 - EQL strategy. 4 - Zero
emissions strategy.
NO.
Baseline
2A
S
e
c
7
12
13
8
10
9
14
15
TIME
Fig. 6. Same as Fig. 5 for NO2.
We find a 70 percent reduction in peak mean ozone con-
centration as opposed to a reduction of 16 percent
averaged over trajectories 1 and 2 of the GRC model.
This difference may result in part from the difference
in the nature of the dependent variables of the two
models. We cannot expect that the mean city-wide 03
NO
Baseline
\-2A
TIME
Fig. 7. Same a? Fig. 5 for NO.
60
Reactive HC
Baseline
52
2 A
48
44
5
I
o.
0.
x (A
40
§
32
28
24
20
TIME
Fig. 8. Same as Fig. 5 for reactive HC.
concentration will necessarily behave in the same way
as O3 concentrations along chosen trajectories. The
difference may also result in part from the fact that
integration along the GRC model trajectories ceased be-
fore an O3 peak was attained. We note that along
14-4
-------
trajectory 1 the EgL strategy loads to increased 03
levels throughout the morning hours and only begins
to lead to reduced 03 at the time integration is ter-
minated. Along trajectory 2 the EQL strategy leads
to reduced o, levels and the difference is still in-
creasing when integration is terminated. The effective-
ness of the EQL strategy is thus increasing along the
GRC model trajectories when integration is terminated
and smaller peak 03 levels would presumably be found
at later times.
IV. Conclusions
Using a generalized single cell model we have con-
ducted a study of the impact of selected emission con-
trol strategies on Los Angeles air quality and com-
pared the results with a study of the same strategies
conducted by Martinez et al. with the substantially
different GRC model. We find that all the strategies
tested result in reductions in mean city-wide ozone
concentrations throughout the day while the GRC model
finds increases in ozone concentration along selected
trajectories and at certain times for some of these
strategies. This probably reflects a difference in
the nature of the concentration variables being
considered rather than a conflicting result from the
models.
In terms of reducing peak 03 concentration the
most effective strategy in the GRC model is the zero
emissions while we find the EQL strategy to be far
more effective.
The greatest difficulty in modeling the impacts
of emission control strategies is the sensitivity of
the calculated results to boundary and initial con-
centration values of the model. The effect of the
former is clearly illustrated in Fig. 3 while the
latter is shown by the difference in our results for
the central park and 30 percent strategies with and
without a corresponding reduction in initial load (Fig.
4). As a consequence of this sensitivity to pre-
scribed inputs, present models can provide only
qualitative guidelines to the relative effectiveness
of alternate control strategies • A truly quanti-
tative analysis of a given strategy will require more
comprehensive calculations capable of modeling the
many interactions and feedbacks which exist between
concentration values within and outside the urban
air-shed.
REFERENCES
'Hanna, S. R. (1973) A simple dispersion model for the
analysis of chemically reactive pollutants. Atmos-
pheric Environment 7, 803-817.
2Reynolds, S. D., Roth, P. M. and Seinfeld, J. H.
(1973a) Mathematical modeling of photochemical air
pollution - I. Formulation of the model. Atmospheric
Environment, 7, 1033-1061.
3Roth, P. M., Roberts, P. J. W., Liu, M. K., Reynolds,
S. D. and Seinfeld, J. H. (1974) Mathematical modeling
of photochemical air pollution - II. A model and in-
ventory of pollutant emissions. Atmospheric Environ-
ment, 8_, 97-130.
Reynolds, S. D., Liu, M. K., Hecht, T. A., Roth, P. M.
and Seinfeld, J. H. (1974) Mathematical modeling of
photochemical air pollution - III. Evaluation of the
model. Atmospheric Environment, 8, 563-596.
5Reynolds, S. D., Liu, M. K., Hecht, T. A., Roth, P. M.
and Seinfeld, J. H. (1973b). Further development of
a simulation model for estimating ground level con-
centrations of photochemical pollutants - Final report.
System Applications, Inc., San Rafael, CA. 91109.
6Lebedeff, S. A., Hameed, S., and Stewart, R. W. (1975)
Ors the validity of a single cell model for photochemical
smog in Los Angeles (to be published).
'Martinet, J. R. , NordsiecX, t\. A., and Eschenroeder,
A. Q. (1973) Impacts of transportation control strate-
gies on Los Angeles air quality. General Research
Corporation Rept. CR-4-273, Santa Barbara, CA. 93105.
14-4
-------
POLLUTIONAL EFFECTS OF SUSPENDED, SEDIMENTED
AND ERODED PARTICULATE MATERIAL IN THE
AQUEOUS ENVIRONMENT
by Hermann H. Hahn and Rudolf Klute
Institut fur Siedlungswasserwirtschaft
University of Karlsruhe
75 Karlsruhe W.-Germany
Abstract
Sediments arid corresponding loads of suspended
material in an aquatic system reflect hydrochemical
and ecological conditions and their change over a
long time period. There is evidence that monitoring
of the dissolved phase alone does not allow a definite
description of any aqueous environment. - There is a
direct pollutional effect from suspended or sedimen-
ted fine particles as demonstrated, for' instance by
changes in biological activity in the river Neckar or
by increased difficulties for various water users. In
addition there are indirect pollutional effects
through various reactions of dissolved polluting sub-
stances with the solid phase as exemplified by the
fate of heavy metals. - The most important processes
responsible for transport and distribution of particu-
late material are aggregation, sedimentation, erosion
as well as aggregate destruction. The extent arid rela-
tive effect of these processes depend upon hydrodyna-
mic factois on one hand and physicochemical factors
on the other. Dissolved substances through their asso-
ciation with the solid phase, mainly due to adsorptive
mechanism are affected by similar processes as the
enrichment of heavy metals in the river Neckar shows.-
There are first attempts to describe pollutional ef-
fects of particulate material and associated dissolved
substances. A simple mass-balance as well as a more
complicated dynamic mociel for the distribution of
suspended solids and heavy metals, as applied to the
river Neckar, show the usefulness and range of applica-
bility of such modelling. From these considerations it
is concluded that present monitoring of the suspended
solid phase alone is insufficient. If data and knowl-
edge existing on the suspended material are to be
used in a meaningful way additional information has
to be gathered.
Similarly diverse as the description of
the complex suspended particulate matter is
its quantitative determination and the stan-
dards based on it. The following ^xgerpts
should serve as an illustration: '
SOURCE
WHO drinking wa-
ter standards
US drinking wa-
ter standards
German drinking
water standard
(DIN 2ooo)
Modern monitoring
equipment
CRITERIA STANDARD
light path frcm 5
inside bottcm
of a glass tube
(Jackson turbi-
dity units)
SiOi scale (4o cm 5
high column of vary-
ing concentration
of Si02 in mg/1)
B-SEDIMETER
filtration and
sedimentation as
mg/1 undissolved
solids
METRAWAIT
turbidity deter-
mined by light
scattering
SIGRIST
turbidity deter-
mined by comparison
with optically
stable standard
25 units
5 units
1o units
I. Nuisances, disturbances and
pollutional effects of parti-
culate matter.
Particulate matter is noticed, deter-
mined and monitored in a number of different,
even contradictory ways, depending upon the
point of view one speaks of turbidity, of
filter residues, of non-dissolved substances
or of colloidal, suspended, eroded etc, ma-
terial. In this discussion particulate matter
shall be understood as non-dissolved inorga-
nic substances, such as metal oxides, clays,
intermediaries in precipitation processes and
non-dissolved organic substances, such as
autotrophic and heterotrophic microorganisms
as well as organic detritus in comparable size
ranges. The natural aqueous environment is
characterized by concentrations of inorganic
particulate matter ranging anywhere from 1oo
to more than 1ooo mg/1 (or about 1o8 to 1o9
particles per milliliter) or by amounts of
organic particulate matter in the range of 1o5
It is apparent that the information derived
from such measurements is only useful for spe-
cific purposes and fails to explain a number
of significant characteristics of the parti-
culate phase. This will be shown through a
number of qualitative and quantitative examp-
les .
1.1. Nuisance created by particulate matter
From the viewpoint of many users, particu-
late matter, specifically if it is mainly in-
organic, creates nuisances and leads to dis-
turbances, for instance in industrial proces-
ses. The following list of quality tolerances
for industrial process waters underlines this
statement: 5
SSfi *** Sf ^ather Paper t£" tile
1o JTU 1o JTU 1o JTU 2o JTU 5 JTU
15-50 2 JTU 5 JTU
JTU
to 1o microorganisms per milliliter.
11
14-5
-------
Even a very brief inspection of this list
will show that not too much insight is gained
by determining the global parameter turbidity
in the way described above.
1.2. Disturbance and pollution caused by
particulate matter.
Disturbances and pollution must not only be
seen from the point of view of direct water
use in household and industry, but also in
comparison with the state of undisturbed natu-
ral waters. Again an example, this time taken
from the natural aqueous environment, should
serve for illustration's sake. Penetration of
light ia often limited by suspended materials,
restricting the photosynthetic zone wherever
habitats have appreciable depths. Turbidity,
therefore, is often important as a limiting
factor. 2* Examples have been reported of 61 %
reduction in primary productivity due to heavy
sediment transport (Cairns). 19 Figure (1a)
schematically depicts this fact.
In addition to such direct effects of sus-
pended particulate matter a number of indirect
effects of suspended solids upon aquatic orga-
nisms has been discussed, such as mechanical
or abrasive action, blanketing action or sedi-
mentation and also toxicity effects of heavy
metals sorbed onto particulate matter on the
biocoenosis. 17' 19 The latter effect will be
discussed in more detail in the following para-
graph.
X.3. Toxicity associated with and indirectly
caused, by particulate matter.
While there is some evidence that particu-
late matter in itself can lead to toxic reac-
tions, the majority of findings relating toxic
effects concentrate on indirect relation-
ships. The association of heavy metals with
particulate matter in its different forms exeirr-
plifies this fact. Heavy metals (as well as
other similarly sorbing substances) can be
transported and distributed in the aqueous en-
vironment by various mechanisms. Gibbs lists
for heavy metals the transport in solution,
through ion exchange, with organic material,as
metallic coating and as crystalline solids. 1°
In this present discussion these transport
mechanisms are summarized and operationally
defined as transport in dissolved form and
transport by "adsorption" on inorganic and or-
ganic particulate matter. Figure (1b) indicates
the rate and the extent of the sorption and
desorption of cadmium from inorganic and or-
ganic particulate matter. 9 Similar findings
on sorption, on desorption and on storage in
sedlmented particulate matter have been repor-
ted by other researchers. 3'6'7' 2 5
exact mechanisms of sorption as well as the
form in which these toxic substances are found
on the solid surface are still discussed. 14 _
However, it can be stated that all particulate
matter affects through its reactions with the
dissolved phase the distribution of such dis-
solved water constituents to a significant de-
gree .
1.4. Pollutional effects of sedlmented particu-
late matter , ~ ~
Depending upon hydrodynamic as well as physi-
cochemical conditions, suspended particulate
matter may sediment. In this way all direct and
indirect pollutional and toxic effects origin-
ally caused by particulate matter are now asso-
ciated with sediments. It is in particular
these sediments that reflect hydrogeochemical
and ecological conditions and their change over
a longer time period. Basically, sediments
can act in two different ways; for one they
may serve as a sink for various dissolved sub-
stances. The much discussed oxygen demand of
organic sediments illustrates this fact. On
the other hand, sediments can serve as a source
of substances that are unwanted and the appear-
ance of which leads to a distubance of the
equilibrium of the aqueous system. Whether one
cites release of sedlmented phosphorous or the
reactivation of heavy metals under changed so-
lution conditions, the role of sediments as a
source of polluting substances is quite clear.
It is in particular for this reason that a bet-
ter understanding of the most important proces-
ses responsible for the distribution of parti-
culate matter is needed. The following para-
graphs contain first proposals on more adequate
methods of characterizing particulate matter,
on the possible reactions that affect their
fate significantly and on simple models that
allow quantitative description of such reactions.
FIGURE 1 : Pollutional Effects of Particulate Matter
a) Reduction in light intensity in nat. waters
b) lransport of toxic materials with parti-
culate matter (adsorption and desorption)
according to Gardiner
a)
photwfrthitic f
5W HO
tWOmv tUlQtty
hectwM Mi tiutac* of ratficats
Itonprw to UH)
TOO (ligMintuHitji)
fiK tony e< tNttwt ttartvbyU
job naatfotbtm
itttm tetto wrfow
ttQcflt Mew tvrta*
b)
adsorption
hgm
in rww&liy
M-J nfl
wbtl
NO (toot mound
(.JO
toarptnrn
IflffV.
iov.
dtfttptofl 4< c
Uvm twmtt wltrial
(originally add«d 23pg Cd/t)
14-5
-------
1.5. Proposal for a more exact description of
particulate matter corresponding to its pollu-
tional effects.
The following simplified first proposals for
a more accurate characterization of specific
pollutional aspects of particulate matter should
serve primarily as a basis for discussion. Further
more, it is understood that it will be difficult
to realize all of this in existing measurement
programs. 12 On the other hand, it is intended
to raise doubts about the usefulness of
some of the parameters measured today in order
to describe particulate matter.
eral form of the equation describing the dis-
tribution of the particulate matter as 3. func-
tion of time and space for given hydrodynamic
conditions is then as follows:
|f- W/? vrJ - (v-c) . (t)
where ^
Qx
0'v J
0
Assessment of
characteristic
aspects
1) Exact amount
of particulate
matter in water
2) Specific pro-
cesses for remo-
val of particu-
late matter
3) Specific ef-
fects in the
aquatic system
4) Transport,
storage and re-
lease capacity
for polluting
material
Desirable
measurements
Number concen-
tration
(numbers/ml)
Size distribu-
tion
(da
etc.
Particle size
distribution
N = N(d)
Specific sur-
face concen-
tration
(m7ml) 8'
Possible method
of determination
Light scattering
measurements (calt-
bration)
Microscopic-scan-
ning methods
Electric resis-
tance measurements
Filtration-refil-
tration methods
22
Sorption arid
desorption
characteristics
Inorganic and
organic fraction
of particulate
phase
Rate and amount of
adsorption of cer-
tain test substan-
ces
Volatilization of
suspended particu-
late matter
IX.
Reactions of particulate
matter determining its
distribution in the aqueous
environment.
D =
concentration of particulate matter
as mg/1 or particles/ml
diffusion respectively turbulent
dispersion coefficients
flow velocities characterizing
convective transport
This equation can be simplified for the use in
a primarily one-dimensional river with segment
wise homogeneity and reads then as follows:
It - D ~ §£ • • ¦ • (2)
This well-known form of the transport equation
appears from an operational point of view sa-
tisfactory for aqueous systems like rivers and
estuarine zones. The solution of the differen-
tial equation, in particular for a realistic
system with a large number of serially connec-
ted quasi-homogeneous river segments (where
one deals with a system of differential equa-
tions) presents no problem and can be found by
such numerical procedures as the forward dif-
ferences methods.
Co.,
s-i , ¦ L -> 2 / , /-
(¦n-
V,-
2h
where:
¦ G)
It has been shown in the preceeding para-
graphs that particulate matter in its suspen-
ded and sedimented form may cause directly and
in particular indirectly grave pollution. For
this reason it is necessary to understand the
factors that govern the motion of particulate
matter. The most important processes which are
briefly to be discussed in the following para-
graphs are horizontal transport through turbu-
lent dispersion and convection and vertical
movements, mainly sedimentation and erosion,
coupled with aggregate formation and aggregate
break-up.
II.1. Horizontal transport of particulate
matter in suspended form.
Brownian motion, turbulent dispersion and
convection may all be invoked for the horizon-
tal and to some degree vertical distribution
of particulate matter. Considering the size
range of the majority of suspended particulate
matter, one can relatively firmly conclude
that for natural aqueous systems hydrodynamic
conditions are such that Brownian motion as a
transport mechanism may be neglected. The gen-
At = time increment
h = interval
superscripts 1 and 2 representing the
conditions at the beginning, respective-
ly end of the time period At
Solving this system of equations for the Neckar
river at specific and characteristic cross-
sections for estimated hydrodynamic parameters
in the range of v = o.1-1.o m/sec and D = 1-2
m2/sec leads to the conclusion that convection
accounts for the most significant amounts of
particulate matter transported. Only close to
concentration discontinuities (sources, inlets,
sinks, etc.) turbulent dispersion may affect
the distribution of particulate matter to any
extent. For the here considered hydrodynamic
conditions this is only the case within stret-
ches not larger than 5o to 15o m downstreams
from the considered point of discontinuity.
Thus, monitoring the movement of particulate
matter is only meaningful if flow velocities
(and turburlence characteristics where neces-
sary) are recorded.
II.2. Aggregation of particulate matter
The preceeding section showed the effects
of hydrodynamic parameters upon the distribu-
tion of particulate matter. In this paragraph,
the significant influence of physicochemical
parameters on the distribution of particulate
matter mainly through aggregation and peptiza-
tion, affecting sedimentation and erosion is
to be discussed. Aggregation has long since
been identified as the result of two indepen-
dent reaction steps: particulate matter, i.e.
suspended particles must be transpor-
t e d to each other and these colliding par-
ticles must be destabilized such
3 14-5
-------
that the collisions will lead to permanent
agglomerations.
Of the various mechanisms invoked for par-
ticle transport, Brownian motion, linear velo-
city gradients in laminar shear flow and eddy-
structures in turbulent flow, Brownian motion
will be only of importance for very small sus-
pended particles. (The ratio of collision fre-
quencies due to laminar shear flow and due to
Brownian motion C3 V 2«7"
where R^j is the collison radius, du/dz the
linear velocity gradient and k the Boltzmann
constant, shows that already for particles lar-
ger than 1 m at du/dz = 1 linear velocity gra-
dients contribute more significantly to the trans-
port) . Laminar flow conditions with linear
velocity gradients on the other hand will rare-
ly be encountered in natural systems, such as
rivers and estuaries. The main transport mech-
nanism for suspended particulate matter in na-
tural waters will therefore be systems of tur-
bulent eddies. It has been found in this au-
thors' laboratory that von Smoluchowski's ori-
ginal formulation and its further modification
by Camp, using the energy input into the aggre-
gating system as a measure of the transport
efficiency does not yield satisfactory results.
Reaction rates, determined from observed chan-
ges in particle size distribution of aggrega-
ting suspensions with comparable energy input
differed as much as 15o %. (At 34 rpm and a
G-value according to Camp the turbine stirrer
produced a reaction rate of 12* 1o-1* sec.-1
while the anchor stirrer yielded at identical
conditions a rate constant of 4,8«1o sec." .
At higher rotational speeds, 84,5 rpm, due to
changes in the flow structure, rate constants
are changed drastically, leading to .
= 4,2 • 1 o-11 sec."1 and k ,= 9,7-lo"* lec?Sl
again for identical conditions). It has been
concluded that the size of the eddies relative
to the suspended particles must be considered
in addition to the overall energy input.
Delichatsios et al. conclude also on the basis
of theoretical and experimental investigations
that there is a reduction in collision frequen-
cy when the ratio of particle size to the Kol-
mogorov micro-scale increases. 4 These deliber-
ations lead to the simplified operational mo-
del for the aggregation process as given in
equation (4).
Particle destabilization has been shown in
the laboratory to come about either through
double-layer compaction, through modification
of surface charge or potential and through
multiple adsorption of bridge-and network for-
ming long chain polymers on more than one par-
ticle. For natural systems one can invoke the
following destabilization mechanisms as the
authors' laboratory has determined:
Inorganic material through double-layer
compaction (aggregation through NaCl addi-
tion 2o - Figure Pa)) and through surface
charge and potential changes (aggregation
through CaCl2 and AICI3 addition).
Organic material through double-layer com-
paction (aggregation through NaCl addition
- Figure (2a) - personal communication from
B. Eppler - these authors' laboratory) and
through surface charge and potential chan-
ges.
FIGURE 2 : Aggregation of Particulate Matter - Collision efficiency
values for a)pure electrolytes and b) Rhine water
samples (according to Neis and Eppler)
0)
—, 1 ..
i
—t
0 On
"B 0.5
-3 -1 -1 0
toSlMot/OKoCI
°Huart2 •MiimnMncuiite "Kaolinite
•MontmoriUonitt a Mile pH=65
'Bacillus NaCl pIM •Ctaetli ItaCI pH=7
-1 0
logIMol/l]CaCI;
0 Quartz •AMmumoxide " Kaolinite
•Montmorillonlte o lllite pH=
'Bacillus Cad: pH=l>
b)
; Q07
! 0.06
005
out
003
ao2
@ Bacillus cer.
-I «=AZ.
- EPM
Kaolinite
100 300 400 500 600
i
100
Chtorello
BBrrvis
• fPM
—i 1 0
700
Rhine km
In the view of generally accepted colloid de-
stabilization mechanisms it is interesting to
note that:
Naturally occurring organic material that
might be expected to lead to bridge forma-
tion, appears stabilizing in particular for
inorganic matter (reduced collision effi-
ciency factors - Figure (2b)).
Organic particulate matter aggregates through
the addition of double-layer compacting sub-
stances as well as through bridge forming
material. It must be noted, however, that
the concentration of the destabilizing re-
agent plays a most significant role: at
higher salt concentrations for instance, re-
duced collision efficiency is observed
(possibly due to the exudance of stabili-
zing cell material).
The effects of different suspensions upon
each other, possibly in the form of of mu-
tual coagulation are not yet investigated
for such systems. There are indications that
enhancement of aggregation takes places when
inorganic and organic particulate matter
come into contact. 16
In summary then aggregation of both, inorganic
and organic material takes place in natural
systems at rates that are lower than expected
from laboratory experiments. Knowing actual
> 14-5
-------
collision efficiency factors for aquatic systems
as determined for instance by Neis 2o one can
formulate a simplified kinetic model for the
aggregation (particle number reduction) of par-
ticulate matter:
9i
C
where: k = %¦¦$>¦ c*-/3£
= particle volume fraction
cx = in situ observed collision efficien-
cy factor
/3E = effective energy input leading to
collisions
Combining this physicochemical model for par-
ticle aggregation with the transport equation
(2) one obtains:
Ot
D
a*2
- m.
3! Wflc^cm/wc
Oaptli on
B jaalk»w91 VttoUtra/m
J-HI IK-1
* hp
II.3. Aggregation and sedimentation of parti-
ulate matter.
After investigating the actual role of ag-
gregation of particulate matter in natural
systems it is now possible to discuss sedimen-
tation of suspended particles and its change
through ongoing aggregation. Two different
cases of sedimentation must be distinguished:
discrete settling of individual particles or
aggregates and hindered settling where partic-
les are spaced so closely that the liquid dis-
placed by moving particles is confined and
exerts additional friction on the sedimenting
particle. According to Fair et al. hindered
settling becomes significant under the follow-
ing conditions: vi»w.w/vjw/. *= Cn)*-
(n = porosity). 5 Thus, only at porosity values
lower than 99,75 % (or volume concentrations
larger than o,25 % or 2,5«1o-3 - compare floe
volume ratios of aggregating suspensions as in-
vestigated in the order of size of o,75»1o-'*)
corresponding to about 66oo mg/1 suspended
solids, hindered settling must be considered.-
Discrete settling can be described by consider-
ing gravity effects and fluid drag, the latter
one under due consideration of laminar or tur-
bulent flow conditions.
If one considers one particle or aggregate
that moves downward a certain distance under
the influence of gravity in a given time one
finds that the number concentration of parti-
culate matter in a given volume changes in the
following way:
c
h
where: v
sed
fs)
the respective sedimentation rate
under laminar or turbulent flow
conditions
h = depth of the system under consi-
deration
Taking into account that particulate matter
aggregates and increases its overall size
(according to a "liquid drop model" the fol-
lowing relationship holds d„-x /d„-(/\/r,/
as well as its sedimentation rate, one finds
the following rate law for the reduction of
particulate matter (number concentration) in
turbulent media. The Reynolds number for the
Neckar at station Obrigheim (km 87) with
v = o,1 m/sec. R = 318 cm is Re = 3-1o5 i.e.
turbulent flow:
-fr~ - 6m- cs/st. (7)
where:
sed
ps,p = specific gravities of sedimen-
ting particles respectively
suspending medium
do,Co = size and number concentration
of initially present "primary"
particles
Combining equation (7) with the transport
model equation (2) one obtains for the over-
all change in the number concentration of par-
ticulate matter due to dispersion, convection,
aggregation and sedimentation:
W ~ D "Is** ^ ¦§* ~ C /e (6)
With respect to practical use and numerical
solution of this equation the statements made
in the discussion of equations (2) and (5)
apply. In order to describe or predict chan-
ges in the particulate matter concentrations
the following parameters have to be known or
determined:
Physical parameters of suspending medium:
absolute viscosity of suspending medium
flow rate (or detention time)
energy dissipation and its efficacy (eddy structure)
specific gravity of suspensa
5
14-5
-------
Chemical parameters of suspending medium:
inert counter ions
charge and potential changing ions
bridge or coat forming substances
Parameters characterizing the suspensa:
particle size
particle number concentration
floe volume ratio
The intensity of the influence of these va-
rious parameters upon the rate of change in
the content of particulate matter is different,
These authors' laboratory made estimates on
the relative effect of different parameters
through laboratory studies (see Figure (4)).
FIGURE h: Aggregation and Sedimentation of Kaolinite
200mg/1 In 10'2M/I CaCI2
at pH=6 energy input 25 sec-1
Curve I : 5min aggregation 5mm sedimentation]
Curve II :5min aggregation 30min sedimentation I dist. HjO
Curve III :30min aggregation 30min sedimentation J
Neckar |
Rhine t 5min aggregation 30min sedimentation
Alb J
Temperature
Absolute visco-
Turbidity reduc-
tion
2o
1o°C
15°C
2o°C
25°C
1/31
1,14
1,o1
o,89
29
34
38
42
2,63
2,58
2,6o
2,67
45
100
lurbidity t HUJ
¦t«d>ur*d tr
-I
twwin
In a second phase of the investigation, aggre-
gation and sedimentation in actual river waters
under controlled hydrodynamic conditions was
studied. Again, it becomes apparent that par-
ticle number reduction (due to aggregation and
/or sedimentation) in natural water systems is
less efficient than in comparable ionic media
of only anionic constituents. It is particular-
ly noteworthy that two different investigations,
those summarized in Figures (2) and (4), lead to
results that compare very well as the follow-
ing sample calculation shows:
Kaolinite stability value in lo-z M/l CaCl2 medium:
a = o,38
Kaolinite stability value in Rhine water sample with
comparable electrolyte conditions and at D0C=6 mg/1:
a r o,o35
Change in particle number reduction due to change in
collision efficiency factor C/C* <*-xp [-%$/<£(cx*-
Oi)i] i C/C*~0-76 (for 1- SCO sec />E-
2s^rc -0.
These conditions are the same as reported in Figure
(4) where depending upon the specific temperature
similar changes in particle number reduction due to
change from pure electrolyte to Rhinewater (DOC =
7,8 mg/1) can be seen:
Temperature 5°C
N / N* 0,84
lo°C
o,8o
15°C
0,75
2o C
0,74
35 PCitaiin
In a first phase of experimentation aggrega-
tion and sedimentation of kaolinite in a CaCl2
medium of known concentration and pH value was
studied. Curves I through III of this figure
show that aggregation alone as well as sedi-
mentation alone are not as effective as the
combination of aggregation and sedimentation.
There is a clearly discernible temperature
dependence with an increase in particulate
matter removal at higher temperatures. Compar-
ing the change in the absolute viscosity of
the medium (see above) at increasing tempera-
ture with the change in turbidity reduction
one concludes that most of the changes in
turbidity reduction can be explained in terms
of viscosity changes alone. This means that
temperature changes mainly affect the sedi-
mentation step.
This confirming experimental evidence on the
role of aggregation and sedimentation in na-
tural waters suggests that the change in par-
ticle number reduction due to changing solu-
tion conditions must be ascribed primarily to
the aggregation step since all results repor-
ted in Figure (2) are based exclusively on
aggregation studies.
XI.4. Erosion of particulate matter.
Records on suspended sediment loads of rivers
at varying discharge show that not only reduc-
tion in particulate matter takes place, but
also an increase can be noticed depending upon
the discharge in the river. Data from the river
Neckar at the station Poppenweiler (Figure (5))
or reports in the literature ' suggest that
the amount of particulate matter in suspension
is not only determined by aggregation and sedi-
mentation alone, but influenced significantly
by erosion and resuspension. This means that
generally suspended sediment loads increase
with increasing river discharge. For reasons
discussed in the previous paragraph such appa-
rent correlations must not be observed in all
instances; examples of an apparent relation-
ship between suspended sediment load and water
temperature have also been reported.11
6
14-5
-------
2 Aw
FIGURE 5: Sedimentation and Erosion in the leckor in Conjunction with River Discharge
(insert: calculation ol flow-velocity (ram river discharge I
cn/lKubU tnatlw and
liscMrgt al
tmatn!
amufi (low wtouij tx
tfiiriiorgn
lliWl
0 it il H I I ITS I» ej»<"{
The problem of describing erosion in quanti-
tative terms has been discussed at great length
(see for instance Graf ), but is still not
solved. This is particularly true for such sub-
stances as disscussed here. In an attempt to
formulate a first and very simple model for the
erosion of particulate matter that previously
underwent aggregation and sedimentation the
following assumptions must be made:
The sedimented material is non-cohesive.
The deposited particles or floes do not
consolidate or change otherwise in shape
and size.
Only the total number of particles in sus-
pension is of interest and the identifica-
tion of individual particles is unimportant.
The considered systems, i.e. river segments,
are not stratified and can be considered as
completely mixed.
Conceptually two different situations can be
envisioned for the erosion process;(I) There
is an infinite reservoir of particles of all
sizes at the bottom of the river and (II) only
such amounts of particles will be resuspended
that have sedimented in a reasonable previous
time-span. In the first approximation the first
case can be formulated as follows:
£C= JL 5L
¦dt h 9x (s)
where: S = particle concentration of size d,
eroded at velocity v In numbers
per unit area
v = average longitudinal velocity
It is understood that this is only one of
many possibilities to formulate the stated fact
of erosion. Since the parameter S is not known
for realistic situations and may lead to comp-
lications in its determination, this case has
been abandoned, even though it is conceptually
attractive,
Case (II), where erosion is considered as
the reverse of the sedimentation process (tak-
ing the integrating role of time into account)
can be formulated as follows:
¦|£ = -t d-Rsed C fJ0)
where: dC/9> represents all particles that are
resuspended up to a critical size,
for instance given by the empiri-
cal equation of Mavis (cited by
Graf 11) vj.ta = o. is ((fi - ?>/? ) ¦
d W with v in m/s and d
in mm.
The factor 2 accounts for the pre-
viously sedimented particles up to
that critical diameter
f e + (id
sphere ; ^
With this conceptually derived, very simple ero-
sion model, the overall change in particulate
matter content of an'aqueous system due to dis-
persion, convection, aggregation, sedimentation
and erosion can be described by:
- *
-------
where particulate matter of a concentration
Cn_i enters and at the outflow the concentra-
¦Lr V9/tCn
-O
~&r.
- r,
r> -J
-o... (13)
tion of particulates is C )
. r jl ,rs*
t>f v~ "" s' y n
or:
7 f'n r ¦ 5 - /" /*: r*~ V
In the view of all other uncertainties in the
characteristic parameters, the assumption of
a steady-state continuous flow reactor appears
acceptable. The resulting equation (13) carv be
solved easier than the differential equation
(12) for each individual and consecutive river
segment.
The overall model describing the change in
particulate matter in an aquatic system (equa-
tion (12) or (13)) is very complex. Upon fur-
ther inspection it is seen that a series of
reaction steps affect a decrease of particu-
lates concentration (aggregation, sedimenta-
tion etc.) while other reaction steps (for
instance erosion) tend to increase the con-
centration of particulates. It appears logical
to simplify the conceptually derived model for
operational purposes (data fitting, informa-
tion compaction, data reproduction) by postu-
lating a "two-parameter-model" where one glo-
bal parameter accounts for all concentration
reducing processes while the other parameter
represents all concentration increasing pro-
cesses :
of -
City
k*
3,5-4,6-lo'6 sec.-1
k* /am. = 7,2*1 o~"sec. ~1
/or
fcJLtf/Ma-.- 3/0'J°~6/7'2'1O"'= 4,86.1o"2 =
for particulate matter of an average size
of 8-1o_G m as used for experiments repor-
ted in Figure (4).
The global rate constant for increase in parti-
culate matter (due to erosion etc.) can be
estimated in a similar way for flow velocity
v = o,2 m /sec. (see Figure (6)). For the
measuring campaigns of February 27, 19 74, and
February 12, 1975, one finds the following
values:
v = o,36 m/sec. k**£4,2-1o-6sec."
v = o,4o m/sec. k**S4,6 1o~6sec."
3£ ,0 .'32r
~ U 'dX* -ax
where: k* = global reaction rate constant
describing particulate matter
decrease
k#*(v) = global reaction rate constant
describing particulate matter in-
crease as function of the flow
velocity
The two parameters k* and k** must be deter-
mined in-situ from actual measurement AC at
given time increments At for characteristic
river segments. In the following paragraph,
such calculations are shown for illustration's
sake for the Neckar.
From empirical equations, such as Mavin's or
from.similar diagrams cited by Graf or Postma
' one can conclude that at flow velocities
below o,1 to o,2 m/sec. no significant erosion
will take place anymore. Taking two neighboring
Neckar cross—sections (Poppenweiler and Aldin—
gen -Figure (6)) one can determine k* from
the change in suspended solids for conditions
where v = o,1-o,2 m/sec. This is the case for
the measurement campaigns of February 2 and
March 3, 1975. A global rate constant for the
reduction in particulate matter k*=3,5-4,6«1o~6
sec. ~1 is found. This compares very well to
sedimentation rates determined in the labora-
tory (Figure (4)) if one takes into account
that in-situ sedimentation took place under
non-laminar conditions while laboratory sedi-
mentation presumably took place under laminar
conditions
These sample calculations and their results
are to be used with caution. They serve mainly
the purpose of illustration and should furnish
rough estimates of the relative impact of these
reactions on the distribution of particulate
matter. Furthermore, they demonstrate the fea-
sibility of quantitative modelling in2addition
to the large amount of existing data ' and
indicate at the same time which type of infor-
mation is needed.
XIX. Conclusions.
On the basis of the above discussion a
sample calculation on the transport and distri-
bution of Cadmium in the Neckar is presented.
In the course of such calculation a number of
assumptions have to be made which are listed
explicitly and which should serve as guide-
lines for further deliberations on monitor
programs and environmental assessment.
III.1. The fate of Cadmium in the Neckar as
an illustration for the role of particulate
matter in determining water quality.
The date given in Table (1) show impressive-
ly how dissolved cadmium as a representative
of other sorbing toxic materials through its
reactions with particulate matter behaves in
unforseen ways leading to unexpected phenomena.
While the calculations have more demonstrative
character, it may be assumed that the order of
size given by these figures is realistic.
REACTION
CONVECTION
(DISPERSION)
AGGREGATION
SEDIMENTATION
ERO-
SION
transport
in dissol-
ved form
L o,o74 ppm
H o,o74 ppm
Transport
with par-
ticulate
matter
L 37,oo ppm
H 1 11o,oo ppn
"Half-life"1
(relative
to total
amount)
clay 1,75-2,23
1o°C days
Clay o,35 days
3o C
"Doubling- I
time" (de- 1
sorption J
not can- J
sidered) |
Clay
1o°C
1,75
days
Table (1): The fate of cadmium in the Neckar.
Sample calculations for the illustration of
various significant, reactions caused by parti-
culate matter. The calculations are based on
the following assumptions:
1) particulate matter (filtrable material)
= 5o mcrfat Q" 6o-7om3/sec.
8 14-5
-------
2) dissolved fraction of Cd o,o74 ppm in
keeping with a reported concentration of
37 ppra (Forstner7) in inorganic sediments
3) erosion rate (assumed) v = o,5 m/sec.
4) L = low concentration factor (silica)=5oo
H = high concentration factor (humic ma-
terial) = 15ooo (Gardiner ).
It becomes convincingly clear that signifi-
cant amounts of cadmium are transported other
than in dissolved form (see also Gibbs )
that there may be an apparent reduction or re-
moval of cadmium due to its adsorption on ag-
gregating and sedimenting material with re-
markably short (half-life) times and that on
the other hand cadmium can be reactivated
through erosion and desorption (in this
instance only erosion was considered) with
significant rates expressed as "doubling time".
III. 2. Assumptions made throughout this
discussion leading to first suggestions
for further sensing programs.
(a) There is much information on particulate
matter in natural systems which is rarely
compatible (see paragraph X) and cannot
be exploited in strictly statistical ways.
(b) Quantitative information on pollutional
effects of particulate matter (change in
photosynthetic activity, equilibria and
rates of adsorption and desorption, etc.)
is scarce and confined to a few case stu-
dies without any indication on their ge-
neral validity.
(c) Despite a number of ongoing studies and
much previous work, the reactions of par-
ticulate matter itself are still not fully
understood or predictable in a quantita-
tive fashion. This is in particular true
for horizontal transport and for the rela-
tive contribution of turbulent dispersion,
for the aggregation behavior with respect
to the role of turbulence and complex in-
organic and organic media and finally for
the sedimentation and erosion processes
where the general validity of conceptually
derived sedimentation and erosion models
needs to be tested and empirical relation-
ships between flow velocity and sedimenta-
tion respectively erosion rates must be
reconfirmed.
From this brief summary of some of the as-
sumptions that had to be made, the following
very rudimentary and basic recommendations
concerning future sensing programs and improve-
ment of environmental assessment campaigns can
be deduced:
(1) Standardization of measurements (for compatability
of data)
(2) Change towards more characteristic and adequate
parameters for the description of particulate
matter as substitution for the very global para-
meter turbidity.
(3) Concurrent determination of water quality (parti-
culate matter etc.) parameters and water quantity
parameters for specific river stations:
particle siae distribution
specific surface concentration
adsorption/desorption characteristics
dispersion rate
convection rate
discharge rate
C+) Measurement of flux at specifically selected river
cross-sections with defined sources and sinks bet-
ween these cross-sections with all hydrographic
data available so that reliable and statistically
sufficient estimates on the global parameters
describing particulates reduction and particulates
increase can be made.
These first suggestions do not represent all
necessary measurements and investigations to
answer even the majority of the questions
raised by this discussion. On the other hand
it appears feasible to consider this list of
recommendations in ongoing or newly conceived
assessment programs, in particular if one
takes the significant role of particulate mat-
ter into account.
REFERENCES
(alphabetically)
(1) American Public Health Association: "Stan-
dard Methods for the Examination of Water,
Sewage and Industrial Wastes". APHA Inc.
New York.
(2) Applequist M.D., Katz A., Turekian K.K.:
"Distribution of Mercury in the Sediments
of New Haven (Conn.) Harbor". Environmen-
tal Science & Technology £>, 1 123-1 124 (1972)
(3) Bilinski H., Schindler P., Stumm W. , Zo-
brist J.: "Kupfer und Blei in natiirlichen
Gewassern" - Jahrbuch vom Wasser 4 3,1o7-
116 Verlag Chemie Weinheim (1974).
(4) Delichatsios M.A., Probstein R.F.: "Coa-
gulation in Turbulent Flow - Theory and
Experiment" MIT Fluid Mechanics Laboratory
Publication No.74-5 (1974).
(5) Fair G.M., Geyer J.C., Okun D.A.: "Water
and Wastewater Engineering" Volumes I and
IX, John Wiley & Sons Inc. New York (1966
and 1968).
(6) Fassbender C.P.: "Zur umwelthygienischen
Bedeutung des Schwermetalles Cadmium" Um-
weltschutz 6, 168-17o (1975).
(7) FSrstner U. , Miiller G.: "Schwermetalle in
Fliissen und Seen als Ausdruck der Umwelt-
verschmutzung". Springer Verlag Heidelberg
(1974).
(8) Gadow S., Sch&fer A.: "Die Sedimente der
Deutschen Bucht: KorngrSBen, Tonmineralien
und Schwermetalle". Senckenbergiana marit.
5, 165-178 (1973).
(9) Gardiner J.: "The Chemistry of Cadmium in
Natural Waters - II. The Adsorption of
Cadmium on River Muds and Naturally Occur-
ring Solids". Water Research 8, 157-164
(1974).
(10) Gibbs R.J.: "Mechanism of Trace Metal
Transport in Rivers", Science 18o, 71-73
(1973).
(11) Graf W.H.: "Hydraulics of Sediment Trans-
port", McGraw-Hill Book Company New York
(1971).
(12} Haberer K., Normann S.: "Metallspuren im
Wasser - ihre Herkunft, Wirkung und Ver-
breitung", Jahrbuch vom Wasser 2§» 157-182
Verlag Chemie Weinheim (1972).
(13) Hellmann H.s "Die Adsorption von Schwerme-
tallen an den Schwebstoffen des Rheins".
Deutsche GewSsserkundliche Mitteilungen 13,
9 I*'5
-------
1o8-113 (1969) and U, 42-47 (197o),
"Die Charakterisierung von Sedimenten
aufgrund ihres Gehaltes an Spurmetallen",
ibid. U, 16o-164 (197o) .
(14) Hohl H., Stumm W.: "Interaction of Pb+2
with Hydrous Y-AI2O3" (Preprint of Manus-
kript) EAWAG ZUrich (1975).
(15) Jernel£5v A.: "Release of Methyl Mercury
from Sediments with Layers Containing in-
organic Mercury at Different Depths".Lim-
nology and Oceanography 1_5,958-96o (197o).
(16) Lahav N.: "Adsorption of Sodium Bentonite
Particles on Bacillus Subtilis", Plant
and Soil 17., 191-2o8 (1962).
(17) Laubenberger G., Hartmann L.: "Speicherung
von Schwermetallsalzen in FluBsedimenten
und ihr EinfluB auf die Schlammbiocoenose"
Die Wasserwirtschaft JJ_,372-375 (197o).
(18) Lodemann C.K.W., Buckenberger U.: "Schwer-
metallspuren im Bereich des oberen Neckars"
GWF-Wasser/Abwasser 1 1 4,478-487 (1973).
(19) Mitchell R. Ed.: "Water Pollution Micro-
biology" John Wiley and Sons Inc. New
York (1972).
(20) Neis U.: "Experimentelle Bestimmung der
StabilitSt anorganischer Schwebstoffe in
nattir lichen GewSssern", Dissertation Uni-
versitat Karlsruhe (1974).
(21) Odum E.P.: "Fundamentals of Ecology", w.
B. Saunders Company Philadelphia (1959).
(22) Oliver B.G.: "Heavy Metal Levels of Otta-
wa and Rideau River Sediments", Environ-
mental Science and Technology 1_, 135-137
(1973).
(23) Postma H.: "Sediment Transport and Sedi-
mentation in the Estuarine Environment"
in "Estuaries" H. Lauff Ed. AAAS Publi-
cation No. 83 Washington (1967).
(24) Sakamoto W.: "Study on the Process of
River Suspension from Flocculation to
Accumulation in Estuary". Bull. Ocean
Res. Inst. Univ. Tokyo No.5 (1972).
(25) Schneider H.: "Zur Frage der Belastbar-
kait des Rheines mit radioaktiven Nukli-
den. IVj Sorption von Radionukliden durch
den planktonischen Anteil der Schwebstof-
fe des Rheines". GWF-Wasser/Abwasser 11o,
647-652 (1969).
10
14-5
-------
AUTOMATED FORECAST PROCEDURES
FOR RIVER QUALITY MANAGEMENT
Co-authored By
David A. Dunsmore
and
Robert J. Boes
The Ohio River Valley Water Sanitation Commission
(ORSANCO), an interstate compact created in 1948 by
eight states — Illinois, Indiana, Kentucky, New York,
Ohio, Pennsylvania, Virginia and West Virginia — to
advance water pollution control programs In the Ohio
River Basin, has made long-term use of a robot monitor
network which gathers water quality data from sites
along the Ohio River and its tributaries. In 1969,
in order to further enhance ORSANCO's pollution
abatement effectiveness, a three-year research
project was initiated. Its goal was the development
of automated water quality forecast procedures,
which would function with the electronic monitor units
already in use at ORSANCO. The outcome of this
effort, a totally dynamic water quality model
eventually named STREAM, could continuously evaluate
and analyze the influence of, for example, municipal
and industrial wastewater discharges, hydropower
generation, tributary waters, surface run-off and
stormwater overflows, and in addition provide timely
predictions of water quality resulting from discharges
to the receiving stream.
STREAM has been designed to answer three general
questions:
How does a given slug of water move and
* how does its quality state change with
time? [Called a RUN.]
What is the quality of the river in any
* given reach on a given day? [Called a
PROFILE.]
What was, is and will be the quality of
* the river at a particular place? [Called
a FORECAST.]
Each of these questions implies something about time
and distance. In a RUN, STREAM varies time and
distance simultaneously as the slug of water moves
downstream. In a PROFILE time remains constant and
distance varies between two locations. In a FORECAST
time changes from past to future at a given place.
The relationships among these distinct mechanisms are
graphically depicted in the following figure to show
how water quality states are determined as functions
of both time and place.
QUALITY SURFACE
In this model, water quality is defined as a set of
values (1. e. water quality states) which constitute
a surface above a distance coordinate in one direction
and a time coordinate in another. The distance
traveled by a slug of water in a specified period of
time is dependent upon stream hydraulics and stream
flow.
Whereas other mathematical models are either time or
event dependent, but not both, STREAM is distinguished
from these because it is both time and event dependent.
Events take place at specific locations along a river
while time dependent characteristics, such as river
flows and natural temperatures, relate to river
reaches, not specific locations. In STREAM, two
"clocks", acting as dual drive mechanisms, tick away
simultaneously, one indicating time and the other de-
picting distance. The time "clock" regulates con-
ditions which are time dependent and the distance
"clock" energizes appropriate events as they are en-
countered. When events are cyclically time dependent,
STREAM accommodates to weekly as well as seasonal
fluctuations. Additional "clocks" keep track of the
day of the week and day of the year, allowing a
Monday in January to be different from a Monday in
July, for example. The day-of-week clock also pro-
vides for weekly variations in daily waste loads from
specified discharges such as sewage treatment plants
or power generating facilities.
The realization that some modeling matters relate to
time, some to distance, and others to both time and
distance led to the conclusion that the model must be
capable of accepting time dependent information with
little regard for place, and place dependent informa-
tion with some regard for time. STREAM, therefore,
provides a place dependent data bank which has some
regard for time and a time dependent data bank which
is oriented to river reaches. Just in case some class
of data not readily adaptable to these data bases
exists, fish counts for example, a "user data bank"
has been made part of the model to facilitate user
processing of such information. Access to this and
existing user data banks is provided via user exits
and user oriented linkages.
Design Considerations and Philosophy
Initial modeling.considerations involved certain prac-
tical constraints, including the number of personnel
and existing machinery which could be utilized. Since
ORSANCO leased a small IBM 1130 and only employed a
staff of three data processing people, the size of
both the computer and data processing staff required
expansion. Computer facilities were subsequently
doubled and two additional persons hired, one devoted
to data processing and the other to automatic monitor-
ing. The basic project objectives for the model could
thus be implemented. These include the following
design criteria in order of importance. The model:
1) Must be easy to use - no computer experience
required; easy to obtain answers to "WHAT IF"
questions;
tRXVEB. MOD®1,
2) Must be flexible and easily adaptable to
change - modular in nature and coded in
FORTRAN IV;
3) Must be "macro" in nature - one dimensional
(only length);
4) Should be totally dynamic to represent the
"real" world in addition to the "static"
world;
1
14-6
-------
5) Should run in a reasonable amount of tine on
a small to medium size computer - leas than
30 minutes; and
6) Should be able to model any river for any
water quality characteristic desired.
Specific modules must be provided for
temperature, dissolved oxygen, conductivity
and chloride for ORSANCO use and the ability
to forecast must be Included.
Various modeling procedures, including continuous,
discrete, deterministic and stochastic techniques
were reviewed, leading to the conclusion that a
deterministic modeling approach would be best suited
to "real" world and "WHAT IF" situations. This de-
cision also meant that computer run times would be
kept to a minimum. The penalty associated with the
selection of any one choice, deterministic in this
case, is that some flexibility and adaptability to
change is reduced or at least temporarily given up in
favor of definition. The choice to model only river
quality naturally restricted STREAM to dealing with
flowing rivers rather than lakes, large reservoir or
estuaries, since modeling these would require mod-
ification of basic data bases and associated sub-
routine modules. Another natural limitation, the
use of normal American units of measurement, may
prove inconvenient to some, but this can be resolved
by the addition of appropriate mathematical conver-
sions.
The development of a deterministic model which has as
a primary goal ease of use, resulted in the following
significant decision - all modeling parameters and
variables would be assigned defaults which would de-
pict average river conditions during moderately low
flows experienced in the Ohio River near Cincinnati,
Ohio. This decision proved its worth many times
during the development and coding of STREAM because
it provided a convenient means of testing all model-
ing modules.
Overview of STREAM
The accompanying diagram is a general systems over-
view of STREAM. In reality, some 75 different com-
puter programs and subroutines comprise the actual
software required to fulfill all design objectives
set forth initially. The flow chart may be sub-
divided into three logical sections. The first,
CREAT, is broadly classified as the creation of the
river to be modeled, and is responsible for the es-
tablishment of actual discharges to be modeled in
addition to the hydraulic characteristics of the
river. This basic data base is referred to as the
default vector (DV) and includes a milepoint name
file and keyword file. The second, RDR (reader), is
given the task of identifying and then causing changes
to the original creation via an override mechanism,
to be temporarily established in a data area called
the single Btream default vector (SSDV). After a
determination of what modeling is desired, the reader
edits the SSDV and excludes any unnecessary informa«t4
tlon. The resultant data set is called an interrupt
vector (IV). The reader also accepts all modeling
control Information and all time dependent informa-
tion (DATA2). The third, EXEC (executive), performs
the actual modeling desired and only activates those
modules necessary for the specific run. For example,
if dissolved oxygen were to be modeled, then only the
first three modules, TRAVEL TIME, TEMPERATURE and
DISSOLVED OXYGEN, would be energised. Other unre-
quired modules are temporarily ignored.
STREAM SYSTEM OVERVIEW
INITIAL
DEFORMATION
CARDS
CHEAT
1 DEFAULT
MILEPOINT
KEYWORD I
(VECTOR
NAME FILE
FILE I
INTERRUPT
VECTOR
USER OPTION
DEFAULT FILE
AND CONTROL
INFORMATION
BATA2
piaa8 f
i USER EXIT TO ADDITION USER LfCKMATIOH I
modeling moduter
travel time
TEMPERATURE
dissolved CKYQEN
CONDUCTIVITY
CHLORIDE
PH
USER
USER EXITa
PRINTER
document
OUTOWJffiJTEgl
Modeling Modules
STREAK contains seven (7) explicit modules with six
(6) user exits, one exit between each module. These
modules, as shown on the systems overview diagram,
are 1) TRAVEL TIME, 2) TEMPERATURE, 3) DISSOLVED
OXTCEN, 4) CONDUCTIVITY, 5) CHLORIDE, 6) PH and
7> USER.
The travel time module (TRAVL)
This routine is the "heart" of STREAM. It regulates
all "clocks" by determining both time and distance.
The entire coordination of all events and their re-
lationship to time and place are totally dependent on
this module. No matter what water quality character-
istic is modeled or mode of operation is specified,
this module is always activiated. As such, it is
vital to STREAM and works In the following way. For
a given river reach, large or small, a basic premise
is assumed; average river velocity is linearly pro-
portional to distance at low river flows. Mathemat-
ically, then, this is expressed as velocity ¦ A plus
B times milepoint at a specified low flow. It is
assumed furthermore that a linear velocity increase
will result from higher river flows, For example, if
river flows are twice that of the specified low flow,
then velocity would be double the low flow velocity.
The coefficients 'A' and are symbols chosen to
2
14.6
-------
reflect average river hydraulic characteristics. The
accurate computation of these coefficients requires
river cross-section data. If this information is not
available, arbitrary coefficients may be selected and
a trial and error approach used in conjunction with
travel time study information. Hydraulic Engineers
may prefer to include the slope of the energy gradient
in addition to cross-section information as a means of
extending the river reach applicable to these two co-
efficients.
The temperature module (TEMPX)
This routine is activated when either temperature or
dissolved oxygen is modeled. The reason for activation
while modeling dissolved oxygen stems from considera-
tions used in the D.O. model, namely the temperature
compensation of various reaction rates and the need to
compute the temperature dependent oxygen saturation
levels in the river. The mathematical formula of the
temperature model is as follows:
T(new) » T(old) plus T(delta) times 10 raised
to the power of -k(t-t°).
Where T(new) is the new temperature, T(old) is the old
temperature, T(delta) is the temperature differential
from normal, t is present time, t° is the former time
and k is a decay constant whose default value is one
(1). The validity of this formula has recently been
verified by Novotny and Krenkel at Vanderbi.lt Univer-
sity.
The dissolved oxygen module (DOBOD)
This routine is activated when dissolved oxygen or
oxygen deficit is modeled. The basic mathematical rep-
resentation used to model DO is a modified Streeter-
Phelps equation and Includes carbonaceous oxygen de-
mand, nitrogenous oxygen demand, land run-off and
benthai oxygen demand terms. Provision is made within
this module for the selection of one of two different
methods of dynamically computing reaeration rates,
either that used by Churchill (the default method) or
that used by 0'Conner. Both of these reaeration
methods require a knowledge of river velocity and
depth. Velocity is available from the travel time
module, and depth must be expressed as a function of
river milepoint using two coefficients, A and B, in a
fashion similar to velocity.
The chloride module (CLMOD)
This routine is active only when chloride is modeled,
because it may be assumed that once chloride enters
the stream it never leaves. The module is basic mixing
model and the chloride concentration is held constant
if no chloride loads enter the river.
The conductivity module (COND)
This routine is invoked only if conductivity is to be
modeled, since it is basically a mixing model (like
the chloride module). However, both chlorides and
dissolved solids loads affect conductivity. If such
loads are encountered, these are converted to "con-
ductivity equivalents" and mixed into the overall con-
ductivity of the river.
The last two modules are not implemented, as no re-
liable means of modeling pH has been identified and
the USER module must be tailored to a particular pur-
pose and suited to the modeler's needs. The USER
module has been included as a convenient vehicle for
modeling water quality variables not explicitly in
STREAM. In connection with this, a user data file
has been integrated with STREAM, thus permitting any
water quality information to be modeled. The user
exits between explicit modeling modules are provided
to .illow for easy modeling of specified aberrations
associated with unique river phenomena. For example,
at ORSANC0 the modeling of reaeration at high level
dams has utilized user exit three(3) in a fashion
that permits dissolved oxygen modeling to continue
up to a dam and then increases the oxygen concentra-
tion as the water flows over the dam. User exit
two(2) has been utilized to change tributary BOD
loads in proportion to mainstem river flows.
Any combination of these modeling modules may be
selected during any single model run. The default is
to model all conditions.
A brief review of some additional features not part
of the system overview but nevertheless contained
in STREAM is included here. These illustrate how
ease of use has been accomplished. Most significant
is the command structured dialogue used by STREAM in
conjunction with a KEYWORD FILE. This tool or medium
establishes a means of telling STREAM what to do in a
manner resembling English. For example, MODEL DIS-
SOLVED OXYGEN is how STREAM may receive an input
record. This record is then translated by the reader
and appropriate computer instructions are issued to
the executive, thus causing modeling of dissolved
oxygen. An associated KEYWORD FILE extension provides
for the inclusion of keyword synonyms and is particu-
larly useful for accepting abbreviations and commonly
misspelled words and as a language Interchange mechan-
ism. A Frenchman, for instance, may elect to provide
the French words for DISSOLVED, MODEL, and OXYGEN, in
addition to all other keywords, and then communicate
with STREAM in his native tongue. The major con-
straint here is the syntax of the English language,
because the order of command statements may not be
altered. Three of STREAM'S modules would require
changes to accommodate syntactical differences between
languages if English syntax proved to be too awkward.
Another feature provided as an aid to the modeler
using STREAM is the Error Diagnostic Message handler.
It transmits three types of messages:
1. Operator messages - informing an operator
what to do!
2. Warning messages - informing the modeler
that there is some inconsistancy with
respect to requested modeling activities
but that certain assumptions have been
made and conditionally carried out;
3. Fatal error messages - informing the modeler
that STREAM cannot resolve a defined
situation, such as a misspelled keyword or
its synonym, and that modeling efforts have
been terminated.
One additional feature deserves mention because of
differences in hardware configurations. STREAM pro-
vides the user with an opportunity to specify a partic-
ular machine configuration unique to the installation
of his own system. This routine -SYGEN- is part of the
initial system startup procedure. It permits STREAM
to be independent of particular card readers or
printers, thereby enhancing portability and transfer-
ability to other machines. In this regard, it is
appropriate to note that STREAM has been installed on
the following kinds of computers:
3
14-6
-------
1. CDC 6500, 6600;
2. IBM 1130, 360/65, all 370's using virtual
systems;
3. Meta IV; and
4. GA 1830.
Model Calibration Techniques and Experiences
Ideally, one would like to measure all water quality
characteristics above and below each and every dis-
charge continuously, in addition to all discharge
characteristics. This information could be modeled
and the model "tuned" to fit "real world" values. In
reality, however, additional information is required
including such basic facts as river flows and cross-
sections. This still may not be enough. The temp-
erature module may also need wind velocity, solar
radiation intensity or cloud cover information. In
addition, the following questions may need quantifi-
cation: Has a thermal mixing zone been defined as
a function of flow? Do various discharges adhere
to one side of the river? Is there any evidence of
stratification? How many sample sites should there
be in each cross-section? Is the sample represent-
ative? The necessity of quality assurance on every
data item is equally important. Even if all this
were possible, how does one assess the contribution
from land run-off (non-point sources)? Can land run-
off be segregated into natural and man made components?
ORSANCO's solution to the problem of model calibration
leaned heavily toward the practical side of all these
issues. First, daily average values of many cooperat-
ing facilities would be used as discharge loads, dally
average river flows would be sufficient, hourly water
quality data from automatic monitors would be
utilized, macro modeling techniques including complete
mixing at the discharge point would be assumed and
land run-off would be presumed to be zero during
periods of low flow and little rainfall.
The next question relates to the comparative accuracy
of the model. Many would prefer that a model always
be exactly right, especially a computer model. If this
feat could be accomplished, a model would not be re-
quired. The set of facts would be known always and
forever. What then is good enough - 5 percent, 10
percent, 25 percent, one degree or one mg/1? The
answers to this question vary with each person asked;
generally, however, attitudes lean heavily toward a
graduated scale defining plus or minus 5 percent as
superb, 10 to 20 percent as usual and 25 percent as the
lower limit of general acceptability. One caution must
be considered when judging a model's worth via calibra-
tion; it does not necessarily follow that a model is
valid because it meets a predetermined set of criteria.
It must also be mathematically correct or varying
circumstances will yield totally unrealistic results.
This occurred at ORSANCO as the result of a keypunch
error in the dissolved oxygen modeling module which
transposed two digits in a constant. The modeling re-
sults were better with the keypunch error than with
the corrected value. This then raised the further
question of whether or not the default to Churchill's
expression was still to be preferred over O'Conner'a,
No change has been made in the original default
choice, but the differentials are no longer as great
as they once were.
MODEL APPLICATIONS
Practical uses of STREAM at ORSANCO are based in pre-
vious applications which have included investigations
of temperature and dissolved oxygen. The very first
study involved an evaluation of a major new power
generating facility which has a 2,400 magawatt capac-
ity (1,800 megawatts are operated with once-through
cooling). This plant's potential effect on river
temperatures needed to be appraised in order to de-
termine whether or not ORSANCO pollution control
standards would be violated.
The second study was aimed at helping the eight
compact states and the Federal Environmental Pro-
tection Agency to assess what, if any, waste load
allocations would be required to keep dissolved
oxygen levels for the Ohio River at, or above, estab-
lished standards upon completion of required treatment
facilities. This effort entailed the evaluation of
carbonaceous and nitrogenous oxygen demand and their
associated reaction rates, temperature effects on
dissolved oxygen, effects of differing reaeration
equations, effects of river flow variations and of
reaeration at 19 high level navigation dams.
Three other completed dissolved oxygen studies deal
with specific river reaches identified in the waste
load allocation report. In each case, study areas
are associated with oxygen levels which occasionally
fall below established standards. The area between
Huntington and Cincinnati was selected for an inten-
sive appraisal of the socio-economic impact of the
1972 Federal legislation (PL 92-500).
Other reported uses of STREAM
A primary goal in model development, overall flex-
ibility and adaptability to change, has been achieved
through the use of modularity. Since modules are
relatively easy to change, STREAM has been used for
three other cited purposes - the modeling of copper
tailings, schoolroom facilities over a five year
period in the FORECAST mode, corporate fiscal resources
relying upon one modular change. Although the model
was not initially designed for these purposes, they
established the versatility of STREAM.
CONCLUSIONS
The successful development of a general and dynamic
river quality model, STREAM, enhances the use of
electronically received water quality data at ORSANCO
and effectively supplements water quality management
activities throughout the Ohio River. Because it
operates on the basis of appropriate river geometry
definitions in conjunction with waste load discharges,
STREAM can be applied without the use of sophisticated
data acquisition systems. Depending on the modeling
objective and the UBer's choice, STREAM nay accept
any combination of constant or dynamically variable
data to model travel time, river temperatures, dis-
solved oxygen, conductivity and chloride in one of
three operational modes (RUN, FORECAST or PROFILE).
Additional variables may be modeled by means of op-
tional user exits and the user data bank, thus giving
STREAM the capability of depicting any water quality
state. The free form of input, combined with the
availability of the Command Structured Dictionary,
allows the individual who may have little computer
expertise to operate the model effectively and
efficiently. The scope of STREAM is further expanded
because the size or complexity of the river being
modeled is unlimited,
CREDITS
The development of STREAM was partially supported by
4
14-6
-------
the Office of Research and Monitoring, U. S. Environ-
mental Protection Agency, Washington, D.C. (Project
16090 DHX).
5
14-6
-------
A STRATEGY FOR AQUATIC POLLUTANTS FATE AND TRANSPORT DETERMINATION
Walter M. Sanders, III
Environmental Research Laboratory
U. S. Environmental Protection Agency
Athens, Georgia 30601
The U. S. Environmental Protection Agency
was established by Congress in 1970 with the
mandate to develop, coordinate, and execute a
national environmental protection program.
Although many complex problems confront the
Agency, one of the largest and most difficult
to solve relates to the release of chemicals
into the environment through deliberate,
incidental, or accidental discharges.
The sheer numbers of chemicals that must
be considered are staggering with
approximately 2,000,000 unique materials
reported in the literature between 1965 and
1972. Rapid advancements in analytical
techniques have made it possible to isolate
and identify hundreds of additional chemicals
in environmental samples each year. in
recent EPA studies, for example, more than
200 compounds were identified in public water
supplies, several of which are potentially
dangerous halogenated hydrocarbons resulting
from the long accepted practices of purifying
drinking water by chlorination.
Many of the pollutants in the environment
are present in microgram or nanogram per
liter quantities and only a small percentage
may produce significant adverse human health
or environmental impacts. Therefore, the
development of effective and efficient
Srotection programs depends on the ability to
etermineor predict these adverse impacts so
that potentially harmful compounds can be
identified and corrective action can be taken
before serious damage occurs.
The operating program of EPA, in carrying
out their mandated advisory, regulatory, and
enforcement duties within the time and
budgetary constraints imposed by the Federal
Water Pollution Control Act Amendments of
1972 and Congress, must provide the
scientific data required to answer the
following questions:
• Which chemicals pose a significant
risk to human health or the
environment?
• How are the risks related to their
concentrations?
• Do the chemicals persist or concen-
trate in the environment?
• Can the materials be removed or
controlled?
• What are the treatment costs in
dollars, resources, and energy?
• Which materials should be regulated or
barred from the environment, and what
are the socio-economic implications of
such actions?
Because of time and budget constraints
and the large numbers of compounds to be
considered, priorities must be established
and a hierarchal system of tests developed so
that the investigative efforts expended on a
particular chemical will be proportional to
the degree of environmental hazard and/or the
significance of the material in the national
economy or in energy production.
Initial toxicity screening tests may be
sufficient to identify the most hazardous
materials. Other chemicals that produce
chronic or more subtle health or
environmental effects may require detailed
transport fate and effects data for
regulatory or enforcement decisions.
One strategy for obtaining such transport
and fate information for aquatic pollutants
while attempting to minimize the expenditure
of time and resources is being applied and
evaluated at the EPA's Environmental Research
Laboratory, Athens, Georgia. Although
developmental research is still underway on
some of the processes and procedures, a
discussion of the strategy should be useful
to others involved in similar research.
In developing the strategy, the following
basic assumptions were made:
First, the potential environmental impact
of a specific pollutant in an aquatic system
is assumed to be a function of its location,
concentration* and chemical structure at any
point in time after its initial discharge.
Second, the location, concentration and
chemical form of the pollutant at a given
point in time are assumed to be the net
result of a set of basic physical, chemical,
and biological transport, degradation, and
transformation processes that act on the
pollutant either simultaneously or in
sequence.
Third. although all of the basic
physical, chemical, and biological processes
could theoretically affect the pollutant,
only a few may be environmentally
significant. The magnitude and rates of
those actually involved are functions of both
the chemical structure of the pollutant and
the biotic and abiotic conditions that exist
in the particular aquatic environment into
which it is discharged.
Fourth, it is assumed that the most
significant physical, chemical, and
biological processes affecting the pollutant
can be identified and their individual
mechanisms, rates, and interactions
determined in the laboratory.
Fifth. it is assumed that these
individual process data can be incorporated
into "evaluative" models that can be used to
estimate the net result of the competing
processes for various sets of environmental
conditions.
1
14-7
-------
Sixth, with the current state-of-the-art,
evaluative models cannot be expected to
provide precise "realtime" rate and
concentration data for given water bodies.
However, with proper simulations, decision
makers can be provided with a range of values
including estimates of upper and lower limits
that can be extrapolated to natural water
systems.
The initial or laboratory phase of the
strategy includes a sequence of coordinated
research tasks designed to minimize the
resources and effort expended on each
pollutant. Studies are conducted in four
major process areas to determine whether or
not degradation or transformation reactions
occur and if their rates are fast enough to
be environmentally significant. When
negative results are obtained, research in
that particular area is terminated. When
positive results occur, detailed mechanistic
and kinetic studies are conducted with each
succeeding task direction based on the
previous task results. The four major
process areas are
A. Microbial Uptake and/or Degradation
Processes,
B. Chemical Degradation and/or Trans-
formation Processes,
C. Photochemical Degradation and/or
Transformation Processes,
D. Physical Transport Processes.
The research tasks in each area are as
follows:
A. Microbial Processes
1. Initial microbial research includes
screening studies for pollutant
degradation and uptake-accumulation
using mixed populations of aquatic
microorganisms (bacteria, fungi,
and plankton).
2. Degradation studies include
enrichment techniques for degrading
populations from five aquatic sites
with acclimation periods up to six
weeks. The five aquatic sites are
chosen to represent
• an unchlorinated sewage effluent
• a site receiving discharges of
the pollutant under study, if
possible
• a eutrophic water body
• an oligotrophic water body
• another relevant site of the
researcher's choice.
Further microbial studies are
eliminated if no degradation is
observed.
3. Detailed studies include
determination of degradation rates
using pollutants as the sole carbon
source under highly dilute aqueous
conditions (below the water
solubility of the pollutant).
Rates are reported as functions of
microbial mass.
4. Rates and extent of sorption and
desorption are determined when the
accumulation is greater than 10-
fold.
5. Primary degradation products are
identified.
B. Chemical Processes
1. Initial chemical degradation
research includes screening studies
to determine if chemical
degradation and transformation
reactions occur in dilute aqueous
soluti on.
2. For hydrolytic reactions, studies
are terminated for materials that
are not 50? degraded in 48 hours at
6S°C and at pH's 2 and 12.
3. For those materials that degrade, a
pH rate profile is determined from
pH 3 to 11.
4. Activation parameters are measured
for each significant hydrolytic
mechanism between pH's 5 and 8.
(Rate constants measured in 3 and 4
above should have a precision of +_
204. If carrier solvents are used as
buffers, their effect on the reaction
must be shown to be negligible.)
5. Reaction products formed at one and
ten half-lives are identified.
6. Stability to molecular oxygen is
established by determining the
extent of degradation over a six-
week period in air-saturated
distilled water in the presence of
light. In the event of
degradation, products are
identified.
C. Photolysis studies in dilute aqueous
solutions are conducted for all
materials having a hydrolysis half-
life greater than 5 hours at pH's
between S and 9 and temperatures below
35°C.
1. Electronic absorption spectrum is
determined for each compound in
water solution and the extinction
coefficients are integrated over
10-nm bands from 300-800 nm.
2. Quantum yields are measured for
direct photolysis in air-saturated
water. Monochromatic light of
wavelengths greater than 300 nm is
used. Effects of pH (3-9) on
quantum yields should be
determined.
3. Products resulting from photolysis
in air-saturated water are
identified.
2
14-7
-------
4. Photodecomposition rates (X > 30O
nm) of the pollutants in distilled
wateT are compared with the rates
of degradation in three freshly
collected natural waters and
distilled water containing humic
acid. One of the natural waters
must contain detectable amounts of
chlorophyll. For those solutions
in which photolysis occurs more
rapidly than in distilled water,
products are identified.
D. Physical Processes
1. Physical transport from water to
air is evaluated as a function of
oxygen reaeration coefficient.
2. Henry's law constant and the
solubility of each compound in
water are measured and a solubility
pH profile is developed.
5. The extent and rate of sediment
sorption from water should be
determined with sediments
representing extremes in organic
carbon content and pff.
The final or evaluative phase of the
strategy includes the collection and
incorporation of the significant process data
into a rational framework so that the net
results of the various competing or enforcing
reactions can be observed simultaneously.
Since it is virtually impossible to handle
such complex data manually within a
reasonable period of time, the various rate
processes are described mathematically and
the resulting equations coupled to form
"evaluative" computer models. These models
do not represent any particular aquatic
system but are structured so that each
important environmental factor (light,
temperature, pH, flow, biomass, etc.) can be
varied within the ranges observed in nature.
Thus, with judicious model simulations
combining hypothetical environments and
observed process data, a range of net
transport, transformation and/or degradation
rates can be estimated. These model results
can be presented as concentration
(persistance) and distribution (accumulation)
profiles for parent compounds and for
degradation products.
The evaluative models can also be used to
indicate which of the major fate and
transport processes are most significant for
a given chemical under specific conditions so
that the effectiveness of various management
strategies can be evaluated.
This strategy has been developed over the
past several years through the combined
efforts of an interdisciplinary team of ERL
scientists studying the aquatic transport,
degradation, and transformation processes for
both organic and inorganic chemicals. During
this period, thirteen organic compounds and
one element (mercury) were evaluated using
vairious combinations of the analytical
procedures outlined._ However, since the
methods have been continually evaluated and
improved and additional processes studied and
included in the strategy, no single chemical
has been processed through the complete
sequence of screening and kinetic
investigations. This is particularly true of
the sediment sorption studies that have just
been added. Table 1 shows a list of the
chemicals tested and the significant
transport or transformation processes that
have been identified for each. A more
complete description of the research
procedures used and the data obtained are
presented in the EPA Reports and scientific
papers listed in the Bibliography under
Additional References,
Table 1. SIGNIFICANT DEGRADATION PATHWAY
Compound
Microbial
degradation
sorption
chemical
hydrolysis
tUtrosaticn
photochemical
degradation
Vdlfltilit&t iAll
Axraiine
X
Toxapherie
X
Diczinon
2,4-D Es?«t
X
X
X
X
Methoxychlor
X
X
X
•X
CIPG
X
IPC
X
HCB
X
X
X
MzlAthion
X
X
PCNB
x
X
X
Carbaryl
X
Vinyl Chloride
Cvotan
The first or laboratory phase of the
strategy is currently being applied by an EPA
contractor to three organic compounds. The
investigation represents the first in a
series of studies of a large number of energy
related pollutants. Results should be
available in early 1976,
In addition to the laboratory studies
described, microcosms' also have great
potential for use as a rapid and inexpensive
technique for screening pollutants for
possible fate and transport pathways.3 After
more thorough evaluations, they may be
included in the initial phase of the
strategy. At this time however, microcosms
cannot provide the detailed fate and
transport data required to support regulatory
and enforcement decisions for most
pollutants. It is practically impossible to
include all of the significant processes and
driving forces in simple microcosm systems,
and the required process rate data cannot be
obtained from microcosms in which either the
components are added sequentially or the
biological compartments are maintained under
nutritional or environmental stress
conditions.
The evaluative models being developed by
ERL scientists as part of the strategy
include (1) a systems analysis model for
material that only undergoes volatilization,
and (2) an organic chemical transport and
degradation process model.
The physical, chemical, and biological
screening studies conducted with vinyl
monomer indicated that the only
process that the chemical
undergoes is volatilization from an aquatic
system. Detailed laboratory volatilization
studies were conducted and
chloride
significant
a).*, the resulting
first-order rate coefficient was expressed in
terms of the re- ""
since both
terms
oxygen
of the reaeration coefficient for
processes involve the
14-7
-------
exchange of gas molecules at the air-water
interface.
Simple flow and mixing models were
developed .for theoretical lake and river
ecosystems4 and the vinyl chloride half-life
was estimated for a range of conditions for
each system.
A much more complicated process model is
currently under development and evaluation
using the organophosphorus pesticide
malathion as the prototype chemical. In this
case, laboratory screening studies indicated
that malathion can be degraded by chemical
hydrolysis and humic acid sensitized
photolysis reactions and by bacterial,
fungal, and algal metabolic processes. Rate
data from the detailed kinetic experiments
showed, however, that only the chemical
hydrolysis, bacterial degradation, and
possibly algal degradation processes are fast
enough to be competitive under most
environmental conditions. These experiments
also showed that (1) the hydrolysis rate is
strongly affected by pH and temperature, (2)
bacteria can use malathion as a sole carbon
source (much higher degradation rates are
obtained, however, when additional organic
substrates are present), (3) for degradation
to occur in algae cultures, malathion must be
in contact with the cells during periods of
photosynthesis, and (4) the major degradation
product in aquatic systems is the malathion
monoacid (the biological processes result in
the g isomer while chemical hydrolysis
produces the a isomer of the monoacid).
The malathion model has been developed
around a river ecosystem and, to date,
subroutines for both chemical hydrolysis and
bacterial degradation have been incorporated
and tested. A subroutine for algal breakdown
is nearing completion. Results from this
model are being compared with data obtained
from experiments conducted in the Aquatic
Ecosystem Simulator (AEcoS) where similar
environmental conditions are established and
maintained. Both processes were tested
individually and in combination and, in each
case, the model predictions of the rate of
malathion degradation were within
approximately 251 of the observed AKcoS
data,^ In each case the rates observed in
the AEcoS were lower than the model
predictions based on the laboratory studies.
The discrepancy may be due to a difference in
mixing rates between the slow-moving stream
in the AEcoS and the rapidly stirred
laboratory reaction flasks and chemostats.
This hypothesis is being tested.
The agreement between the model outputs
and the AEcoS data thus far gives a strong
indication that information obtained in
separate laboratory process experiments can
be combined in simulation models to provide
good estimates of the persistence,
distribution, and form of pollutants under
investigation.
Research is continuing at ERL-Athens to
identify and describe other fate and
transport processes that may be significant
in aquatic environments and to improve the
evaluative models by developing,
incorporating, and testing subroutines for
phytoplankton, sorption, and photolysis
processes.
In conclusion, this strategy provides a
logical system for obtaining and evaluating
significant fate and transport information
for specific pollutants. It is still under
development but preliminary results indicate
that it can be applied successfully.
Acknowledgement s
The author gratefully acknowledges the
assistance of his colleagues on the ERL-
Athens staff for their contributions in
developing and evaluating this strategy and
in the preparation, review, and editing of
this paper.
Bibliography
1. Metcalf, R. L., G. K. Sanghan, and I. P.
Kapoor. Model Ecosystem for the
Evaluation of Pesticide Biodegradability
and Ecological Magnification.
Environmental Science and Technology.
_5: 709, 1971.
2. Sanborn, J. R. The Fate of Selected
Pesticides in the Aquatic Environment.
EPA Research Report. EPA-660/3-74-02S,
December 1974.
3. Sanders, W. M. EPA Needs in Microcosm
Research. Symposium: The Role of
Microcosm Research in Ecology. To be
published in the proceedings of the AIBS
Meeting, Oregon State University,
Corvallis, August 17-22, 1975.
4. Hill, J., H. P. Kollig, I). E. Paris, N.
L. Wolfe, and R. G. Zepp. Dynamic
Behavior of Vinyl Chloride in Aquatic
Ecosystems. EPA Report. In preparation.
5. Falco, J. W., D. L. Brockway, K, D.
Sampson, H. P. Kollig, and J. R.
Maudsley. Models for Transport and
Transformation of Malathion in Aquatic
Systems. Symposium: Freshwater Quality
Criteria Research of the EPA. To be
published in the proceedings of the AIBS
Meeting, Oregon State University,
Corvallis, August 17-22, 1975.
Additional References
Baughman, G. L., M. H. Carter, N. L. Wolfe,
and R. G. Zepp. Gas-liquid
Chromatography-Mass Spectrometry of
Organomercury Compounds. Journal of
Chromatography. 76:417-476, 1973.
Baughman, G. L., N. L. Wolfe, R. G. Zepp, and
J. A. Gordon. Chemistry of
Organomercurials in Aquatic Systems. EPA
Research Report. EPA-660/3-73-012,
September 1973,
4
14-7
-------
Cox, M. F., H. W. Holm, H. J. Kania, and R.
L. Knight. Total Mercury and
Methylmercury Concentration in Selected
Stream Biota. Submitted to the Journal
of the Fisheries Research Board of
Canada.
Hill, J. and R. R. Lassiter. Development of
Transport Model for Pesticides in Aquatic
Environments. Published as part of an
EPA Research Report. EPA-66Q/3-74-024,
December 1974. pp. 40-54.
Holm, H. W. and M. F. Cox. Simple Method for
Introducing Elemental Mercury into
Biological Growth Systems. Applied
Microbiology. 2j^(3) :662-623, March 1974.
Holm, H, W. and M. F. Cox. Mercury in
Aquatic Systems: Methylation, Oxidation-
Reduction, and Bioaccumulation. EPA
Research Report. EPA-660/3-74-021,
August 1974. 39 p.
Holm, II. W. and M. F. Cox. Transformation of
Elemental Mercury by Bacteria. Submitted
to Applied Microbiology.
Lewis, D. L. and D. F. Paris. Direct
Determination of Carbaryl by Gas
Chromatography Using Electron Capture
Detection. Journal of Agricultural and
Food Chemistry. 2j2(l):148, January-
February 1974.
Lewis, D. L., D. F. Paris, and G. L.
Baughman. Transformation of Malathion by
a Fungus, Aspergillus orvzae. Isolated
from a Freshwater Pond, Submitted to the
Bulletin of Environmental Contamination
and Toxicology.
Paris, D. F. and D. L. Lewis. Chemical and
Microbial Degradation of Ten Selected
Pesticides in Aquatic Systems. Residue
Reviews. 4_5:95-124, 1973.
Paris, D. F., D. L. Lewis, J. T. Barnett, and
G. L. Baughman. Microbial Degradation
and Accumulation of Pesticides in Aquatic
Systems. EPA Report 660/3-75-007, March
1975.
Paris, D. F., D. L. Lewis, and N. L. Wolfe.
Rate of Degradation of Malathion by
Bacteria Isolated from an Aquatic System.
To be published in Environmental Science
and Technology.
Paris, D. F. and D. L. Lewis. Rates and
Products of Degradation of Malathion by
Bacteria and Fungi from Aquatic Systems.
To be published in the Journal of
Environmental Quality and Safety.
Wolfe, N. L., R. G, Zepp, J. A. Gordon, and
G. L. Baughman. Chemistry of
Phenylmercury Compounds in the Aquatic
Environment. Chemosphere. If6):273-278,
1972.
Wolfe, N. L., R. G. Zepp, J. A. Gordon, and
G. L. Baughman. Chemistry of
Methylmercurials in Aqueous Solutions.
Chemosphere. £: 147-152, 1973.
5
Wolfe, N. L., R. G. Zepp, G. L. Baughman, and
J. A. Gordon. Kinetic Investigation of
Malathion Degradation in Water. To be
published in the Bulletin of
Environmental Contamination and
Toxicology.
Wolfe, N. L., R. H, Cox, and J, A. Gordon.
Synthesis and Carbon-13 Studies of
Malathion Acid Derivatives. Submitted to
the Journal of Agricultural and Food
Chemistry.
Wolfe, N. L., R. G. Zepp, J. A. Gordon, and
R. C. Fincher. N-Nitrosamine Formation
from Atrazine. Accepted for publication
in the Bulletin of Environmental
Contamination and Toxicology. 197 5,
Wolfe, N. L., R. G. Zepp, G. L, Baughman, and
J. A. Gordon. Kinetics of Chemical
Degradation of Malathion in Water. To be
submitted to Environmental Science and
Technology.
Wolfe, N. L., R. G. Zepp, J. C. Doster, and
R. C. Hollis. Captan Hydrolysis. To be
submitted to the Journal of Agricultural
and Food Chemistry.
Zepp, R. G., N. L. Wolfe, and J. A. Gordon.
Photodecomposition of Phenylmercury
Compounds in Sunlight. Chemosphere.
3:93-99, 1973.
Zepp, R. G., N. L. Wolfe, G. L» Baughman, and
J, A. Gordon. Chemical and Photochemical
Alteration of 2,4-D Esters in the Aquatic
Environment. To be published in the
Journal of Environmental Quality and
Safety.
Zepp, R. G., N. L. Wolfe, J. A. Gordon, and
G. L. Baughman. Dynamics of 2,4-D Esters
in Surface Waters: Hydrolysis, Photo-
alteration, and Vaporization. Accepted
for publication in Environmental Science
and Technology.
Zepp, R. G., N. L. Wolfe, J. A. Gordon, R. C.
Fincher, and G. L. Baughman. Chemical
and Photochemical Transformations of
Selected Pesticides in Aquatic Systems.
In preparation for EPA Research Report.
Zepp, R. G. and N. L. Wolfe. Direct
Photolysis of Methoxychlor. In
preparation.
14-7
-------
REMOTE SENSING OF GROUND AND SURFACE WATER
CONTAMINATION BY LEACHATE FROM LANDFILL
by
Dwight A. Sangrey
Wi 11 iam L. Terig
Warren R. Phi 1ipson
Ta Liang
School of Civil and Environmental Engineering
Cornell University, Ithaca, New York
Summary
This paper describes a research activity in which
a group of 15 landfill sites, in central New York
State, were monitored during a coordinated program of
five remote sensing missions. Aircraft flights and
ground sampling were scheduled at key times during a
one-year period, and the aerial data included color
and color Infrared films and thermal imagery. The
primary research objective is to develop remote sensing
procedures for detecting and evaluating ground and sur-
face waters contaminated by landfill leachate.
Each type of aerial data proved useful in monitor-
ing leachate or contaminated water, but their effect-
iveness varied considerably--especia11y with season.
The initial results and conclusions are summarized.
This work is part of an ongoing study being supported
by NASA, New York State Department of Environmental
Conservation and EPA.
Introduct ion
Leachate produced from the land disposal of solid
waste constitutes a significant environmental problem
(1). Selecting the most appropriate sites, designing
landfill to minimize leachate production and develop-
ing methods to collect and treat leachate are elements
of the problem that are receiving considerable atten-
tion. One unique aspect of the problem is the develop-
ment of an effective approach to monitoring sites to
determine whether leachate is being produced and to
what degree ground and surface waters are being con-
taminated.
Several characteristics of leachate production
point to the appropriateness of remote monitoring tech-
niques. The first is the widespread use of the land as
a solid waste disposal medium. Sites are typically
scattered over large areas, such that control or even
cataloging by regulatory agencies is difficult. Sec-
ondly, at any site leachate is generally not a point
source in the usual sense of effluent discharges.
Leachate may break out of the top or side of a landfill
directly or it may flow beneath the ground surface for
some distance before coming to the surface. Ground in-
spection of the area around a landfill for leachate
contamination is expensive, tedious and usually not
practical. The third important parameter in leachate
production is time. Until sufficient moisture has
infiltrated a landfill to reach its field capacity, no
leachate will flow from the refuse. The time necessary
to achieve field capacity varies considerably depending
on many site and climatic factors, but several years
will often pass before a landfill begins to produce
leachate. By this time the site may have been closed,
with the operators and regulatory agencies assuming
that there is no leachate problem.
The primary objective of this study is to develop
remote sensing procedures for detecting and evaluating
ground and surface waters contaminated by landfill
leachate. While some limited use of remote sensors for
monitoring of leachate has been attempted by previous
investigators (2), no comprehensive research effort had
been directed toward the subject. The Investigation is
being conducted in central New York State, an area
characterized by a humid continental climate. Of par-
ticular interest in this study was the relative effect-
iveness of various sensors during different times of
the day and year. Ground control in support of the
aerial missions included temperature measurements of
air, land and water, plus comprehensive physical and
chemical testing of leachate and contaminated waters.
Site and Scope of the Study
The site of the remote sensing research study is a
corridor roughly encompassing the area between Syracuse
Table I. Date, nominal scale and other parameters of imagery
obtained In EPA flights for the landfill investigation
A11 itude
Date Color Color IR Thermal above ground Notes
ft.
Sept.
26
1974
1:5,000
1:10,000
1:9,000
1:17,000
1:9,000
2,500
5,000
2,500
Quality of thermal 6
Partial coverage (8 sites) i
Night (2100 hr.) ' photo data poor
Dec.
6
1974
1:2,000
1:2,000
1:3,500
1:5,500
1:9,000
1,000
1,500
2,500
Partial coverage (6 sites)
Night (2100 hr.); partial coverage (5 sites)
Night (2000 hr.)
Mar.
23
1975
1:5,000
1:5,000
1:9,000
1:9,000
2,500
2,500
Night (2300 hr.)
Apr.
22
1975
1:7,000
1:7,000
1:12,000
1:9,000
3,500
2,500
Predawn (0600 hr.)
June
22
1975
I:5,500
1:10,000
1:5,500
1:10,000
1:9,000
2,500
5,000
2,500
Partial coverage (3 sites)
Predawn (0500 hr.); partial coverage (6 sites)
15-1
-------
&
¦ i^'
\
v*m
2%^. r-'
s. \T*'"K-*V
' v • *s
••• V'- r
\ Jta- x
^ofer f/ow -
li •
?•#' •
b',,'
*% '
• ** :ls
^,«a'
Ssfafe:
t& -
* ® Wetness
$ Discoloration
Wetness
$< VeQetotive stress
X Areas Interpreted as not wet on one sensor but wet on the other sensor.
Fig. 1. Overlay interpretation of leachate contamination.
a_ Reference photograph, b. Interpretation of two characteristics on color film,
c; interpretation of two characteristics on color infrared film. Note the use of
two sensors to refine wetness interpretation.
and Cortland, New York. Fifteen solid waste disposal
sites within the corridor were monitored during a ser-
ies of flights between September 197^ and June 1975.
Details of the missions, scales and sensors are
included in Table I. Other coverage of the sites-
including high altitude, multispectra 1 photographs and
panchromatic coverage dating back some 35 years--has
been collected for use in the study.
The disposal sites were selected to represent a
variety of solid waste materials, standards and meth-
ods of disposal as well as a variety of site condi-
tions and soil types. Most sites were used for muni-
cipal refuse only, however, several had received mix-
ed Industrial and municipal refuse. One site was used
exclusively for the disposal of industrial solid waste.
Seven sites were open and receiving refuse during the
time of this study; the remaining sites had been clos-
ed, some areas for as long as twenty years prior to
the remote sensing flights.
The entire study corridor had been glaciated, and
the variety of site conditions reflected this geologi-
cal history. Refuse disposal sites were located in a
variety of landforms including lacustrine clay plains,
outwash, ice-contact stratified drifts and till-cover-
ed bedrock highlands. In most cases the occurrence
and movement of leachate was very site specific and
depended on the details of the site geology.
A second research project not reported in this
paper was performed in cooperation with the remote
sensing. Supported by New York State's Department of
Environmental Conservation, this project Included
ground sampling and testing of leachate, ground and
surface waters from disposal sites and nearby areas.
The details of sampling varied from site to site with
some being monitored very extensively. For these
sites the time of ground sampling was scheduled to co-
incide rather closely with the aerial missions. Re-
sults of the field and laboratory measurements were
used to determine the extent of leachate contamination.
Interpretation Methods
The general procedure was to examine positive
film transparencies of the landfill sites with a zoom
stereoscope, analyzing each sensor's coverage Inde-
pendently, comparat ively and temporally. For the
comparative image analysis, acetate overlays depicting
interpreted wetness categories, anomalously colored
water, and stressed vegetation were made directly from
the color films. Features on these overlays were com-
pared to features imaged on the color Infrared films.
Overlays of the thermal imagery--daytime, nighttime
and predawn data—were compared to the photographic
overlays using a BL Zoom Transfer Scope.
For the color and color infrared photography,
acetate overlays were prepared directly from the nine-
inch film format. As illustrated in Figure 1, differ-
ent information was derived from the two types of data.
For some Image characteristics, such as the red-orange
colored stains in color film or vegetation stress in
the color infrared, all or most of the useful informa-
tion was on one sensor. For other characteristics the
two sensors provided complementary information. As an
example, Interpretation of wetness could be done for
either sensor, but in each case with some potential
for error due to misinterpretation of features which
appeared similar to wetness. By comparing overlays
prepared from the two types of film, Fig. lb 6 lc, the
areas of wetness were more precisely defined and mis-
interpretation could be corrected.
The thermal Imageries were analyzed separately
from the photography. Film densities indicating rela-
tively warm areas were recorded on overlays. The
thermal data were recorded on a five-Inch film rather
than the nine-inch used for the photography.
Evaluation of the Sensors and Methodology
In order to provide a basis for discussion of the
merits and limitations of various remote sensors
applied to the monitoring of landfill leachate, it Is
necessary to bring several other factors into perspec-
tive. Some of these were noted in the introduction,
but another Important factor is the variation In lea-
chate production with climate or season of the year.
Climate and season also affect the use of remote sen-
sors with some periods being distinctly better than
others. The conclusions of this study were developed
to include consideration of these factors.
The overall evaluation of remote sensing as a
method for monitoring leachate from landfills is sum-
marized In Table II. Information is tabulated in col-
15-1
*
-------
umns representing characteristic periods of the cli-
matic and seasonal variations in the study area.
These characteristic periods do not correspond di-
rectly to chronological periods although there is an
indication of when these periods usually occur in cen-
tral New York. The rationale for defining these char-
acteristic periods was to select units which maximized
the seasonal and climatic influence on leachate pro-
duction and use of remote sensors.
The four characteristic periods can be described
¦in more specific detail:
a. Wet. The major factor in this period is a-
bundant moisture both at and immediately prior to the
time of sensing. Abundant moisture leads to greatly
Increased production of leachate so that the poten-
tial numbers and size of target areas are maximized.
Wet areas on the ground as well as springs and drain-
age courses are numerous and as clearly defined as
they will be during the year. Moisture can come from
rainfall, snowmelt or both.
Vegetative cover varies but a dense vegetative
canopy is absent. Because vegetation cover can vary
from none to moderate, this characteristic period is
subdivided on the basis of vegetative cover. At the
extreme there may be some interference with the analy-
sis but the ground is generally observable.
b. Full Canopy. Dense vegetative canopy is the
overriding factor. The degree of wetness on the
ground, and the production of leachate, can vary; but,
since the visual analysis is restricted, the interpre-
tations are severely limited.
c. Dry - little or no canopy. Climate is dry
and both surface and groundwater flows are limited.
Leachate production is at a low level. There is lit-
tle or no vegetative canopy, however, low vegetative
cover interferes somewhat with the interpretation.
Temperatures are normally above freezing, but when the
ground is frozen water flow is restricted and this ad-
versely influences Interpretation.
d. Snow. The major factor in this period is
snow, but because of the variation in snow thickness,
the class ification period is further subdivided. Con-
tinuous light snow cover produces interference but
also accentuates some characteristics of leachate con-
tamination, Warm and saline leachates, as well as wet
areas, will melt a light snow cover or prevent accumu-
lation. Heavy snow covers everything and interference
with the interpretation is almost total. The degree
of frozen ground will influence the amount of leachate
and the interpretation.
The information contained in Table II will be
discussed in several sections. The usefulness and po-
tential of each sensor will be reviewed along with
discussion of scale, combinations of sensors and the
roost effective time and season of the year for moni-
toring leachate from landfill. The analysis of land-
fill imagery focused on areas characterized by: (1)
wetness, (2) anomalous tones--either In water, land or
vegetation--and (3) anomalously high temperatures.
Although these characteristics are not positive Indi-
cators of leachate per se, their observed or inferred
association with landfill causes areas displaying
these characteristics to be primary candidates for more
detailed study.
Photography
Color and color infrared photography were both
effective for monitoring some characteristics of lea-
chate (Table II). Moreover, to the extent that stereo-
scopic coverage was obtained, either could be used for
analyzing the landfill site physiography and topo-
graphy.
Although wet areas could be identified with
either the color or color infrared photographs, the
color infrared was generally superior. This is to be
expected since, in contrast to its surroundings, water
reflects little if any infrared radiation.
Discoloration of land and water contaminated by
leachate was best identified using color photography
since the discoloration was due principally to iron
oxides (i.e., reddish orange colors). In color photo-
graphy the red-orange water, in particular, was in
sharp contrast to the normal blue-green uncontaminated
water. Since water absorbs infrared radiation, the
difference between discolored and uncontaminated water
in the color infrared film was only a tonal difference
between blue-green and blue-black.
Color infrared photography has a unique capability
for identifying vegetative stress, and this was found
to be useful during some characteristic periods, par-
Tab
We
No
vegetat ive
canopy
le XX. Rating
Monitoring Un
t
Moderate
vegetat ive
canopy
s of Senso
der Differ
full
canopy
Usefulness
snt Ground Co
Dry,
little or
no canopy
for Leach
nd i t ions
Sn
Light
ate
ow
Heavy
Potential for
quant 1 tat ive
character izat ion
of leachate
Color
1 (2)
2
2
2
l»
Low
Color IR
1
1
3
2
3(2)
it
Low
Thermal*
2(3)
3
if
3
2
k
High
Useful
combinations
and rating
Color
Color IR
1
Color
Color IR
1
None
None
None
None
Legend
>vThe quality of thermal data I. Very useful; recommended
severely limited sensor 2. Useful; some limitations
evaluation 3. Generally not useful; some information
4. Not recommended
2(3)-dual symbol with rating between the two
numbers
15-1
-------
ticularly the wet springtime (Table II). Patches of
vegetation in wet areas, at or near the toe of land-
fill, were significantly lacking in infrared reflect-
ance. The vegetation, stressed by contaminated water
and/or excessive wetness, appeared brown-toned in con-
trast to the bright red of unaffected vegetation. Some
areas of stressed vegetation were identified where lea-
chate contamination could not be the explanation. As
with other indicators, this sensor characteristic can-
not be relied on as a unique identifier of leachate.
As indicated in Table II, the color and color in-
frared sensors had some applicability at almost any
characteristic period throughout the year. In combi-
nation, they are clearly the best overall remote sen-
sing system for identifying leachate from landfill.
While image enhancement for leachate detection would
likely prove worthwhile, the potential for quantita-
tively characterizing leachate with photography seems
low. The spectral variability of leachate is wide and,
In many cases, the leachate is too shallow.
Thermal Infrared Scanning
Thermal infrared is potentially a very attractive
sensor for monitoring leachate. The biological activ-
ity within a landfill produces relatively high temper-
atures and the fluids discharged from a landfill are
generally and significantly higher in temperature than
the surrounding waters, Figure 2.
Leachate Spring
Ponded Water
Ground Surface
9 10
21 Apr. 75
22 Apr. 75
Fig. 2 Variation in measured contact temperature of
four points on a landfill site:
During this research program a thermal scanner
was used on all daytime missions along with the photo-
graphic sensors. In addition special missions were
flown at various times during the night when the tem-
perature difference between leachate and other fea-
tures was expected to be a maximum. As illustrated in
Figure 2 this maximum temperature difference was im-
mediately predawn, however, during most night time
periods the temperature difference was insufficient to
provide any better information than either the color or
color infrared photography.
Predawn and early morning ground fog were so
common in the central New York study area that more
than half of the scheduled predawn missions could not
be flown. Consequently early nighttime flights were
used eventhough the temperature contrast was not max-
imum. Daytime flights provided no useful information.
At an altitude of 2500 feet above ground level
thermal scanner resolution (instantaneous field of
view) is not sufficient to differentiate many small wet
areas of 1-2 square meters.
Scanner calibration and background noise were a
problem on most flights. Use of a radiometer was un-
successful in this program but, in principle, this
technique should provide data which could be analysed
quantitatively as well as qualitatively.
The fourth major complication was introduced by
the variety of ground features besides leachate which
produced a warm signature. Most water bodies, springs
and seeps were warmer than adjacent soil areas during
the nighttime sensing periods. As a result, and with-
out more precise temperature calibration, these areas
could not be distinguished from leachate. In a similar
way snow patches appeared warmer under some conditions
during the winter months. Other interferences for the
thermal scanner included vegetation, shadows and micro-
cIimate.
In spite of these complications the thermal scan-
ner was partially successful as a leachate monitoring
tool. On the larger scale mid-winter flights in par-
ticular the sensor could be used to accurately locate
warm leachate springs and seeps. More research Is
necessary and on the basis of this study future efforts
should concentrate on 1) larger scale and higher re-
solution, 2) radiometer calibration and 3) time periods
of maximum temperature difference.
F1ight Parameters
Within the range of scales used in this study
(Table I) there were some indications of preferred
scales for photography applied to leachate monitoring.
A scale of 1:5,000 was generally satisfactory, combin-
ing the advantages of wide coverage with sufficient
detail for analysis of areas of Interest. At scales
much smaller than 1:5,000, specific details were lost
and 1:10,000 was generally too small. Larger scale
photography is useful for very detailed analysis, but
probably should be flown over very specific target
areas rather than for routine monitoring. In all
cases, stereoscopic coverage of the complete landfill
site is essential to the analysis.
The altitude of flights is noted in Table I. This
information may seem superfluous given scale, but It
should be mentioned that altitude proved to be an im-
portant factor in several missions because of weather
conditions. This was particularly critical during the
winter and spring seasons when the landfill targets
were optimum from the standpoint of leachate produc-
tion. On several occasions missions were flown at al-
titudes lower than planned because of clouds. When
such deviation in flight plan had not been anticipated,
the quality of data suffered. The significance of this
problem will vary in other climatic areas.
Preferred Periods for Remote Sensing of Leachate
It Is apparent from the summary In Table II that
certain times of the year, or characteristic periods,
are to be preferred in monitoring of leachate. Two
major factors Influence this variable: the amount of
leachate being produced and the ability of sensors to
monitor the leachate. In the humid climate of the
study area, there is a reasonably direct correspond-
15-1
-------
ence between infiltration and leachate production.
Consequently, from this standpoint the wetter periods
during the late fall and particularly In the spring
are to be preferred.
The sensors have various advantages and limita-
tions as noted above. In general, the periods of full
vegetative canopy and heavy snow cover are so limiting
that remote sensing of leachate cannot be done. Dry
periods with little or no vegetative canopy are satis-
factory for the use of most sensors, however, much
less leachate can be expected during these periods.
Overall, the best time for remote monitoring of
leachate Is during a wet period when there is little
or no vegetative canopy. In the study area that per-
iod occurred most consistently In the spring and at
other times during the late fall and winter.
Some other periods, usually much shorter and less
consistent than the above, were also found to be sat-
isfactory for leachate monitoring (Table II). For ex-
ample, the condition of light snow cover, particularly
in combination with wet ground, was useful since lea-
chate contamination could be associated with (a) melt-
ed snow areas, (b) color-stained snow, (c) wet areas,
and (d) warm spots monitored using the thermal sensor.
Conclusions
Remote sensing can be used effectively to monitor
the contamination of ground and surface water by lea-
chate from landfill. Color, color Infrared and ther-
mal sensors were all demonstrated to be useful to
some degree, but the combination of color and color
infrared photography was the most effective system.
Wetness, discoloration of land and surface waters,
and vegetative stress were useful ground characteris-
tics Imaged by the remote sensors, in no case were
they uniquely indicative of leachate but, in associa-
tion with landfill, they defined prime areas for more
detailed study and ground level confirmation. No
methods were found to quantitatively define leachate.
Remote sensing was most effective during wet
periods in early spring.
Acknowledgements
Support for this project was provided in part by
the New York State Department of Environmental Conser-
vation (Contract C-79273) and the NASA - sponsored
Remote Sensing Program at Cornell (NASA Grant NGL-33-
010-171). The five remote sensing missions and re-
sulting data were provided by the U.S. Environmental
Protection Agency.
References
(1) Garland, G.A. and Mosher, D.C. (197*0 "Leachate
Effects from Improper Land Disposal" Waste Age
5(H) (Nov.)
(2) Souto-Maior, J. (1973) "Applications of Thermal
Remote Sensing to Detailed Ground Water Studies"
Remote Sensing and Water Resources Management
Am. Water Res. Asso. Urbana, 111.
5
15-1
-------
LAND DISPOSAL OF WASTES: POTENTIAL FOR GROUNDWATER POLLUTION
Richard A. Carries
Dirk R. Brunner
Robert E. Landreth
Mike H. Roulier
Disposal Branch
Solid and Hazardous Waste Research Division
Municipal Environmental Research Laboratory
Cincinnati, Ohio
Abstract
Land disposal of municipal and industrial wastes
has the potential of polluting surface and groundwater
supplies. Studies have identified certain geographic
regions as being more susceptible than others to
groundwater contamination. Case histories have been
studied identifying incidences of pollutant migrating
into surface or groundwater. The shortage of well-
documented cases of water pollution resulting from
poor waste disposal practices is primarily due to the
lack of adequate surface and groundwater monitoring
systems. Research is ongoing to Identify migration
potentials in various soils under different leaching
media. Control technology studies are evaluating
fixation processes, landfill liners, encapsulation,
and leachate collection and treatment as means to
prevent surface and groundwater contamination.
Recently passed legislation indicates that Congress
is concerned over disposal practices and how they
affect surface waters and groundwaters.
Summary
Although methods have been developed for con-
trolling or eliminating potential environmental Insult
caused from disposing of vast amounts of solid and
hazardous wastes on land, we must continually seek to
improve the design of and operating conditions at land
disposal sites. The volume of leachate 1s highly de-
pendent on site hydrogeology as Influenced by design,
operation, and local climate. Following a detailed
hydrogeologic and engineering design study, correctly
operated disposal sites will have much less potential
of polluting the groundwater than open dumps. By the
very nature of the material hazardous waste, leachates
must always be under positive control. The pollution
of groundwater and surface water can be found in the
literature, but considering the many waste disposal
sites that exist, the number of incidents 1s rela-
tively few. However, this should not give rise to
complacency but serve to drive us on to even better
disposal practices so that future generations will
have a groundwater supply unpolluted from man's
activities on the earth's surface. Monitoring capa-
bilities must be advanced beyond those presently in
force to assure that the groundwater remains an asset.
Introduction
Municipal solid waste generation 1n the U.S. dur-
ing 1971 was estimated to be 3.3 lb/cap/day (1.5 kg/
cap/day), or an annual production of 113 million
metric tons.* About 90 percent of the total was dis-
posed of on the land, the remainder was Incinerated,
composted, reused, or dumped into the ocean.
It Is estimated that some 10 million tons of non-
radioactive hazardous wastes (Industrial) were gener-
ated In a year and that the total 1s Increasing by 5
to 10 percent annually.^ Most of these wastes are
also disposed of on the land, sometimes along with
municipal solid wastes.
A recent survey estimated that some 18,539 land
3
disposal sites were in operation nationwide, but it
would be fallacious to believe that all of them even
vaguely met the general definition of a sanitary land-
fill. Given the fact that land disposal activities
are so prevalent and that about 50 percent of the
Nation's population draws groundwater for domestic use,
it seems only appropriate that sound engineering prin-
ciples should be applied to all stages of land dis-
posal .
Regional Considerations
Several studies have investigated groundwater pol-
lution in various areas of the country and reached the
same general conclusion, that 1s, if a land disposal
activity is placed over or near a groundwater table,
there is a real threat that the latter will become
polluted. A report published by the EPA^ discusses
groundwater problems in Arkansas, Louisiana, New
Mexico, Oklahoma, and Texas, and found very few docu-
mented instances 1n which leachate from sanitary land-
fills had seriously contaminated groundwater. The
report goes on to say, however, that the major solid
waste threat to groundwater quality will probably be
from the land disposal of industrial wastes.
Another report5 indicates few problems are ex-
pected 1n Arizona, Nevada, and parts of California due
to the deep water table and lack of rainfall. Provo,
Salt Lake City, and Logan, Utah were identified as
having landfills located 1n lowlands where the ground-
water table is fairly high and is actually near the
refuse disposal area at certain times of the year.
The EPA's report entitled "Groundwater Contamina-
tion in the Northeast States" discusses 60 cases of
damage to groundwater caused by dumps.® The informa-
tion, which was obtained from interviews, public
agency files, published sources, and unpublished re-
ports, does not represent an exhaustive survey of all
sites in the area.
The states of Colorado, Idaho, Montana, Oregon,
Washington, and Wyoming were also surveyed and, In
general, groundwater pollution from landfills was
7
found to be minimal. Several instances were found 1n
which improper disposal had polluted groundwater, but
these were generally In areas that receive more rain-
fall than elsewhere In the study area. In the arid
portion of the Northwest, landfills were generally not
considered to be a problem because little or no
leachate is produced.
Q
Weddle and Garland published an assessment of
dumps as a threat to groundwater supplies, and Garland
15-2
-------
q
and Mosher published an article relating leachate to
improper land disposal practices. They provided a map
showing the magnitude of potential infiltration and
leachate production in the U.S. Surface runoff was
not included since its effects vary with local condi-
tions. Thus, a reduction in the quantity of leachate
formed may be affected by encouragement of runoff.
In essence, they indicated that those areas receiving
the greatest amount of rainfall and having the lowest
evapotranspiration rates have the greatest potential
for groundwater pollution from waste disposal sites.
EPA Research Activities
The Solid and Hazardous Waste Research Division
of the USEPA is responsible for investigating new and
improved solid and hazardous waste management systems.
To this end, studies are underway on settling rates,
decomposition of material, leachate formation, migra-
tion and treatability of leachates, liner evaluations,
and landfill gas production.
In addition, we are actively involved in a series
of projects directed toward achieving sound management
of hazardous waste. Specific projects are: (1)
determining how and to what extent these materials
might migrate through different soils; (2) chemically
stabilizing and encapsulating wastes to prevent haz-
ardous constituent leaching; (3) evaluating various
liner materials to see if they can be used to prevent
hazardous wastes from migrating into groundwater at
disposal sites.
Legislative Action
On December 16, 1974 the Congress passed Public
Law 93-523 entitled "Safe Drinking Water Act." By
this action the Congress showed it wanted not only to
preserve surface water quality, but also to assure
that our natural groundwater supplies are not contami-
nated by land disposal activities.
Section 1421--Protection of Underground Sources
of Drinking Water - regulates the activities of State
underground injection programs. The many sections and
subsections are too numerous to discuss here, but the
important implication is that groundwater is recog-
nized as a drinking water source that must not be con-
taminated.
Section 1442--General Provisions - calls for the
Administrator of the EPA to conduct research into im-
proved methods of protecting underground water sources
(groundwater supplies) from contamination. The sec-
tion also requires him to conduct a survey and study
of: (1) waste disposal (including residential) which
may endanger underground sources which supply, or can
reasonably be expected to supply, any public water
systems; (2) ways to control such waste disposal.
The Act calls for research mandates and require-
ments, but the message is that Congress has recognized
that land disposal of even residential waste can pre-
sent a danger to groundwater supplies.10
Leachate from land disposal activities, which
might contaminate aquifers, is a growing problem in
many areas, and many states now require that the
groundwater quality in the vicinity of land disposal
sites be monitored. Since July 1974, for example,
New Jersey has required that monitoring wells be
drilled at all new landfill sites, or any of the
existing 307 landfill locations within the state if
there is any indication that their operation may pose
an actual or potential threat to the groundwater. The
regulations call for three wells per site - two down-
stream from the landfill and one upstream. It'is an-
ticipated that the test wells will enable state
personnel to check water quality to determine if land-
fill contaminants have permeated the water table.
Delaware's solid waste disposal regulations require
that systems for collecting, treating, and disposing
of leachate be installed at all landfills. Also, the
solid waste disposal area must contain an impermeable
liner of natural or synthetic material to keep leach-
ate out of groundwater supplies. Treated leachate
must meet the effluent standards of the state's regu-
lations governing the control of water pollution.
During its operation, and for at least five years after
the landfill closes, the operator is required to moni-
tor groundwater quality in its vicinity. He must also
maintain the leachate collection and treatment system
during this period. The new regulations provide
standards for both sanitary landfills and industrial
landfills.11
Monitoring
Legislation requiring disposal site groundwater
monitoring is a step in the right direction, however,
there still are plenty of problem areas to be identi-
12
fied and corrected. Walker suggests that site moni-
toring involves much more than merely sinking a few
observation wells. Little, if any, attention is given
to possible adverse effects of banked pollutants re-
entering the environment from soil or plant storage
areas. Frequency of water sampling is not standar-
dized, normally sampling is monthly or bimonthly.
Samples are seldom taken during and/or immediately
after major periods of precipitation when the major
quantity of toxic chemical removal by these routes
would be expected to occur.
The monitoring system themselves need to be stud-
ied and updated. For instance, in fine-grained, low-
permeability earth (material generally considered to
be most desirable for land disposal sites) a repre-
sentative groundwater sample may not be obtainable from
a properly installed monitoring well for several weeks
or even months after it is emplaced because such mate-
rial has low water-yielding characteristics.
Recent studies indicate that observation well
monitoring systems may not be the most effective means
to trace chemical pollutant flow paths or to determine
groundwater chemical concentrations at any time or
depth. Instead, these studies show that chemical
analysis of core samples from the underlying earth
material profile permits a positive definition of any
chemical constituent within the profile at any given
location.
Walker goes on to identify six possible avenues
for pollutants to reenter the environment from toxic
waste land disposal sites. He suggests monitoring
techniques for each avenue directed at maximizing the
usefulness of the data collected.
What is difficult to establish, since definitive
data are still very much lacking, is how long after a
site is closed should monitoring continue? In some
cases it was several years after a site was closed be-
fore any groundwater contamination was noted; estab-
lishing an arbitrary monitoring period may, therefore,
be a very risky business.
2
15-2
-------
References
1. Second Report to Congress: Resource Recovery and
Source Reduction. USEPA 1974 publication (SW-122)
prepared by Office of Solid Waste Management
Programs.
2. Swift, W. H. Feasibility study for development
of a system of hazardous waste national disposal
sites. Vol. 1 [Richland, Washington]. Sattelle
Memorial Institute, Mar. 1, 1973.
3. An Overview of the Land Disposal Problem. Waste
Age, Vol. 6, No. I, Jan. 1975, pp. 17-24.
4. Scalf, M. R., J. W. Keeley and C. J. LaFevers.
Groundwater Pollution in the South Central States.
EPA-R2-73-268, NERC-Corvallis, Oregon 97330, June
1973.
5. Fuhrinian, Barton and Associates, Provo, Utah
8460J, Groundwater Pollution in Arizona, Califor-
nia, Nevada, and Utah. U.S. EPA Water Pollution
Control Research Series 16060 ERU 12/71,
6. Scalf, M. R. Groundwater Contamination in the
Northeast States. EPA-660/2-74-056, U.S. EPA,
Washington, D.C. 20460. June 1974.
7. van der Leeden, F. et al. Groundwater Pollution
Problems in the Northwestern United States.
EPA-660/3-75-018, May 1975. U.S. EPA, Washington,
D.C. 20460.
8. Weddle, 8. and G. Garland. Dumps: A Potential
Threat to Our Groundwater Supplies. Nation's
Cities, Oct. 1974,
9. Garland, G. A. and D. C. Mosher. Leachate Effects
of Improper Land Disposal. Waste Age, 5 (11) Mar.
1975. pp. 43-48.
10. Safety of Public Water Systems - Public Law
93-523, 93rd Congress, S. 433, Dec. 16, 1974.
11. Land Pollution Reporter. Freed Publishing Co.,
P.O. Box 1144, FOR Station, New York, New York.
Mar/Apr. 1975.
12. Walker, W. H. Monitoring Toxic Chemicals in Land
Disposal Sites. Pollution Engineering. Sept.
1974, pp. 50-53.
3
15-2
-------
A Proposed National Monitoring System
for Organics in Water
John M. McGuire
Analytical Chemistry Branch
Environmental Protection Agency
Environmental Research Laboratory
Athens, Georgia 30601
Summary
Computerized gas chromatography/mass
spectrometry can identify specific organic
compounds isolated from natural water and
manufacturing effluents. Computerized
spectra matching programs extend the system
effectiveness by providing rapid
identification of compounds whose spectra
are in a 35,000+ collection of reference
spectra. Application of this system to
monitoring has revealed a significant number
of previously unidentified pollutants.
Addition of record-keeping facilities to
a central computer can provide an ideal tool
for State, EPA, and industrial monitoring
facilities. The geographic frequency
distribution of identified compounds (and
even unidentified compounds that occur
repetitively) can be tabulated
automatically.
************
Tabic I
Com ' iri sons of Compounds Reported t)y Discharger and
compounds Identified by ERA in an Industrial Discharge
Products or
Raw Materials
Reported
Compounds
Identified
Propylene
all xylenes*
1,5-cyclooct ad ler.o
Ethylene
i sopropy 1 benzene
styreno*
Butadiene
o-cthy 1 toluene
0 - me t h y 1 a t.y re n rj *
Butane
diacetone alcohol
indan*
Octane
2-butoxyethanol
,!~mo.thylstyrene
Ethylene glycol
i ndono*
dime thylfuran isomer
Ethylene oxide
n-oentadocane
1-me thyl indeno *
Polyglycols
3-mothylindono
acetophenonc
Ammonia
n-hexadecane
M-terpineol
Raw gas
naphthalene*
both methyl
Ethane
benzyl alcohol
naphthalene isomers*
Refinery gases
ethylnaphthalone
rt-methylbonzyl alcohol
Refinery C2
2,6-dimethy1-
phenol*
Stream
naphthalone*
me thylcthylnaphthalene
Refinery C3
cresol isomer
acenaphthcne
Stream
acenaphthalene
methylbiphenyl isomer
Propane
fluorene
two pluhalate diestcrs
Hydroformer gas
3,3-dipheny1-
Platformer gas
propanol
•Identification was confirmed with a standard.
Concentrations of less than 1 mg/1 for
organic compounds in water can be far in
excess of environmentally safe levels or
can be totally insignificant, depending upon
the toxicity of the specific compound. For
example a concentration of 10 Vg/1 of
palmitic acid is relatively harmless in
water; however quinoline at the same
concentration is toxic to fish. This
concentration, equivalent to a total organic
carbon (TOC) concentration of 8.4 jjg/1, is
two orders of magnitude below the detection
limit of the tentative standard method for
TOC. For such toxic materials, specific
compound identifications are necessary for
establishing water quality criteria.
EPA laboratories have shown that
monitoring of industrial effluents for
specific compounds is essential for the
enforcement of water quality standards. The
first column of Table 1 lists pollutants
suggested by the representative of a
refinery company as possible waste products
from the plant's operation. The second
column is a partial list of compounds
actually found in the effluent by EPA. The
two columns bear little resemblance to each
other. In the case of a pesticide plant,
the effluent was found to contain 9 highly-
chlorinated by-products of unreported
toxicities as well as 5 expected chlorinated
pesticides. Most of the unexpected
materials were present at substantially
higher concentrations than were the
pesticides. The identification of specific
compounds in cases such as these aids in
assessing the environmental impact of an
effluent, and in setting and enforcing water
quality standards.
over the past 7 years the EPA and its
predecessors have developed and evaluated
techniques for the identification of
specific pollutants. Gas chromatography has
been used to separate complex mixtures of
moderately volatile compounds. However,
although it is a sensitive method for
separating many mixtures, GC does not
provide reliable identifications. To extend
the utility of GC, detectors have been
investigated that yield spectra or finger-
prints of the organic compounds eluting from
the gas chromatograph. The detection method
of greatest utility at this time is mass
spectrometry. A gas chromatograph/mass
spectrometer (GC/MS) system can resolve
complex mixtures of water pollutants and
produce a mass spectrum of each peak as it
elutes from the chromatograph. The
Environmental Research Laboratory in Athens,
Georgia, demonstrated this technique with a
manually operated GC/MS in 1968.
Modern instruments can be interfaced to
dedicated computers to provide computer
control of data acquisition as well as
computerized data reduction. A computer-
controlled system is not only much faster
than a manual system, but also is free of
errors caused by operator carelessness.
To evaluate the feasibility of
interfacing the GC/MS system to a computer
for EPA laboratories, the Athens Laboratory
and the Environmental Monitoring and Support
Laboratory in Cincinnati (EMSL) purchased
identical computerized GC/MS systems in
1971. Since that time, more than 20 EPA
1
15-4
-------
laboratories have chosen essentially the
same system. A PDP-8 minicomputer in this
system controls the operation of a
quadrupole mas3 spectrometer and associated
output devices. With the computerized
system, the applied time required to gather
data and prepare it for interpretation was
decreased by more than an order of magnitude
compared to manual GC/MS systems.
A computerized system can also be
employed to shorten the time required for
spectral interpretation and compound
identification. Several methods for doing
this have been suggested. In 1971 an EPA
research grant was made to Battelle to
develop one of these methods1 into a
computerized semi-automatic spectra matching
program and to develop a reference library
of mass spectra. The successful program was
made available to EPA laboratories on a
trial basis in 1972 and an improved version
was made available in 1973.
In the final version2, the two most
intense peaks in every 14 mass unit grouping
of an unknown spectrum were automatically
transmitted from a remote minicomputer to a
central large computer. The large computer
selected the best fingerprint matches to the
input data from a file of 9,000 reference
spectra. The matches were ranked on the
basis of their similarities to the input
spectrum, transmitted to the remote
minicomputer, and listed on an output device
in the user's laboratory. The system was
rapid and efficient. A typical search is
shown in Figure 1.
figure 1
Computerised Spectra Matching Program Dialog for
a Component of an industrial Plant Extract
S OR P? S
OPTION TTttT-A
PHILADELPHIA CCE#1, RUN 2' 4-16
47,161
45,138
48,429
50,140
65,69
73,64
83,999
85 ,648
93,28
111,202
113,175
141,31
0,0
EDIT? N
FILE KEY = 8569
0 1 AA-1344-2
SI = .678
FILE KEY » 4388
0 1 AA-813-2
SI = .378
FILE KEY * 21379
0 1 AA-813-3
SI - .325
FILE KEY » 21867
0 1 AA-1026-2
SI - .148
FILE KEY - 37854
0 1 EP-0-U380
SI - .117
1,1,3,3-TETRACHLOHQ-2-PROPANONE
DICHLOROACETYL CHLORIDE
DICHLOROACETYL CHLORIDE
BROMODICHLOROMETHANE
IODODICHLOROMETHANE
All user inputs are underlined. The fourth
through the eighteenth lines of the dialog,
generated and transmitted by the
minicomputer, comprise the unknown input
spectrum and a terminator. The following
line shows that the user accepts the input.
The five most probable identifications are
then given along with references to the
complete spectrum and a measure of the
similarities of the unknown to the reference
spectra.
In parallel with the development of this
system, workers at the National Institutes
of Health developed an interactive program
to screen essentially the same reference
file of mass spectra for an operator-
selected pair of mass and intensity
values. Successive screenings resulted in
narrowing the number of spectra passing the
process. An evaluation of this program by
the EMSL11 for typical water pollutants
showed that an average of 5 pairs of mass
and intensity data were required to reduce
the number of screened spectra to less than
10. A typical dialog is shown in Figure 2.
Figure 2
NIH PEAK Search for a Petrochemical Compound
TYPE PEAK, INT
CR TO EXIT, 1 FOR ID,MW,MF, AND NAME
USER 105,100
# REFS M/E PEAKS
893 105
NEXT REQUEST 120,38
# REFS M/E PEAKS
96 105 120
N2XT REQUEST 91,11
# REFS M/E PEAKS
25 105 120 n
NEXT REQUEST 79,16
# REFS M/E PEAKS
7 105 120 91 79
NEXT REQUEST 1
ID#
2135
4867
16799
20472
20473
-20474
20475
W
MF
120 C9 H12
125 C9 H12 02
388 C12 H12 03 W
120 C9 1112
120 C9 H12
120 C9 Hi2
120 C9 H12
NAME
ISOPROPYLBENZENE
CUMENE HYDROPEROXIDE
PI-1,3,5-TRIMEtHYLBENZENE-
TRICARBONYL TUNGSTEN
ISOPROPYLBENZENE
ISOPROPYLBENZENE
1-METHYL-2-ETHYLBEN2ENE
O-ETHYLTOLUBNE
Application of both of these search systems
by EPA resulted by 1973 in identification in
industrial effluents of more than 200
compounds that had not been identified
previously.
In both search systems the number of
spectra in the reference library obviously
restricts the number of identifications that
can be made by spectra matching. The
Management Information and Data Systems
Division of EPA (MIDSD) therefore purchased
use rights for the largest single collection
of mass spectra s and contracted with one of
the mass spectrometrists who had compiled
2
-------
that collection, Prof. F. W. McLafferty of
Cornell, to expand and improve it from its
initial size of 25,000 spectra.
Dr. S. R. Heller of MIDSD was
instrumental in arranging for the
EPA/Battelle and NIH mass spectral search
programs to be combined into a single system
located on a commercial computer. This Mass
Spectral Search System6, which uses the
current enlarged EPA data base of 35,000
spectra, is now available to users in state
and local monitoring laboratories, to
industries, and to academic institutions, as
well as to 24 EPA laboratories, and to other
Federal agencies. It is a joint effort of
NIH, FDA, EPA, and the Mass Spectral Data
Centre of the U.K.
During the past year, Battelle
researchers developed record-keeping
modifications to the original EPA/Battelle
search programs. The modifications are
designed to provide statistics on how
frequently a given compound is identified,
where particular compounds are identified,
and how often the same unidentified spectrum
is encountered. These modifications, when
incorporated into the Mass Spectral Search
System, will provide an ideal nucleus for a
national organics monitoring system. It can
accept inputs from and provide outputs to
local, state, regional, and Federal
enforcement officials. This system can give
them current tabulations of which organic
compounds are most widely distributed, which
ones are associated with particular bodies
of water, and which unidentified spectra are
common enough to justify special
identification methods.
In the actual implementation of the
monitoring system, provision can be made for
those industrial users who may not want to
put their sample source data into the
central file, but only want a spectrum
identification. However, for the
representative user concerned with
monitoring, a typical dialog is given in
Figure 3.
Figure 3
Dialog Showing User
Input of Sample Source
Identi fication
LAB. NO.? 5j0
YOUR NAME/ MCGUIRE
MAIN FILE? Y
S, E, P, R, OR C? S
I.D.? 2X45C1 ~
WATER POLLUTION SAMPLE? Y
STATE? LOUISIANA ~
NEAREST CITY? NEW ORLEANS
RIVER? MISSISSIPPI
DATE? 3/1/75
TIME OF DAY? 1410
PAPER TAPE? N
A model output of the geographic
distribution program is shown in Figure 4.
Figure 4
Portion of an Hypothetical
Distribution Retrieval
FILE KEY? 49999
USER/LABORATORY-JONES/NYS
STATE/CITY-NEW YORK/WATERFORD
RIVER/RIVER MILE-HUDS0N/1(3T1
DATE/TIME-3/15/76//1255
SIMILARITY INDEX = .712.
USER/LABORATORY-SMITH/REGION II
STATE/CITY-NEW YORK/NEW YORK
RIVER/RIVER MILE-HUDSON/181
DATE/TIME-3/16/76//0842
DIBENZYL HYPOTHECATE 368 C26.H24.02
SIMILARITY INDEX = .572
USER/LABORATORY-GOLD/NEIC
STATE/CITY-UTAH/SALT LAKE CITY
RIVER/RIVER MILE-RES. #2/0
DATE/TIME-4/2/76//1030
DIBENZYL HYPOTHECATE 368 C26.H24.02
SIMILARITY INDEX ¦» .420
This output could be used to show the need
for obtaining toxicological information on
"dibenzyl hypothecate."
For the analyst in the monitoring
laboratory, the system might note that his
input spectrum is unidentified, but is
similar to another laboratory's unmatched
spectrum by giving, for example, an
identification of "unmatched spectrum #1234"
and the source data for that spectrum. For
the monitoring official, the corresponding
listing might show that spectrum #1234 was
found in 3 locations on the Mississippi
river and in the drinking water supply of
New Orleans within a particular period of
time. This information may then be used to
locate the source of the pollutant, to chart
its movement in a water system, or to
trigger a high priority effort to identify
it.
REFERENCES
1. Hertz, H. S., Hites, R. A., and Biemann,
K., Identification of Mass Spectra by
Computer-Searching a File of Known
Spectra, Anal. Chem., 43, 681 (1971).
2. Hoyland, J. R. and Neher, M. B.,
Implementation of a Computer-Based
Information System for Mass Spectral
Identification, EPA Research Report EPA-
660/2-74-048 (1973T.
3. Heller, S. R., Conversational Mass
Spectral Retrieval System and Its Use as
an Aid in Structure Determination, Anal.
Chem., 44, 1951 (1972).
4. Budde, W. L. and Eichelberger, J. W.,
personal communication. July 1974.
5. Stenhagen, E., Abrahamsson, S., and
McLafferty, F. W., Registry of Mass
Spectral Data, Wiley-Interscience, New
York, 1974.
3
15-4
-------
6. Heller, S. R., McGuire, J, M., and
Budde, W. L. / Environ. Sci. Technol., 9^
210 (1975).
15-4
-------
CAPABILITIES AND LIMITATIONS IN IDENTIFYING AND
MEASURING AQUATIC POLLUTANTS
William T. Donaldson
Environmental Protection Agency
Environmental Research Laboratory
Athens, GA 30601
Summary
An increased concern about the health
effects of pollutants has called attention to
a broad range of organic compounds and
chemical elements. Qualitative organic
analysis and multielement analysis therefore
becomes essential in assessing the
significance of pollution sources and in
determining acceptability of water for human
consumption.
Gas chromatography-mass spectrometry (GC-
MS), with computerized data interpretation,
has opened the door to the identification of
extractable, volatile organic compounds; but
identification of polar and high-molecular-
weight compounds requires additional
research. Spark source mass spectrometry
will identify and measure simultaneously all
chemical elements except carbon, hydrogen,
oxygen, nitrogen, and the six rare gases, but
it is neither fast nor inexpensive.
A program at the Athens Environmental
Research Laboratory is reducing the cost of
GC-MS analysis, developing techniques for
separating and concentrating polar organic
compounds, and evaluating multielement
methods that are faster and less expensive.
Introduction
In the past decade the approach to the
analytical chemistry of aquatic pollutants
has changed significantly—largely for two
reasons: (1) Applied analytical chemists
have developed the capability to identify and
measure hundreds of chemicals at lower
concentrations than was previously possible.
(2) The public has become concerned about
chronic human health effects and effects on
the aquatic environment caused by trace
levels of chemicals in water. This change
has brought into the field of environmental
analytical chemistry many specialists, who
have applied their skills in augmenting those
of "water chemists", who applied mostly wet
chemical techniques to a broad spectrum of
problems. Subsequently a gap in
understanding has arisen between the
analytical chemists and specialists in other
environmental fields. Because the techniques
used by the chemist are specialized, their
capabilities and limitations in identifying
aquatic pollutants are not fully understood
by those whom they could help most.
The two most rapidly changing areas of
environmental analytical chemistry are the
identification and measurement of specific
organic pollutants and simultaneous
multielement analysis. A discussion of some
relevant observations about these two areas
may help to provide a better perspective of
current analytical capabilities for
environmental problems.
The Number of Chemicals in Water
As our ability to separate organic
chemicals was improved and detection limits
were lowered we observed an interesting
phenomenon at the Athens Environmental
Research Laboratory: when the detection level
is lowered an order of magnitude, the number
of organic chemicals observed increases an
order of magnitude. For example, using one
technique we observed 3 organic compounds in
New Orleans drinking water at estimated
concentrations between 1 and 10 lag per liter.
Between 0.1 and 1 pg per liter 17 were found
and between 0.01 and 0.1 pg per liter, 170
more were found.1 We were unable to separate
or to detect compounds at lower
concentrations; but if one extrapolates this
relationship to a concentration of 10~6 ug/H>,
the corresponding number of compounds is 10®.
In other terms, one may expect to find in a
liter of water at least 1010 molecules of
each of the 106 organic compounds that have
been reported in the literature.
Considering solubilities and other
physical properties of inorganic materials
one might expect a similar relationship for
the chemical elements. We have observed
about half of the chemical elements at
concentrations above 1 Ug/liter in most
surface waters examined.2 By no means are the
same 40 or 50 elements always present.
Capabilities
Analytical chemists and other
environmentalists play leapfrog. The
chemists learn more about the nature of the
environment, other scientists use this
information and discover a need for yet more
information, and the proverbial monkey on the
back goes back and forth. In the case of
some organic compounds the analytical
chemists have currently placed the monkey
squarely on the backs of the toxicologists.
Organic compounds that can be sorbed from
water onto activated charcoal or
macroreticular resins, back-extracted and
concentrated in organic solvents, and passed
through a gas chromatograph can be identified
and measured at concentrations down to 0.01
yg/liter. The toxicologists, however, have
not yet proven that any compound has adverse
health effects in drinking water below a
concentration of 0.01 yg/liter. The chemists
are now providing them with long lists of
compounds found in drinking water at
concentrations above 0.01 yg/1, a significant
challenge.
1
15-5
-------
Similar lists of chemical elements can be
provided for various samples of drinking
water. Chemists can now determine
simultaneously the concentration of all
chemical elements (except carbon, hydrogen,
oxygen, nitrogen, and the six rare gases)
present in water at concentrations above 1
yg/liter. 3 Although toxicologists have
information on a relatively high percentage
of the chemical elements, the ecological
significance of some elements now being
observed by simultaneous qualitative and
quantitative multielement techniques is not
established.
Limitations and Suggested Approaches
to Reduce Them
Organic Analysis
There are several limitations in the
identification and measurement of organic
compounds in water. In using gas
chromatography-mass spectrometry (GC-MS) for
analysis, we are essentially looking through
only one "window",4 and that window exposes
only about 10 percent.of the total mass of
carbonaceous material in some waters.
The first limitation lies in the sampling
method—sorption on carbon or resins. Only
the less polar compounds sorb and their
sorption efficiencies vary from compound to
compound, both on carbon and on resins.6 a
thorough investigation must be made to
develop optimum procedures for sampling and
for extraction of the sorbed materials from
sorption columns.
Of those compounds that are sorbed and
extracted into solvents, not all will pass
through a gas ehromatograph without changing
in composition. Methods to separate these
compounds by liquid chromatography or by
derivitization to stable compounds and
subsequent gas chromatography should be
established.
Compounds that can be gas chromatographed
are usually identified by obtaining their
electron impact, low-resolution mass spectra
and comparing the spectra empirically with
those of previously identified compounds.
Several limitations are encountered in this
process. Because of the thousands of
compounds to be considered, computer-assisted
spectra matching programs are used.7 Not all
compounds likely to be encountered in water
are represented in the computer files. As
the files are increased, the cost of computer
searches increases, and the time for a search
becomes longer. The spectra for some
compounds are not significantly different
from those of others, so neither the computer
nor the analyst can identify the compounds
from their electron impact spectra alone.
All identifications must be considered
tentative until confirmed by comparing at
least one additional parameter.
To compile needed computer files of
spectra representing a high percentage of
compounds that are in environmental waters,
chemists must know more about which compounds
are present in water. More efficient search
programs must be developed by improving
algorithms and by separating compounds into
high-probability subfiles that can be
selected by the analyst on the basis of
sample history. Capabilities should be
developed for generating other spectra
(infrared, chemical ionization mass spectra,
high resolution mass spectra, etc.) so that
compounds not identified from low-resolution
electron impact mass spectra can be
identified or so that identifications can be
confirmed. Marker compounds for gas
chromatography are needed to provide precige
relative retention times and to provide
reference intensities for quantifying
identified compounds.
Polar compounds are not removed from
water by sorption columns or by solvent
extraction. Techniques for concentrating,
separating, and generating characteristic
spectra for polar compounds should be
developed. Freeze concentration, low-
temperature distillation, and high-pressure
ion exchange liquid chromatography® are some
of the techniques that show promise for
application to polar compounds.
High molecular weight compounds, even
after separation, are not amenable to
conventional mass spectrometry. Atmospheric
pressure ionization mass spectrometry.
Fourier transform infrared spectroscopy,9
nuclear magnetic resonance spectrometry, and
Raman spectroscopy offer some hope for
identifying these compounds.
Finally, a study should be made to gain
insight into the nature of the 80 to 90
percent of organic matter in water that we do
not observe with our GC-MS window. In such a
study as this compounds should first be
separated into classes based on chemical or
physical properties, and some evaluation of
the ecological significance of each class
should be made. This approach will permit
setting priorities in the tedious research of
characterizing each class.
Multielement Analysis
Current multielement analysis
capabilities are hampered by limitations that
should be removed. Spark source mass
spectrometry is the only instrumental
technique capable of measuring simultaneously
up to 82 elements at concentrations as low as
1 ng/1. However, it is expensive (250
thousand dollars), slow (4 samples per 8
hours), and imprecise (±30-50% relative
standard deviation in multielement
determinations). Also, a highly skilled
analyst is required to operate the
instrument.
Two multielement techniques, X-ray
fluorescence spectrometry1" and inductively
coupled plasma emission apectrography, 1 1 show
promise of overcoming most of these
shortcomings. However, neither possesses the
uniform detection levels for all elements
that spark source mass spectrometry exhibits
and neither covers as broad a spectrum of
elements. A third technique, argon plasma
source mass spectrometry, may ultimately be
15-5
-------
the best of all techniques currently under
consideration; but it is still in the
relatively early developmental stages. 12
Energy dispersive X-ray fluorescence can
be accomplished with a relatively inexpensive
instrument equipped with a computerized
detector and data reduction system. Because
this technique affords no capability to vary
the excitation time, some elements can not be
determined when present at low concentrations
relative to other elements. The wavelength
dispersive x-ray technique overcomes this
problem, but instrumentation is more complex
and costly. Also, since water samples can
not be analyzed with adequate sensitivity,
they must be concentrated by ion exchange.
Solids analysis is difficult since the
relatively shallow penetration of photons
excites only those elements near the
surfaces. However, if matrix and sample
preparation problems can be overcome, X-ray
fluorescence can be successfully applied
directly to soils and sediments.
Inductively coupled plasma emission
spectrometry can be applied directly to
water, without preconcentration, with
sensitivities of 1 Vg/liter for most elements
of interest, although only about 20 elements
can be determined simultaneously. Although
this technique is usually applicable to
drinking water, workers at our laboratory
have discovered adverse matrix effects when
applying it to wastes from metals plating
industries. The cost and complexity of the
instrumentation is about the same as that of
the energy dispersive X-ray fluorescence
spectrometer, and both techniques are capable
of analyzing a hundred or more samples in
eight hours.
The argon plasma source mass spectrometer
should be capable of accepting either liquid
or solid samples at atmospheric pressure.
Sensitivities and elemental coverage are
expected to be comparable to those of the
spark source mass spectrometer, and precision
and sample throughput should be comparable to
that of X-ray fluorescence and plasma
emission spectrometry. As mentioned earlier,
it is in the early stages of development.
Programs for reducing the limitations of
current capabilities in organic and
multielement analysis are underway at EPA's
Environmental Research Laboratory in Athens,
Georgia.f* Working closely in an iterative
manner with other environmental scientists we
hope to keep the monkeys on their backs to
determine the significance of chemicals that
the analytical chemist identifies and
measures in the environment.
References
1. Garrison, A. W. New Orleans Area Water
Supply Study—Analysis of Carbon and
Resin Extracts. U.S. Environmental
Protection Agency, Athens, Georgia. In
press. 48 p.
2. Taylor, C. E. Survey Analysis for Trace
Elements. Proceedings, ASMS 23rd
Conference on Mass Spectrometry and
Allied Topics. Houston, Texas, May 24-
30, 1975. In press.
3. Taylor, C. E. and W. J. Taylor.
Multielement Analysis of Environmental
Samples by Spark Source Mass
Spectrometry. U.S. Environmental
Protection Agency, Athens, Georgia.
Publication Number EPA 660/2-74-001.
January 1974. 23 p.
4. Webb. R. G., A. W. Garrison, L. H. Keith
and J. M. McGuire. Current Practices in
GC-MS Analysis of Organics in Water.
U.S. Environmental Protection Agency,
Athens, Georgia. Publication Number EPA
R2-73-277. August 1973. 91 p.
5. Keith, h. H. Analysis of Organics in Two
Kraft Mill Wastewaters. U.S.
Environmental Protection Agency, Athens,
Georgia. Publication Number EPA 600/4-
75-005. In press, p. 72.
6. Webb, R. G. XAD Resins, Urethane Foams,
Solvent Extraction for Isolating Organic
Water Pollutants. U.S. Environmental
Protection Agency, Athens, Georgia.
Publication Number EPA 660/4-75-003. In
press. 20 p.
7. Hoyland, J. R. and M. B. Neher.
Implementation of a Computer-Based
Information System for Mass Spectral
Identification. U.S. Environmental
Protection Agency, Athens, Georgia.
Publication Number EPA 660/2-74-048.
June 1974. 43 p.
8. Pitt, W. w., R. L. Jolley and S. Katz.
Automated Analysis of Individual
Refractory Organics in Polluted Water.
U.S. Environmental Protection Agency,
Athens, Georgia. Publication Number EPA
660/2-74-076. August 1974. 98 p.
9. Azarraga, L. V. and A. C. McCall.
Infrared Fourier Transform Spectrometry
of Gas Chromatography Effluents. U.S.
Environmental Protection Agency, Athens,
Georgia. Publication Number EPA 660/2-
73-034. January 1974. 61 p.
10. Camp, D. C. and A. L. VanLehn. X-Ray
Spectrometry £:123, 1975.
11. Fassel, V. A. and R. N. Kniseley.
Analytical Chemistry 46:1110A, 1974.
12. Gray, A. L. Analytical Chemistry 47:600,
1975.
13. Donaldson, w. T. Environmental
Analytical Chemistry 3:1, 1973.
3 15-5
-------
INVESTIGATION OF EXIT AREAS OF GROUNDWATER RELATED TO ANTHRACITE DEEP MINES
Carolyn A. Petrus
HRB-Singer, Inc.
Environmental Analysis Department
State College, Pennsylvania 16801
Introduction
Detection of surface and groundwater discharge
has been shown** to be a direct application of thermal
scanning. By using a thermal scanning system, it is
now feasible to locate mine water discharges that
were once both difficult and inefficient to establish
solely through extensive ground reconnaissance. The
Eastern Pennsylvania Southern Anthracite Field,
where the major anthracite reserves are located, was
selected as the study area to demonstrate this
application.
In the Southern Anthracite Field, the relation-
ship of groundwater to the level of the mine water
pools and the subsequent exit of such water into
the major streams is detected by the application of
established interpretation techniques to airborne
thermal imagery. Development of the method enables
a rapid and efficient means of locating point sources
and seep zones prior to the initiation of a ground
survey. Various studies have demonstrated the appli-
cability of infrared imaging as an effective tool for
recognizing coal industry-related environmental pro-
blems, including burning coal refuse banks and
underground mine fires. Since the mine water dis-
charge problem is probably the single most critical
physical factor limiting the development of increased
anthracite production, this study was undertaken to
produce a compendium of surface and groundwater dis-
charges in the area of anticipated anthracite coal
production.
Acknowledgments
The author acknowledges support by the United
States Bureau of Mines for contract support of a
portion of the work summarized in this paper. HRB-
Singer, Inc. also funded portions of the research.
Mr. John J. Rosella and Mr. Glen Evely,
Schuylkill Haven Environmental Affairs Field Office,
United States Bureau of Mines, gave generously of
their time to participate in field checking, water
sampling, and to discuss mining-related environmental
problems of the Southern Anthracite Field.
Ms. Jody Emel, HRB-Singer, Inc., participated in
data collection and field checking.
Principles of Thermal Scanning
Data for the study was obtained by using the HRB-
Singer, Inc. RECONOFAX XVI Calibrated Environmental
Science Infrared Scanning System. This system con-
sists of a precision, internally calibrated scantling
radiometer whose rotational axis is mounted parallel
to the longitudinal axis of the aircraft. The ground
surface of the study area is scanned perpendicular
to the flight direction.
Essentially, the system depends upon a sensitive
detector (Mercury-Cadraium-Telluride) whose peak
response is in the middle infrared region (8-12
micrometers). Surface emitted radiation is focused
through reflecting optics onto the detecting element
generating a weak electrical signal which is ampli-
fied, processed, and stored on magnetic tape.
In the laboratory, the taped signal is used to
modulate a cathode ray tube film printer producing
thermal film imagery. Thermal imagery shows surface
features of different omittance, therefore, illumi-
nation of the target or object is not necessary. In
fact, the data were collected at night to eliminate
the effects of solar heating.
Data Collection
Initially, the infrared data were collected in
May 1974 at 2000 hours at an altitude of 1219 meters
(4000 feet) above terrain. The ambient air temper-
ature during this flight was 7.77 degrees Centigrade
(46 degrees Fahrenheit). At this temperature, the
apparent temperature differences between the general
background and water discharges were at a minimum.
This tended to blend the water features on the
imagery into the general background making delineation
and interpretation difficult.
An analysis of the flight conditions resulted in
the recommendation that the study area be reflown at
a time when the ambient air temperature was below
freezing. A second flight was conducted in November
1974 at 1900 hours at an altitude of 610 meters
(2000 feet) above terrain producing imagery on the
scale of 1:24,000. The ambient air temperature was
-2.2 degrees Centigrade (28 degrees Fahrenheit). The
imagery resulting from this flight showed a signifi-
cant difference between the background temperature
(-2.2 degrees Centigrade) and the water discharges
with the areas of water discharge imaging signifi-
cantly warmer than the background (12.5 - 22 degrees
Centigrade or 54,5 - 72 degrees Fahrenheit). Black
bodies in the scanning system were set at 5 degrees
Centigrade (41 degrees Fahrenheit) and 24 degrees
Centigrade (75.2 degrees Fahrenheit). The imagery
produced was of very high quality related to this
application.
Data Analysis
The actual flight lines of the second data col-
lection mission were plotted on 1:24,000 topographic
base maps, and the imagery was analyzed in detail to
delineate all sources of groundwater, runoff, and
surface water. The original image negative was also
analyzed to insure complete delineation of all water
features.
The imagery was collected over the areas in the
Southern Field known to contain major discharge
sources. Only those areas within the flight lines
were analyzed in detail. All water sources, seeps,
entry and exit points, surface streams, standing
water, and drainage points delineated from the
imagery were plotted on standard 1:24,000 topographic'
maps.
Information provided by the United States Bureau
of Mines'' * as to the locations of known discharge
stations was cross-checked for their locations on the
imagery. If the known discharge was evident on the
imagery, it was assigned a number identification?
used by the Bureau of Mines. Other known discharges
as yet unidentified by the Bureau were assigned an
1
15-6
-------
TABLE 1 FIELD DATA FOR SELECTED STATIONS
75-13
MINE NO.
D 2)
NAME OF MINE DISCHARGE
MINE WATER DISCHARGES
ALT 1TUOE
OF
SURFACE
AT MINE (FT.)
ALTITUDE
OF
LOWEST
LEVEL (FT. )
ALTITUDE
OF
MINE WATER
POOL (FT.)
ALT 1TUOE
OF
DISCHARGE
(FT.)
ESTIMATED
VOLUME
gpm 3)
TEMP
°C
pH
IRON
(ppm)
28
REPPLIER -DARK WATER DRIFT
1 .830
15°
5.7
11.2
1 ,095
423
885
879
30
WADESV1LLE SHAFT
1 ,690
17.5
6.7
6.0
782
59
493
782
J
MT. LAFFEE BOREHOLE (NEW ST.)
1 ,520
12.5
6.1
6.0
-
-
-
-
27
PINEKNOT W. TUNNEL OVERFLOW
9,050
17.5
3.0
10.0
950
208
760
711
5
BROOKS IDE TUNNEL OVERFLOW
2,520
19.5
6.2
23.2
1 .412
-409
985
915
4
MARKSON SLOPE OVERFLOW
960
22.0
3.1
25.2
884
441
876
876
33
RANDOPLH SLOPE OVERFLOW
540
17.5
6.1
19.2
724
160
724
724
32
SALEM HILL WATER LEVEL TUNNEL
500 4)
18.5
7.0
5.2
636
355
636
636
43
TAMAQUA LANDS S.D1P REEVESDALE
OVERFLOW
830
16.0
4.1
2.7
959
531
962
959
1) SAME NUMBERS AS USED IN MICROFILMING PROJECT
2) BASED ON MEASUREMENTS OF GRAVITY DISCHARGES DURING THE PERIOD 1963 TO 1870
3) BUREAU OF MINES UNPUBLISHED INFORMATION
4) ESTIMATE
ST CLAjp:
i" !i> ^
St CLAIR- PlHt fOI EST
EAGLE [MILL
WAOESVILLt
PALMERf VS
-------
alphabetic notation. All known and identified dis-
charges were plotted on the base map along with their
identification code.
All related discharge data provided by Bureau of
Mines field work®> ® were tabulated^ to be used as
part of a field validation study. Information on
43 discharges included mine number, name, description
of discharge, altitude of the surface at the mine,
the lowest mine level, and the mine water pool level;
community name and minimum altitude; mine water dis-
charge altitude, estimated volume in gpm®, location,
and U.S.G.S. quadrangle name.
Field Validation
In order to validate the results of the imagery
interpretation and to attempt to establish a
relationship between the amount of flow, temperature,
and quality of the discharge, and its resolution on
the imagery, a field sampling program was undertaken.
The base maps and infrared imagery were used to
delineate selected sites oti which pH, temperature,
and iron content were measured. Iron and pH samples
were included because of the relationships of these
parameters to mine drainage permitting and control.
Sites were chosen as representative of different
types of discharges (boreholes, drifts, veins, water
level tunnels, and overflows). The sites were
selected and field studies were conducted with the
cooperation of personnel from the Schuylkill Haven
Office of the U.S. Bureau of Mines. Ground photo-
graphs were taken at each site to document the site
inspections. All data collected in the field along
with available Bureau of Mines' data for each site
were tabulated^.
The weather during the two days (February 4-5,
1975) of field work approximated that of the over-
flight. Subfreezing conditions, a 2.5-7cm snow
cover, 100% cloud cover, and intermittent freezing
rain dominated both days. Ten centimeters of fresh
snow fell on the night of February 4, 1975.
Data Presentation
The results of the interpretation of the thermal
imagery were presented in a final report to the
Bureau oi Miaes^. All surface waters delineated
from the imagery were plotted on 1:24,000 topographic
maps and differentiated as to type of surface water.
These data were presented on 11 maps which represent
the extent of the study area. Boundaries of mining
company land were outlined on the maps. The extent
of aerial coverage was also delineated. In addition,
annotated examples of thermal imagery were included
to correspond with the information plotted on the
topographic maps. Figure 1 is an example of the data
output for the Pottsville, Pennsylvania U.S.G.S. quad-
rangle map. Figure 2 represents a sample strip of
imagery of the Pottsville quadrangle on which the
interpretation was based.
The results of the field validation are presented
in Table 1, Field Data for Selected Stations.
Conclusions
An analysis of the airborne and ground data has
resulted in the following conclusions:
Environmental Limitations
It is of utmost importance for the maximization
of discharge detection that environmental conditions
be optimized. Imagery of highest quality is obtain-
able when there exists a minimum of vegetation, that
is, during late fall or winter. Maximizing the
temperature differentiation between background and
water features is imperative for sharp resolution.
Ground temperatures below freezing are optimum,
and light snow cover does not effect the ability of
the scanner to detect the warmer discharges. Flight
altitudes no higher than 610 meters (2000 feet)
above terrain should be maintained to insure detec-
tion of the smaller discharges.
Water Temperature and Flow
Detection of discharges by thermal infrared imagery
is primarily accomplished through temperature dif-
ferentials and flow rates. Based on a limited
sampling, when ambient air temperatures are sub-
freezing, the highest detection rate is achieved
through high temperature, high flow conditions. Of
these two factors, high temperature is more important
than high flow. In the case of high temperature and
low flow, the discharge is evident on the imagery,
but the opposite case is not. A low temperature,
high flow condition Is not as easily detected. Worse
case conditions combine both low temperature and
low flow.
Another consideration in detection or nondetection
of water discharges is the immediate background.
When background and discharge are of relatively equal
temperature, the background will tend to mask the
discharge. In instances where this is not the case
and the temperature of ground and discharge vary,
flows of as low as 50 gpm have been detected.
Water Quality
Identification of the type of discharge source
whether vein, drift, borehole, etc., is not generally
possible except in the case of mine pool overflows
where the temperature of the discharge source waa
noticeably higher.
Iron content and pH were measured for all field
stations. Iron values of samples requiring a 1:4
dilution with distilled water were considered as
unreliable and no conclusions were drawn from these
consistently high concentrations of iron in the
samples. Pool discharge, in addition to having
higher temperature, had higher pH readings than
other discharges. However, in the opinion of the
author, if there are no quantitative data, the
relative tones depicted on the infrared imagery are
usually too subtle to differentiate acid from non-
acid waters*'
The feasibility of the application of low level
infrared imaging to discharge water pipeline detec-
tion was inadvertently demonstrated on the Southern
Anthracite Field imagery. The existence of and dif-
ferentiation between operative and inoperative
pipelines in coal preparation plant operations was
clearly represented. This demonstration could
extend the capability of the infrared scanning
system to detection of leaking or otherwise damaged
discharge pipelines in remote or inaccessible areas.
Recommendations
Consideration should be given to utilizing thermal
scanning as an integral part of any airborne system
used in planning and monitoring surface effects of
coal mining.
3
15-6
-------
A specific problem exists in the Pennsylvania
Anthracite Region, that is, the problem of mine
water discharge, entry point and seep sreas. The
precise location of such water in the entire region
has not been completely delineated despite consider-
able field work by a number of state and federal
agencies. The water problem is probably the single
most critical physical factor limiting the develop-
ment of increased anthracite production. The
pumping costs (and treatment if required) even in
stripping operations are becoming prohibitive to
increasing production.
Coverage, at low altitude, of the remaining
anthracite fields is necessary to fully comprehend
the water discharge problem in the anthracite fields.
Further Investigations are needed to correlate physi-
cal parameters with naturally occurring discharges.
Infrared imagery at higher altitudes in conjunction
with existing aerial photography could be utilized to
establish whether a correlation exists between the
discharge locations and the occurrance of fracture
traces, lineaments, and other geologic phenomena.
Any further work would necessitate complimentary
field work.
It is recommended that a study be undertaken
whereby the entire anthracite region be overflown
with a thermal scanner at an optimum time. All
water areas should be delineated using the scanning
data and correlated with field work already being
conducted and with additional field validation
studies. The results of such a study would serve as
the base from which to validate past studies and to
identify other water sources which were not identi-
fied in previous studies. The study would mark
another operational use of remote sensing data to
attack a mining related environmental problem.
As has been previously documented for mine fires^
and thermal plumes^, digital computer processing of
the thermal data and the related outputs, including
thermal contour maps, grey scale representations, and
calibrated thermal profiles, are included in the
capabilities of the system"*. Potential for develop-
ment of a statistics package exists. First, second,
and third moments of frequency distributions of
temperature over a given area, in addition to curve
fitting using regression analysis techniques, can be
used to recognize patterns of distribution. In this
way, discharges with low resolution may be detected.
In summary, aerial infrared imagery was proven to
be an effective, comprehensive, and efficient tool
for the detection and delineation of fluvial dis-
charges related to mining activity. The inaccessibil-
ity of many discharge sources makes ground surveys
time consuming and costly. Thermal scanning in con-
junction with aerial photography can be used as a
valuable tool in detecting, delineating, and monitor-
ing environmental phenomena associated with coal
mining. Thermal scanning can supplement other data
by producing information on burning activity and
water discharge source.
2. Deely, D. and others, Applications of Aerial and
Orbital Remote Sensing to the Study of Mined
Lands. Proc., Research and Applied Technology
Symposium on Mined-Land Reclamation, Pittsburgh,
March 1973.
3. Knuth, W. and Charmbury, H. B., Remote Sensing
Techniques for Analysis of Burning in Coal
Refuse Banks. Proc., NCA-BCA Coal Conference
and Expo I, Coal and the Environment, Louisville,
Kentucky, October 1974.
4. Petrus, C., Detection of Surface and Groundwater
Discharge in the Southern Anthracite Field Using
Thermal Scanning. HRB-Singer, Inc., Report
4821F, U.S. Bureau of Mines, June 1975.
5. Stingelin, R., Airborne Infrared Imagery and Its
Limitations in Civil Engineering Practice.
Highway Research Record, Number 421, Remote
Sensing for Highway Engineering, 51st Annual
Meeting, 1972.
6. Taylor, J. I. and Stingelin, R., Infrared Imaging
for Water Resources Studies. Journal of the
Hydraulic Division, Proc., American Society of
Civil Engineers, January 1969.
7. U.S. Bureau of Mines, Microfilming Maps of
Abandoned Anthracite Mines (mines in the Southern
Field), unpublished information.
8. U.S. Bureau of Mines, unpublished information
based on measurements of gravity discharges
during the period 1963 to 1970.
9. U.S. Bureau of Mines, unpublished information.
References
1. Ahmad, M. U. and Ghosh, B. A., Temperature Survey
of Coal Mines Producing Acid Water. Proc., Seventh
International Symposium on Remote Sensing of Envi-
ronment, University of Michigan, Ann Arbor, May
1971.
4
15-6
-------
ASSESSMENT OF THE BENEFITS OF ENVIRONMENTAL REMOTE SENSING
Alex Hershaft, Ph.D.
Director of Environmental Studies
Enviro Control, Inc.
Rockville, Maryland
The objective of this paper is to present a
method for estimating the incremental benefits accru-
ing from the application of remote sensing information
to the more promising areas of environmental manage-
ment. The scope of these benefits is defined both by
the needs of environmental management decisions and by
the projected capabilities of the remote sensing system.
The method is illustrated by a case study to assess the
environmental benefits of data supplied by a continuing
Earth Resources Survey mission in the 1977-1986 time
frame.
Introduction
Purpose and Scope
In the fall of 1972, the U.S. Geological Survey
commissioned a study to determine the benefits and
costs of continuing the Earth Resources Survey (ERS)
mission beyond the two experimental Earth Resources
Technology Satellites (ERTS, now renamed Land Sat).
The bulk of the study consisted of six case studies
designed to examine the detailed benefits accruing from
the application of ERS information to land use planning,
range land management, forest management, agricultural
production assessment, water resources forecasting, and
environmental management. The study was completed late
last year and its conclusions naturally became the
subject of widespread controversy. Still, the environ-
mental case study can serve as a useful illustration of
an effective analytical approach to assessing the bene-
fits of environmental remote sensing.
Specifically, the objective of the environmental
management case study was to estimate the incremental
environmental benefits of the additional data provided
by a continuing ERS mission in the 1977-1986 time frame.
Environmental management is defined here as the manage-
ment of commonly owned earth resources (e.g., air and
water and wilderness areas) that are adversely affected
by human activities. The scope of the benefits is de-
fined both by the anticipated needs of specific environ-
mental management decisions and by the projected capa-
bilities of the ERS system. Senefits are credited only
for those ERS data that would not be provided anyway by
other competing sources. They are further reduced by
the cost of acquiring, processing, and applying these
data.
Analytic Approach
The specific approach adopted may be outlined as
follows:
• Select application areas
• Collect information on data use
• Analyze information for each application area
• Estimate benefits.
The major applicat ion areas were selected by
matching their general monitoring requirements against
the capabilities of the ERS system, projected to the
1977-1986 time frame on the basis of the results of
ERTS experiments and current R&D efforts. The major
areas selected were condition of disturbed lands,
condition of wetlands, and air and water quality.
Information on potential data use was obtained
through a comprehensive review of published literature
and in-depth interviews with some 70 environmental
data users representing 30 government agencies and
other institutions. The users were asked to evolve
"what-if" scenarios postulating likely environmental
management decisions and actions in the 1977-1986 time
frame, if the anticipated ERS data became available, as
well as the price they would be willing to pay for these
data. These scenarios were based largely on historic
legislative and budgetary trends, but necessarily also
on some "blue sky" speculation.
In the analysis phase, the demand for ERS data was
estimated through a match of detailed monitoring re-
quirements for all the anticipated management decisions
and actions in each application area against the specific
projected ERS system capabilities. The estimate of
monitoring requirements was based on the results of
user interviews, as well as on our own extrapolation of
current requirements in light of anticipated changes in
legislation and public attitudes.
Three types of environmental benefits of ERS data
in the 1977-1986 time frame were recognized: cost
savings, process benefits, and outcome benefits. Cost
savings would be credited when the ERS data could meet
at a lower cost data requirements currently supplied by
other sources. In this case, the "sunk" costs of the
other sources would not figure in the calculation, but
the "fixed" costs would, where applicable. Process
benefits would accrue when the ERS data could improve
the efficiency of the decision process. Outcome bene-
fits would be realized when ERS data made possible a
beneficial management decision that had not been feasible
before. Both process and outcome benefits were reduced
by the full projected cost of acquiring, processing, and
applying the ERS data. In each case, the final incre-
mental net benefits were estimated by applying the
"with" vs. "without" criterion. This involves comparing
the information available with and without the ERS
system, to isolate the incremental benefits of ERS data
and exclude benefits due to data available from competing
satellite systems and other sources.
Monitoring Requirements
The large number of laws enacted in recent years to
protect the nation's environmental resources has escalated
drastically the demand for environmental data. State
and local authorities must meet formal and stringent
requirements for data collection, storage, and retrieval.
In an idealized information flow, these raw data are
translated into the so-called earth resource information,
which then leads to management information, and eventually,
to an environmental decision and/or action undertaken on
the basis of this information. In assessing the utility
and benefits of a monitoring system, it is important to
trace its impact along this entire sequence.
The specific monitoring capabilities of the ERTS
system have been covered widely and need no reiteration
here. The major systemic advantages of satellite remote
sensing vis-a-vis in situ sampling and monitoring are
synopticity - the capability to present a simultaneous
pictorial representation of the entire area, serendipity
1
16-1
-------
the ability to oberve at a later date phenomena
that were not sought at the time the data were being
collected, rapid coverage, and low cost. The princi-
pal shortcomings, on the other hand, are low and
fixed pass frequency (18 days) and inability to
penetrate cloud cover (unless multiple satellites
are used), general dependence on reflected light,
poor penetration beneath water surface, poor spectral
and spatial resolution, and inability to yield
reliable quantitative estimates.
Environmental monitoring activities are normally
carried out in pursuit of one or more of the following
objectives:
• Detection and measurement of pollutants
discharges
• Evaluation of the effects of management
actions
• Assessment of condition and trends of
environmental quality.
Detection and measurement of pollutant discharges
demands the highest spatial resolution and frequency of
coverage and the lowest breadth of spatial coverage
among these three objectives. The area of interest is
highly localized and the contrast between the offending
plume and the surrounding medium is quite high. The
objective here is to detect violations of pollution
control regulations, to evaluate the relative contribu-
tions of various pollution sources, or to determine the
fate and distribution of the pollutant in a given medium.
Evaluation of the effect of a management action, such
as the closing of a source of discharge, or modification
of a stream flow, calls for a slightly lower spatial reso-
lution and frequency of coverage, and a somewhat higher
spectral resolution and breadth of coverage. The area
of interest is less localized and the changes sought are
less clearly defined. The important consideration here
is to establish a definite cause-effect relationship on
the basis of timing and nature of the change.
Finally, assessment of the condition and trends of
environmental quality requires the lowest spatial reso-
lution and frequency of coverage and the highest spectral
resolution and extent of coverage among the three ob-
jectives. The user here is interested in detecting small
changes over a large area, at relatively long time inter-
vals. The major concern is to avoid being misled by
local or short-term conditions and changes.
The more promising environmental applications of ERS
data were selected on the basis of the system capabilities
and monitoring requirements described above. They are
listed in Table 1 in terms of the earth resource informa-
tion obtained from interpretation of the observed
phenomena, the management information derived from the
Table 1. ENVIRONMENTAL REMOTE SENSING APPLICATIONS
Earth Resources
Information
Disturbed lands,
absence of vege-
tation, waste piles,
water bodies
Presence of vege-
tation
Disturbed lands
Land/water inter-
faces, vegetation
Presence of land,
absence of water
Sedimentation,
chemical pollution
Sediment plumes,
chemical plumes
Flow rate, sedi-
mentation, eutro-
phication
Aerosols,
particulates
Aerosols,
particulates
Management
Information
Mined lands
inventory
Status of
reclamation
Occurrence of
disturbance
Wetlands
inventories
Occurrence
of wetlands
disturbances
Occurrence of
discharges
Di spersion
patterns
Quality trends
Occurrence of
discharges
Dispersion
patterns
Management Decision/Action
Enactment of legislation
Enforcement of legislation,
remittance of reclamation bonds
Enforcement of legislation
Enactment of legislation,
establishment of hunting limits
Enforcement of regulations,
persuasion of owners
Identification of nonpoint
sources
Issuance and monitoring of
Users *
BM, regional commissions, states
BLM, GS, states
BLM, SCS, GS, states
FWS, NPS, SCS, states
FWS, NOAA, NPA, states
CG, CE, EPA, SCS, state and regional
agencies
CG, CE, NASA, NOAA, state, regional,
dumping permits, investigation and municipal agencies, disposal firms
of offshore circulation patterns
Establishment of abatement
priorities
CE, EPA, GS, NOAA, state and regional
agencies
Identification of point sources EPA, state agencies
Investigation of impact on NOAA
weather conditions
BLM - U.S. Bureau of Mines
BM - U.S. Bureau of Land Management
CG - U.S. Coast Guard
CE - U.S. Army Corps of Engineers
EPA - U.S. Environmental Protection Agency
FWS - Fish and Wildlife Service
GS - U.S. Geological Survey
NASA - National Aeronautics and Space Administration
NOAA - National Oceanic and Atmospheric Administration
NPS - National Park Service
SES - Soil Conservation Service
2
16-1
-------
earth resource information, and finally, the management
decision/action undertaken on the basis of this informa-
tion. The last column lists some of the more likely
federal and state users of the management information.
Condition of disturbed Lands
Information supplied by a continuing ERS mission
appears capable of updating disturbed lands inven-
tories, monitoring reclamation activities, and detect-
ing disturbances on virgin lands. These applications
are evolved and examined in terms of the nature of the
problem, monitoring considerations, and promising
applications.
Nature of the Problem
The most dramatic land disturbances in the United
States have been produced by surface mining of coal
which generated a major confrontation between the demand
for energy and environmental protection. With the
dwindling supply of natural gas and fuel oil, strip-
mining of low sulfur coal can be expected to grow
rapidly in the western states, which hold some 80 per-
cent of the nation's reserves of low sulfur coal and a
like fraction of all strippable reserves. The total
land area disturbed by surface mining for coal is
approximately 2,500 square miles and it is growing at
an annual rate of about 140 square miles. If all the
estimated available reserves of 128 billion tons of
strippable coal were recovered in this manner, the
disturbed land area could exceed 70,000 square miles.
The environmental effect of strip-mining includes
destruction of wildlife habitats and productive soils,
promotion of erosion and flooding, pollution of water-
ways by sediment and acid mine drainage, as well as the
loss of the aesthetic enjoyment of the countryside. A
meaningful reclamation, i.e., restoration of the land
to a useful condition, can reduce these environmental
disruptions, but it is usually very costly and sometimes
not feasible. The U.S. Congress has failed repeatedly
to enact effective national legislation requiring
reclamation of stripped land, but more than half of
the states do have some requirements to that effect.
Monitoring Considerations
The principal monitoring requirements for disturbed
lands are those associated with performance of inven-
tories and determination of reclamation status. In
addition, there is some call for detection of disturb-
ances in virgin lands held by. the federal government in
the western states. Concern for reclamation in the
eastern states has centered on the large tracts of land
now being mined and a considerable acreage of previously
mined and unreclaimed or "orphan" lands. In the western
states, attention is directed to monitoring of current
mining and reclamation operations and inventories of
resources that will be disturbed by future mining
operations.
The performance of stripped land inventories
requires the capability to discriminate between vegeta-
tive cover and stripped, or otherwise disturbed land,
and to detect waste piles and water impoundments, and
can be performed by remote sensing. Monitoring of the
status of reclamation, on the other hand, presents a
more complex problem, because of the additional informa-
tion required. Many state laws state that spoil ridges
must be graded to a rolling topography or to the approxi-
mate original contour. The vegetative cover usually must
be self-regenerating and compatible with the surrounding
flora. Flowing water must be channeled to reduce
erosion and siltation, and water bodies are to be safely
impounded. Consequently, this type of monitoring is
usually performed by earthbound mining inspectors.
ERTS sensors are capable of differentiating
among three gross types of vegetation, of identifying
and delineating disturbed lands associated with area
surface mining, and of detecting major areas of
contour mining. The smallest area that can be mapped
is about five acres and mapping accuracies in excess of
90 percent have been recorded, especially for areas in
excess of one square mile. Other features of mining
activities, such as refuse piles and slurry ponds have
been identified as well. These results apply largely
to the eastern states, where newly disturbed surface
of the mined land contrasts sharply with the surrounding
vegetative cover. In the western states, where mines
are frequently located in arid areas, the contrast and
the resulting mapping accuracy is likely to be degraded.
Promising Applications
On the basis of the above considerations and in-
depth discussion with users of monitoring data on
disturbed lands, the following applications were
selected as showing the most promise:
• Updating of inventories of strip-mined land
on the basis of observations of disturbed
land, absence of vegetation, waste piles,
and slurry ponds, to assist in enactment of
protective legislation
• Monitoring of the status of reclamation of
mined land, on the basis of observation of
disturbed land and vegetation density, to
assist in the enforcement of permit regulations
• Detection of disturbances of virgin lands, on
the basis of observation of disturbed land
and absence of vegetation, to assist in the
enforcement of protective legislation.
Several other potential applications, such as assessment
of reclamation priorities of orphan land were nut
considered feasible, because of inadequate detection
and resolution capabilities of the ERS system.
Although ERS information alone would not be
adequate to perform a national inventory of strip-mined
lands, it could provide periodic updates on a national
inventory, once the nature and boundaries of specific
strip-mined areas had been initially determined by
high altitude photography. ERS imagery could also
serve to reduce the amount of aircraft flying by
"targeting" the more important areas, but this is a
"one-time" benefit obtainable from the ERTS experimental
satellites, and thus, could not be credited to a
continuing ERS mission.
The application of ERS imagery to monitoring of
reclamation status would not be very useful, because
mine inspectors would still have to visit the area to
check for outeroppings of toxic minerals and to collect
other information not obtainable by remote sensing.
However, ERS information could serve to verify the
inspectors' findings in active mines or to monitor
reclamation of inactive or abandoned mines which are
not visited by inspectors.
Agencies, such as the Bureau of Land Management,
that are responsible for managing lands in western
states could clearly benefit from a monitoring system
that would alert them to the occurrence of major
disturbances by developers and others. Most such
disturbances in excess of several acres should be
detectable by ERTS sensors in time to prevent excessive
damage. Field crews could then be dispatched to the
3
16-1
-------
area to determine the extent of the disturbance and to
initiate whatever remedial action may be necessary.
This would permit the achievement of the current moni-
toring effectiveness with less personnel or of increased
effectiveness with the same personnel.
Finally, the demand for data in each of the appli-
cations during the 1977-1986 time frame was estimated
on the basis of the anticipated legislative climate,
the potential users of environmental data, and the
projected amount of data required. The legislative
climate assumed the continuing growth in the number of
states requiring the reclamation of strip-mined lands
and the enactment of some form of national legislation
to this effect. Potential users include the U.S. Bureau
of Mines, the U.S. Bureau of Land Management, the U.S.
Geological Survey, the Soil Conservation Service, the
various regional commissions and appropriate agencies
of some thirty states. The amount of monitoring data
required was projected by assuming continuing growth
in U.S. energy demand and in reliance on coal to satisfy
this demand, as well as increase in the fraction of
U.S. coal produced by surface mining, and in the
stringency of state reclamation requirements.
Condition of Wetlands
Information supplied by a continuing ERS mission
appears capable of updating wetland inventories and
detecting major encroachments on wetland integrity.
These applications are evolved and examined in terms of
the nature of the problem, monitoring considerations,
and promising applications.
Nature of the Problem
Wetlands, i.e., areas that are either covered by
water or have a very high water table, may be classified
as natural or artificial, permanent or seasonal, inland
or coastal, and fresh or salty. They are found in
every state and territory of the United States, but are
more prevalent in Florida and other South Atlantic
states. In the aggregate, U.S. wetlands cover an area
of approximately 116,000 square miles.
Wetlands represent a prime example of areas of
critical environmental concern, because they serve a
vital ecological and aesthetic function, yet are very
vulnerable to destruction through draining, filling,
or pollution. Wetlands provide for wildlife habitat
and breeding grounds, growth of cash crops, storage of
ground water recharge supplies, shoreline stabilization
and sediment retention, and aesthetic diversity.
Coastal wetlands are prime targets for residential and
commercial development, whereas inland wetlands are
frequently drained and converted into crop lands.
Specific threats to wetland preservation include dredge
and spoil operations, housing and other construction,
surface mining, agricultural cultivation, and waste
disposal,
Monitoring Considerations
Protection of wetland areas requires proper manage-
ment and control of the various threats described above,
under federal and state legislation. In order to react
in a timely and effective manner to these threats,
suitable information must be collected and disseminated
rapidly to the enforcement agencies. The principal
items of management information required for wetlands
management are identification and location of wetlands
and assessment of their condition.
The identification and location of wetlands require
the capability to delineate water bodies and to differ-
entiate among gross types of vegetation. Assessment of
the condition of wetlands requires the capability to
determine the presence of water and its quality and the
condition of the vegetation. Information on vegetation
provides a key indication of the changing role of a
wetland, whereas the presence and quality of Water
provide an early warning of the eventual fate of the
vegetation.
ERTS imagery has been used successfully in deline-
ating coastal and some types of inland wetlands, to
detect encroachments on the integrity of wetlands,
and under favorable conditions, to discriminate among
certain classes of vegetation. ERTS imagery has been
used to identify prairie potholes as small as 2 acres
and to measure wetland areas in excess of 1 square mile
with a 95 percent accuracy. Smaller wetlands disturb-
ances have been detected with the aid of computer
enhancement techniques. On the other hand, attempts to
assess the vigor of vegetation and water quality have
had very limited success.
Promising Applications
The most promising applications of ERS information
for wetlands management were selected as follows:
• Performance of wetland inventories on the
basis of delineation of land/water boundaries
and determination of gross vegetation classes,
to assist in enactment of protective legisla-
tion and in establishment of hunting limits
• Detection of wetlands disturbances on the
basis of observation of the presence of
land (filling) and the absence of water
(draining), to assist in enforcement of
protective regulations and in persuasion of
owners to preserve their wetlands.
The detection of degradation of wetlands, on the basis
of observations of vegetation vigor and water quality,
in order to assist in establishing pollution abatement
priorities, was not considered a practical application
because of inadequate ERS detection and resolution
capabi1i ties.
The ERS system could assist in the compilation of
a wetland inventory or to update an existing inventory.
Interpretation of ERS imagery could reduce the initial
inventorying effort by "targeting" the more important
wetland areas, but, as in the case of stripped lands
inventory, this is a "one-time" benefit that cannot be
credited to a continuing ERS mission. Its capability
to delineate the land/water interface and large roono-
typic classes of vegetation, combined with such other
known parameters as location and salinity, could satisfy
a rather detailed classification scheme. Moreover,
ERS would be capable of identifying the 20 percent
of prairie potholes that are over 2 acres in size,
and thus, improving considerably the accuracy of
waterfowl availability predictions.
The role of the ERS data in the protection of
wetlands would be to detect disturbances and to
direct state inspectors to the site of the violation
in time to minimize the extent of damage. At the
present time, this task is performed by a handful of
inspectors who have a vast territory to cover and
frequently are unable to discharge their responsibility
in a timely and effective fashion. Hereagain, one has
the option of maintaining the same degree of effective-
ness with fewer inspectors or increasing this effective-
ness with the same number of inspectors.
As in the case of disturbed lands, the demand for
wetland monitoring data was estimated on the basis of
the anticipated legislative climate, potential user
institutions, and the projected amount of data to be
4
16-1
-------
collected. The legislative climate will be governed
largely by the Fish and Wildlife Coordination Act of
1956, which has been interpreted to authorize national
wetland inventories, the Coastal Zone Management Act
of 1972, designed to preserve and protect the nation's
coastal zone, including wetlands, tidelands, beaches,
and coastal waters, and the Marine Protection, Research,
and Sanctuaries Act of 1972, providing for the designa-
tion of marine sanctuaries to preserve or restore areas
with conservational, recreational, ecological, or
aesthetic value. In addition, nearly 30 states have
enacted some form of legislation to protect wetlands
within their jurisdiction.
Potential users of wetland monitoring data include
the U.S. Fish and Wildlife Service, the National Oceanic
and Atmospheric Administration, the U.S. Army Corps of
Engineers, the Soil Conservation Service, the National
Park Service, the Smithsonian Institution, and appro-
priate agencies of some 30 states. The amount of data
required has been projected by assuming that ERS
imagery will assist the Fish and Wildlife Service in
conducting its semi-annual survey of prairie potholes
and in updating a national wetlands inventory at
5-year intervals. It was further assumed that ERS
imagery will meet a rising demand for detection of
wetland disturbances.
Water and Air Quality
Information supplied by a continuing ERS mission
appears capable of detecting massive waterborne dis-
charges, characterizing hydrologic circulation patterns,
and determining water quality trends in large lakes and
estuaries and in offshore waters. Its contribution to
management of air quality is expected to be minimal,
because of the satellite's limited capabilities for
detection and spectral discrimination of airpollutants
and low pass frequency relative to the dynamics of air
quality events. These applications are evolved and
examined below.
Nature of the Problem
The deteriorating quality of our waters has been
long considered the most serious environmental problem
facing the nation. Consequently, water pollution
control has received far more funding than any other
area of environmental concern and the improvement of
the water quality management process promises a high
return. Traditionally, government regulatory programs
have focussed on control of discrete "point" sources
of water pollution, such as municipal and industrial
outfalls. More recently, however, attention has been
shifting toward "non-point" sources, which contribute
most of the pollution loading on the national level.
These may be caused by agricultural, silvicultural,
mining, or construction activities, urban runoff, solid
waste disposal, or offshore dumping.
Air pollutants are also emitted both by point
sources, including power plant, industrial, or incin-
erator stacks, and by area or line sources, such as
smokestacks in housing developments or automobile
exhausts along a highway. Major point sources are of
special interest, because they are susceptible to
effective abatement through installation of control
devices.
Monitoring Considerations
The major objectives of water quality monitoring
are detection and measurement of discharges and
determination of water quality. The first objective
serves both to enforce discharge regulations and to
assess the waste load contributions of various sources,
a necessary step in wastewater management planning.
Measurement of the discharge plume reflects the fate of
pollutants, which is particularly important in the
vicinity of such sensitive areas as municipal water
intakes or beaches. Determination of water quality
serves to assess progress resulting from wastewater
management activities. It reveals whether water quality
standards consistent with designated water use have been
met and helps in establishing interim water uses.
Current water quality monitoring practices vary
widely in sophistication from networks that utilize
electronic sensors measuring several chemical and
physical parameters to the collection of a "grab" sample
and its analysis in a field laboratory. Major monitoring
networks and data storage and dissemination facilities
are maintained by the U.S. Geological Survey and the
U.S. Environmental Protection Agency. Water quality
data are also collected by a number of state and local
organizations for a variety of purposes.
The ability of the ERTS system to detect and map
large-scale turbidity, as well as growth and movement
of algae and other aquatic organisms in large lakes,
rivers, and estuaries has been demonstrated by several
investigators and semi-quantitative estimates have been
attempted. High concentrations of industrial acid waste
have been observed, but attempts to detect acid mine
waste and oil slicks have not been successful. The
major drawback of ERTS sensors in this application is
their inability to penetrate beneath the water surface.
The ability of the ERTS system to detect atmos-
pheric pollutants is severely limited by interference
from the spectral response of the background. Conse-
quently, air pollutants can be observed only over
homogenous backgrounds, such as the Great Lakes.
However, ERTS imagery has been used successfully to
detect large plumes from point sources, especially
when these sources emit particulates and aerosols in
addition to gaseous pollutants. Areawide air pollution
in the form of aerosol and haze has been delineated in
some cases by special techniques.
Promising Applications
The following applications were selected as
showing promise for water quality management:
• Detection of sediment discharges from
non-point sources to assist in establishing
and enforcing waste load allocations in
large bodies of water
• Delineation of dispersion patterns from
offshore dumps to assist in the issuance
and enforcement of dumping permits and in
the investigation of offshore circulation
patterns
• Characterization of long-term trends in
sedimentation and eutrophication of large
bodies of water, to assist in establishing
pollution abatement priorities.
The identification of point sources was not considered
a promising application because of the poor resolution
of ERS sensors and the availability of this information
from more conventional sources. Enforcement of dis-
charge regulations was disqualified by the predictability
of ERS passes, in addition to the resolution problem.
Determination of water quality trends was undermined by
the projected poor spectral sensitivity of ERS sensors.
In the first application, the ERS system could
serve to provide preliminary assessment of the discharge,
to direct ground inspection teams to the more critical
or urgent discharges, and to detect unexpected discharges
16-1
-------
in remote areas of the ocean. However, the use of ERS
data for monitoring non-point sources would be degraded
by the fact that discharges from these sources are
typically accentuated after a major rainfall, when
cloudy skies are likely to obscure satellite observa-
tion.
With regard to offshore dumping, the present EPA
monitoring program calls for water quality surveys
around the dumping site at three-month intervals. Since
the waste plume could well cover the distance to shore
in several hours, under adverse conditions, such surveys
clearly do not provide a timely characterization of the
plume's behavior. ERS imagery supplied at 18-day inter-
vals, weather permitting, comes much closer to the ideal
monitoring program. An additional long-range benefit
would accrue from the use of composite ERS observations
of waste plumes to suggest elimination of unsuitable
dumping sites and location of new ones. The specific
utility of using ERS imagery to delineate sediment
plumes in the study of coastal processes is difficult
to assess at this time.
Assessment of national trends in water quality
serves to inform the public of the magnitude of water
quality problems and the rate of their improvement or
deterioration, to identify critical regions and
pollutants and help establish cleanup priorities, to
indicate the effectiveness of past pollution abatement
efforts, and to direct the location and focus of
future monitoring efforts.
Demand for water quality data provided by the ERS
mission was estimated on the basis of the legislative
climate, interviews with potential users, and the
projected amount of data required. The legislative
climate will be shaped largely by the Federal Water
Pollution Control Act of 1972 and any future amendments.
The Act provides for extensive and growing water quality
monitoring programs by federal and state agencies. The
Marine Protection, Research, and Sanctuaries Act of
1972 provides for monitoring of the offshore dumping of
wastes. Potential users of ERS water monitoring data
are the U.S. Environmental Protection Agency, the U.S.
Geological Survey, the U.S. Army Corps of Engineers,
the National Oceanic and Atmospheric Administration,
and numerous regional, state, and local agencies, such
as regional water resources planning commissions, state
and county environmental protection and health depart-
ments, and municipal wastewater management agencies.
In the case of air quality, the promising
applications were:
• Identification of major point sources to
assist in compilation of national inven-
tories for regulatory purposes
The use of ERS data to detect major point source
emissions would benefit primarily EPA's National
Emission Data System by identifying overlooked point
sources and updating the inventory periodically. On
the other hand, much of the information required for
the inventory is not obtainable by remote sensing and
would still require visits by earthbound inspectors.
The utility and benefits of observing air emission
plumes to learn their impact on weather conditions
can not be assessed at this time.
Conclusions
The analysis described here suffers from a number
of uncertainties and assumptions. The major uncertain-
ties lies in the projected capabilities of the ERS
system and availability of competing data sources as
well as in the anticipated nature of future environ-
mental management decisions and actions, and the
resulting demand for data. The principal assumptions
are that environmental data will be obtained in the
most efficient manner from the best available source,
that environmental management decisions will be based
on the best available data, and that these decisions
will produce the anticipated beneficial results. For
these reasons, benefits can only be expressed here in
qualitative terms as follows:
• There are substantial benefits in updating
national stripped-lands and wetlands
inventories
• There are some benefits in monitoring stripped
lands reclamation, and integrity of federal
lands and wetlands
• There are likely benefits in monitoring off-
shore dumping operations and in observing
sediment plumes in coastal waters
• There are minor benefits in detection of
runoff discharges and long-term water
quality trends in major bodies of water
• There may be some benefits in verifying and
updating inventories of major point sources of
air pollution and in observing the impact of
air pollution plumes on weather conditions.
It appears that in all the environmental applications
considered, ERS information would complement, rather
than replace, in situ monitoring data. The major roles
of ERS information would be to guide the ground and
aircraft monitors to the more critical areas, to detect
changes in environmental condition defined by more
precise monitoring techniques, and to correlate informa-
tion obtained by other monitoring techniques over a
broad geographic area.
Acknowledgement
• Observation of plume patterns to assist in
investigating the impact of air pollution
and cloud formation.
The use of ERS data in enforcement of emission regula-
tions was not considered feasible because of the poor
resolution, coverage frequency, and predictability of
satellite passes. The value of ERS data in siting
emission sources was negated primarily by availability
of the required information from other sources. Deter-
mination of air quality trends suffered from lack of a
capability to detect gaseous pollutants and to provide
quantitative estimates.
The guidance of Mr. H. Theodore Heintz, Jr., now
of the Office of Policy Analysis, Department of the
Interior, and his contribution to the development of
the method of analysis are gratefully acknowledged.
During the performance of this project, Mr. Heintz and
the author were employed, respectively, by the Earth
Satellite Corporation and Booz, Allen & Hamilton, Inc.
6
16-1
-------
MONITORING ESTUABINE CIRCULATION AND OCEAN WASTE DISPERSION
USING AN INTEGRATED SATELLITE-AIRCRAFT-DROGUE APPROACH
V. Klemas, G. Davis and H. Wang
College of Marine Studies
University of Delaware, Newark, Del.
W. Whelan and G. Tornatore
ITT Electro-Physics Laboratories, Inc.
Columbia, Md.
SUMMARY
The mounting economic pressure to extract oil
and other resources from the Continental Shelf and to
continue using it for waste disposal is creating a
need for cost-effective, synoptic means of determining
currents in this area. An integrated satellite-air-
craft-drogue approach has been developed which employs
remotely tracked expendable drogues together with
satellite and aircraft observations of waste plumes
and tracers, such as dyes or suspended sediment.
Tests conducted on the Continental Shelf and in Del-
aware Bay indicate that the system provides a cost-
effective means of studying current circulation, oil
slick movement and ocean waste dispersion even under
severe environmental conditions.
INTRODUCTION
There exists an urgent need to better understand
the Continental Shelf environment because of economic
pressures to extract oil and other resources; to in-
crease the harvest of food; to continue using it for
waste disposal; and to route ships or conduct small-
craft rescue operations effectively. The large
concentration of population in the coastal zone and
the accompanying increase in utilization pressure is
likely to have deleterious effects on the shelf re-
gions. The offshore-onshore transport rates of pol-
lutants, sediments and nutrients strongly influence
the ecology of the coastal zone. In order to keep
the enviornmental impact within acceptable levels,
it is important to understand the circulation and
exchange processes on the shelf.
In the last decade the data base on shelf water
movement has increased significantly.* A large number
of continuous records of variability on the shelf
have been produced with self-contained current and
temperature recorders and shelf-bottom pressure gauges.
In addition, the uae of modern instruments such as
CTD's (current-temperature-depth), STD's (aalinity-
temperature-depth), and XBT's (expendable bathythermo-
graphs) has increased the knowledge of water-mass
structure.
The Eulerian method of measuring simultaneously
the current direction and speed at preselected points
in the water column requires many ships and current
meters when synoptic measurements over large coastal
areas are to be made. Therefore an inexpensive, in-
tegrated aatellite-aircraft-drogue approach has been
developed which is based on the Lagrangian technique
and employs remotely tracked drogues and dyes together
with satellite observation of natural tracers, such as
suspended sediment. The overall objective is to em-
ploy this cost-effective technique to improve our
understanding of estuarine and shelf circulation, par-
ticularly at critical sites where disturbances of
shelf environment are anticipated.
CURRENT CIRCULATION PROM SATELLITE IMAGERY
Using suspended sediment as a natural tracer, it
is ^possible to study current circulation in the sur-
face layers of turbid estuaries and coastal waters
by employing satellite imagery and a small amount of
ground truth data.2
Imagery and digital tapes from ten passes of the
Earth Resources Technology Satellite (ERTS-1) and one
successful Skylab pass over Delaware Bay were analyz-
ed. The ERTS-1 imagery used in our work was produced
by the four-channel multispectral scanner (MSS) hav-
ing the following bands'.
Band 4 0.5 - 0.6 Microns
Band 5 0.6 - 0.7 Microns
Band 6 0.7-0.8 Microns
Band 7 0.8 - 1.1 Microns,
From an altitude of 920 km, each frame covered
an area of 165 km by 185 km. In addition to the 9-
track 800 bpi magnetic tapes, reconstructed negative
and positive transparencies in 70 millimeter format
and prints in nine inch format were obtained from
NASA. Before visual interpretation, some of the
imagery was enhanced optically, using density slicing
and color additive techniques. Annotated thematic
maps were prepared by computer analysis of digital
tapes and by direct photointerpretation of the trans-
parencies reconstructed by NASA.2
Only MSS Band 5 images are shown, since the "red"
band was found to give the best contrast In delineat-
ing suspended sediment concentration in the upper one
meter of the water column. Adjacent to ERTS-1 pic-
tures, Figures 1, 2 and 3 contain NOAA-NOS tidal cur-
rent maps for Delaware Bay. Each ERTS-1 picture is
matched to the nearest predicted tidal current chart
within + 30 minutes. The current charts indicate the
hourly directions by arrows, and the velocities of
the tidal currents in knots, the Coast and Geodetic
Survey made observations of the current from the sur-
face to a maximum depth of six meters in compiling
these charts.5'®
The satellite picture in Figure 1 was taken on
October 10, 1973, two hours after maximum flood at
the entrance of Delaware Bay. Masses of highly tur-
bid water are visible around the shoals near the mouth
of the bay and in the shallow areas on both aides of
the bay. Since at that time flood currents were pre-
vailing throughout the bay, some of the sediment in
suspension seems to be locally generated over shoals
and shallow areas of the bay resulting In a higher
degree of backscatter from ahallover waters. During
flood tide at the mouth of the bay, considerable cor-
relation was found between the depth profile and
image radiance, even though the water depth exceeded
the Secchl depth by at least a factor of three in all
areas. No such correlation was found during ebb tide
when finer sands and clay are carried down the rivers
into the bay. At the time of Tnis ERTS-1 picture,
16-2
-------
the wind velocity was 13 to 22 km per hour from the
north causing surface currents in nearly the opposite
direction to the tidal flow. Peak flood current ve-
locity is occurring in the upper portion of the bay,
creating sharp boundaries along the edges of the deep
channels.
Figure 2 represents tidal conditions two hours
before maximum flood ad the mouth of the bay observed
by ERTS-1 on January 26, 1973. High water slack is
occurring in the upper portion of the bay, resulting
in less pronounced boundaries there as compared to
Figure 1. The shelf tidal water is not rushing along
the deep channel upstream anymore as in Figure 1, but
is caught between incipient ebb flow coming down the
upper portion of the river and the last phase of the
flood currents still entering the bay. On the morn-
ing of January 26, 1973 there was a variable wind
over the bay at about 9 to 11 km per hour from the
south-southwest, helping tidal currents push clearer
shelf water against highly turbid water masses in
Mew Jersey's shallow flats.
The satellite overpass on February 13, 1973
occurred about one hour after maximum ebb at the
mouth of Delaware Bay. The corresponding ERTS-1 image
and predicted tidal currents are shown in Figure 3.
Strong sediment transport out of the bay in the upper
portion of the water column is clearly visible, with
some of the plumes extending up to 30 km out of the
bay. Small sediment plumes along New Jersey's coast
clearly indicate that the direction of the nearshore
current at that time was towards the north. The wind
velocity at the time of the satellite overpass was
about 13 km per hour from the west-northwest, rein-
forcing the tidal current movement out of the bay.
In addition to current circulation studies,
ERTS-1 image radiance of Band 5 was correlated with
suspended sediment concentration and Secchi depth
data obtained from boats and helicopters during the
satellite overpass.2 A suspended sediment concentra-
tion map based on ERTS-1 image radiance correlation
with water sample analyses is shown in Figure 4, for
tidal conditions identical to those in Figure 2.
FRONTAL SYSTEMS AND OIL SLICK MOVEMENT
Boundaries or fronts (regions of high horizontal
density gradient with associated horizontal converg-
ence) are a major hydrographic feature in Delaware
Bay and in other estuaries. Fronts such as the ones
shown in Figure 1 have been Investigated using STD
sections, dye drops and aerial photography. Horizon-
tal salinity gradients of 4 loo in one meter and con-
vergence velocities of the order of 0.1 m/sec. have
been observed. Underwater visibility improved from
one meter to two meters as certain boundaries were
crossed. Several varieties of fronts have been seen.
Those near the mouth of the bay are associated with
the tidal intrusion of shelf water. The formation
of fronts in the interior of the bay appears to be
associated with velocity shears induced by differ-
ences in bottom topography with horizontal density
difference across the front influenced by vertical
density difference in the deep water portion of the
estuary. Surface slicks and foam collected at frontal
convergence zones near boundaries were found to con-
tain concentrations of Cr, Cu, Fe, Hg, Pb, and Zn
higher by two to four orders of magnitude than concen-
trations in mean ocean water.7
In order to assess surface currents at a proposed
oil terminal site eight miles east of Cape Henlopen,
nearly 150 1 of Rhodamine WT dye were spread into a
large circular slick about 300m in diameter (Figure 5).
The dye patch was made large enough to be observed
from the Earth Resources Technology Satellite. Sub-
sequent cloud cover prevented the satellite from
imaging the dye slick, but it was traced successfully
by aircraft. The dye was injected about one hour be-
fore the predicted flood current peak at the mouth
of the bay. A frontal system like the one shown in
Figure 6 passed over the dye slick, attracted the
dye into the boundary, and carried it for about two
miles northwest toward the mouth of the bay. An
aerial photograph obtained by enhancing the dye with
a Wratten 25A red filter is shown in Figure 7.
By capturing and holding oil slicks, these fron-
tal systems also significantly influence the movement
and dispersion of oil slicks in Delaware Bay. Recent
oil slick tracking experiments conducted by the
authors in order to verfiy a predictive oil disper-
sion and movement model have shown that during cer-
tain parts of the tidal cycle the oil slicks tend to
line up along boundaries. This effect was illustrat-
ed by an oil spill which occurred on January 10,
1975 as a result of a lightering operation in the
anchorage area off Big Stone Beach in Delaware Bay.
As shown in Figures 8 and 9, at 0930 hours the spill
consisted of four large slicks and numerous smaller
ones almost randomly dispersed throughout the area.
Although for a few hours the movement of the oil was
influenced by wind and current forces, by 1500 hours
most of the oil was aligned along the boundaries and
stretched into two five-mile long slicks. Thus, un-
usual oil slick distribution patterns resulted which
even for a known oil type cannot be predicted on the
basis of wind and tidal current information alone.
SATELLITE OBSERVATIONS OF OCEAN WASTE DISPOSAL PLUMES
Approximately forty nautical miles off the Del-
aware coast is located the disposal site for waste
discharged from a plant processing titanium dioxide.
The discharge is a greenish-brown, 15 to 20% acid
liquid which consists primarily of iron chlorides and
sulfates. The barge which, transports this waste has
a 1,000,000 gallon capacity and makes at least three
trips to the disposal site per month.
The frequency of this dunping made it possible
for the ERTS-1 satellite to image the acid plume in
various stages of degradation, ranging from minutes
to days after dump initiation. As shown in Table 1,
nine photographs were found which show water dis-
colorations in the general vicinity of the waste dump
site. The position of the discoloration, the dump
pattern and the time difference between the dump and
photograph give strong indications that the dis-
coloratlons are the acid plume,
TABLE I
List of ERTS-1 Images Containing Acid Waste Disposal
Plumes and Satellite Overpass Time in Hours After Dump
DATE
I.D. NUMBER
TIME AFTER DUMP
10
October
72
1079-15133
9 hrs 38 min
27
October
72
1096-15081
14 hrs 8 min
25
January
73
1186-15081
4 hrs 3 min
07
April 73
1258-15085
4 hrs 3 min
13
May 73
1294-15083
During Dump
22
October
73
1456-15055
29 hrs 25 min
23
October
74
1457-15113
53 hrs 31 min
15
December
73
1510-15052
5 hrs 45 min
26
May 74
1672-15012
21 hrs 6 min
Carefull examination of the overpass of January
25, 1973 disclosed a fishhook-shaped plume about 40
2
16-2
-------
miles east of Cape Henlopen caused by a barge dispos-
ing acid wastes. The plume shows up more strongly in
the green band than in the red band. Enlarged en-
hancements of the acid waste plumes, prepared from
the ERTS-1 MSS digital tapes (Figure 10) aided con-
siderably in studies of the dispersion of the waste
plume. Currently acid dumps are being coordinated
with ERTS-1 overpasses in order to determine the dis-
persion and movement of the waste materials along the
Continental Shelf.
Sludge disposal plumes in the ocean off the Del-
aware coast have also been detected in ERTS-1 imagery.
However, they are more difficult to track due to their
less frequent dump schedule.
DROGUE DESIGN AND APPLICATION
One important shortcoming of satellite investi-
gations of coastal currents by remote sensing has
been their inability to determine current magnitudes
and to penetrate beyond the upper few meters of the
turbid water column. These objections can be over-
come by complementing the satellite observations with
data from current meters and drogues capable of track-
ing currents at any depth encountered on the Conti-
nental Shelf.
The Eulerian method of measuring simultaneously
the current direction and speed at preselected points
in the water column requires many expensive current
meters if synoptic measurements over large coastal
areas are to be made. Considerable cost-effective-
ness can be attained if Eulerian techniques are com-
plemented by Lagrangian methods employing expendable
drogues in combination with aircraft and satellite
observations.
Two types of drogues were used. The first are
small, compact units which can be dropped and tracked
from low-flying aircraft. Their basic design does
not differ significantly from that of drogues used by
various investigators during the past few decades.®
These small drogues are deployed whenever a detailed
charting of current circulation over a relatively
small area, such as four square miles, is desired.
As shown in Figure 6, the drogues consist of a styro-
foam float and a line to which is attached a current
trap consisting of a stainless steel biplane. The
length of the line determines at what depth currents
will be monitored. The floats are color-coded to
distinguish their movement and mark the depth of the
biplanes. Packs with dyes of two different colors can
be attached to the float and the biplane.9'10 The
movement of the dye and drogues is tracked by sequen-
tial aerial photography, using fixed markers on shore
or on buoys as reference points to calibrate the scale
and direction of drogue movement.
The second type of drogue is a "radio-sonde"
which can be released to track currents over large
areas, such as an entire segment of the Continental
Shelf. The unit was developed by the I.T.T. Electro-
Physics Laboratories with the intention of providing
free-drifting drogue which would be sufficiently eco-
nomical to render it unnecessary to recover the ex-
pended hulk. The sea-sonde employs a low-powered HF
radio transmitter radiating 1 to 50 milliwatts in the
6 MHz region, to telemeter its position and any of
several other optional variables of interest in-
cluding:
a. wind speed
b. water depth
c. temperature
d. conductivity.
From the latter two salinity can be deduced.
The sonde consists of a ten-foot length plastic
pipe somewhat less than two inches in diameter (Fig-
ure 9). Buoyancy is provided by a pair of floatation
chambers so attached that when properly ballasted,
the antenna position of the pipe projects approximate-
ly 39 inches above a still sea datum plane.1
All necessary electronics are carried inside the
pipe below the water line. Power is provided by four
standard D-sized flashlight cells. A light-sensing
phototransistor shuts down the system during the hours
of darkness and conserves battery life. Thus, mea-
surements of ocean drift currents are made only during
daylight hours, more for system operator convenience
than necessity. The sonde's operating life is select-
able from one to almost six weeks through control of
the transmitter power. The four units used in this
test series were programmed for from ten to twelve
days of operation. Range is not particularly sensi"
tive to planned operating life.
Current sensing is achieved by the use of a cur-
rent-trap or biplane which is suspended at a control-
lable depth beneath the sonde hull or pipe. The cur-
rent intercept area Is from 4 to 16 square feet and
is isotropic. A ballast weight of approximately 12
pounds is attached to the bottom of the current trap
by about an eight-foot long slack line. The ballast
provides a righting moment about the system's meta-
center resisting heelover under strong wind condi-
tions and rights the sonde if it is upset by the
actions of large waves.
Position-finding is normally accomplished by the
use of triangulation from two (or more) shore radio
direction-finding stations using the surface-wave or
ground-wave signals. The ground-wave triangulation
routine provides ample accuracy of about +0.5 de-
gree, This translates to errors of less than half-
mile in position coordinates at stand-off ranges of
50 miles from the direction-finding baseline. Moored
sondes at fixed positions have been used to verify
these accuracies.
In a test of severe weather survivability, sev-
eral sondes were released at a waste disposal site
about 40 miles off the Delaware-Maryland coast in
January 1975. For a period of ten days, tracking
was conducted while severe winter sea conditions pre-
vailed in the test site area. Significant wave
heights on several occasions exceeded 25 feet and
averaged 10 to 15 feet during the tests. Air tempera-
tures ranged from the low twenties to the mid-forties.
Sea temperatures ranged from 37 to 42 degrees. This
data was derived ashore and from the weather ship HOTEL.
Ground truth at point of insertion was derived from
Loran-A/Loran-C. Wind gusts exceeded 50 knots over
substantial portions of three days in the period.
The drogues produced tracks as shown in Figures 10
and 11. At one point the drogues had moved 88 miles
off the Delaware coast. Only two fixes per day were
made during a 10-hour operating day, hence these
tracks do not show all the sondes' motions, since
night operations were not conducted.
Since the buoy portion of the sonde is nearly
awash and has only a thin radio antenna protruding
above the water surface, wind drag on the sondes was
not significant. A ratio in excess of 20:1 was main-
tained between the projected areas of the buoy ex-
posed to surface currents and of the current trap ex-
posed to currents at its depth. Carefull hydro-
dynamic design of the buoy hull structure further re-
duced the effect of surface currents.
3
16-2
-------
Compared to the $5,000 average cost of a current
meter, each radio-tracked drogue costs about $125
and each aircraft-tracked drogue about $15. The low
cost makes the drogues expendable, i.e., usually it
is less expensive to leave them in the water, than to
recover them. Each radio direction-finding station
onshore costs about $1,600.
CONCLUSIONS
Satellites, such as ERTS-1, can be used to obtain
a synoptic view of current circulation over large
coastal areas. Since in turbid coastal regions sus-
pended sediment acts as a natural tracer, cost is min-
imized by eliminating the need for expensive injec-
tions of large volumes of dye such as Rhodamine-B.
One of the principal shortcomings of satellite imag-
ing of coastal currents has been its inability to de-
termine current magnitude and to penetrate beyond the
upper few meters of the water column. These objec-
tions have been overcome by complementing satellite
observations with drogues tracking currents at various
selected depths. By combining the satellite's wide
coverage with aircraft or shore stations capable of
tracking expendable drogues, a cost-effective, in-
tegrated system has been devised for monitoring cur-
rents over large areas, various depths and under
severe environmental conditions.
9. J. R. Eliason and H. P. Foote, "Surface Water
Movement Studies Utilizing a Tracer Dye Imaging
System", Proc. of Seventh Int. Symp. on Remote
Sensing of Environment, Ann Arbor, Michigan, 1971,
pp. 731-749.
10. V. Klemas, D. Maurer, W. Leatham, P. Kinner, and
W. Treasure, "Dye and Drogue Studies of Spoil
Disposal and Oil Dispersion", Journal of Water
Pollution Control Federation, Vol. 46, No. 8,
August, 1974, pp. 2026-2034.
11. V. Klemas, G. Tornatore and W. Whelan, "A New
Current Drogue for Monitoring Shelf Circulation",
Abstracts of 56th Annual Meeting of American
Geophysical Union, Washington, D. C., June 16-19,
1975.
ACKNOWLEDGEMENTS
This project is partly funded through NASA-ERTS-1,
Contract No. NAS5-21937; and ONR Geography Programs,
Contract Programs, Contract No. N00014-69-A0407.
REFERENCES
1. P. P. Niiler and C. N. K. Mooers, "Programs in
Shelf Dynamics", Report on the Conference on
Physical Oceanography of the Continental Shelves,
Annapolis, Md., April 3-6, 1974.
2. V. Klemas, D. Bartlett, W. Philpot, and R. Rogers,
"Coastal and Estuarine Studies with ERTS-1 and
Skylab", Remote Sensing of Environment, Vol. 3,
No. 3, 1974, pp. 153-174.
3. R. L. Mairs, and D. K. Clark, "Remote Sensing of
Estuarine Circulation Dynamics", Photogrammetrlc
Engineering. Vol. 39, No. 9, 1973, pp. 927-938.
4. G. A. Maul, "Applications of ERTS Data to Oceano-
graphy and Marine Environment", Proc. of COSPAR
Symp. on Earth Survey Problems, Akademie-Verlag,
Berlin, 1974, pp. 335-347.
5. U. S. Department of Commerce, Tidal Current
Charts-Delaware Bay River, Environmental Science
Services Administration, Coast and Geodetic Sur-
vey, Second Edition, 1960.
6. U. S. Department of Commerce, Tidal Current
Tables, Atlantic Coast of North America, National
Oceanic and Atmospheric Administration, National
Ocean Survey, 1972 and 1973.
7. K. H. Szekielda, S. Kupferman, V. Klemas, and
D. F. Polls, "Element Enrichment in Organic Films
and Foam Associated with Aquatic Frontal Systems",
J. Geophys. Res. 77, No. 27, September 20, 1972,
pp. 5278-5282.
8. E. C. Monahan and E. A. Monahan, "Trends in
Drogue Design", Limnology and Oceanography, Vol.
18, No. 6, November, 1973, pp. 981-985.
4
16-2
-------
I 0 Cuutni (¦blit. Atlantic
< \\
^ N V
f WWN v % (
>0 \ *
A>i \
V 1 3
V \ v
\ N /
\ I) k\ i. a w a n k 0 4
* "V""^ -V-
I WO HOURS AFTER MAXIMUM FLOOD AT DELAWARE BAY ENTRANCE
c i Predicted tidal currents and ERTS-1 MSS band 5 image of Delaware
F,GU Bay obtained o» October 10, 1972 O.D. Nos. 1079-15133).
Jkj:
Md s^"
f'
!&®iKr
TWO HOURS BEFORE MAXIMUM FLOOO AT 0CLAWARE BAY ENTRANCE
Figure 2: Predicted tidal currents and ERTS-1 MSS band 5 image of Delaware
Bay obtained January 26, 1973 (I.D. Nos. 1187-15140). ARE
16-2
-
-------
Figure 3: Predicted tidal currents and EPTS-1 MSS band 5 image of Delaware
Bay taken on February 13, 1973 (I.D. No, 1205-151^1).
ONf HOUR AHFR MAXIMUM I UK Al OHn.v.Wj |,m (NIHAN.i
Figure 5,
A 300-m diam P.hodamine WT dye patch being
spread by R. V, DELAWARE,
Figure 6.
Dye patch caught in a boundary and carried
TOWARD THE BAY.
-------
¦ firldgelon
Dover •
KEY
Concentrations from non-
linear regression, mg/liter
m "
[TTT] unclassified
| | unclassified
Figure 4 SEDIMENT DISTRIBUTION MAP
Figure 7. Aquatic boundary between two water masses of
DIFFERENT PROPERTIES OBSERVED FROM AN ALTITUDE
OF 3,0CG M.
7 16-2
-------
Fit 8 lOSOhrt
?w«"
ft.* VAftTlCLD
_ •oummmv
MJUWTHtL »IH
i»v
Figurl 8. Oil slicks at C93C hours.
TA"
M PI
•wV
PI 4im
..ICI MCAM
VffiWv,LT
FIO. 9 I500hrt
(••0,000
Figure 9, Oil slicks at 1500 hours.
8
16-2
-------
¦ aaiiBsaHaiBBE' S'jsr*?? -At
•' ?^rr
*r ;» ¦ t *.'$*• %J<-.
fCxQ '
L.^ | «KM|)
HI '
v* .r»,
***- ¦MHHsMSS .. --V
J-- Sgga :. ,gffi -
sjss&s* Sight,
• - rsz-z^pi'-*-
sj,~-* i*1:'•¦¦¦?¦
" ,^- -r •
r., ¦k^Jt£
m^s§Msi
w SMS
-
**~§ '•• * * *,, -x^r"W ;V:
• ¦ •?£ii&Bry--C; ' .v •• *5%- 4 V- *s . '" .i V-tffa "'=»«r-ji.' t *. *5ftv'-~w«r*-
Figure 10, Enlarged digital enhancement of acid waste plume imaged
by ERTS-1 on January 25, 1973,
STYROFOAM 4
1
,3/8 THREADED STEEL ROD
B*~W
111/4H PAINTED PLYWOOD
/
V.
DYE PACKET g) WASHER
K.
WOOD SPIDER
SWIVEL SNAP HOOK
-ADJUSTABLE NYLON ROPE
GALVANIZED STEEL
BIPLANE
DYE PACKET
FIGURE II * DROGUE AND DYE EXPERIMENTAL PACKAGE
HF Antenna
Figure 12
DEEP CURRENT DROGUE
WITH LOW PARASITIC
DRAG (Model 3)
4 Ft
61/4 Ft
Electronics
/
Flotation
Chamber
Low-drag
Line
Variable, to 300 Ft
H
Current Trap (Biplane)
Ballast
16-2
-------
Delaware Bay
Can* May. N.U
10 IS 20 25 30
NAUTICAL MILES
•. Cap* hkmoptr. Dn
'q
Indian River Inlet
Ocean City Inlet
Depth Contours In Fathoms
,M0
JO
75°00' W
1030 - 28
1800-21 1620-29
Insertion
" 0910-29
*r\llOO-3g / .
& X rJ 0930" 1100-
1630-23
TOOffi^ ^ <3 1630-23
1200-24 1100*25 110^23
093* ,100-26 q ^.26
1500-31
38° 30'
TRACKING STOPPED
1030-22
40
/' 50''
* '
•: \ \
\ 1
: §
\
100
FIGURE 13.
Trace of Sea-Sonde * 005
Jan. 21 through Jon. 3t, 1975
Current Trap Depth 30 ft
74° 30'
/ /
S/' \-y
'' / ! °
/'woo'w \
/
38°00'N
73° 30'
100
7
Cape May, N.J.
//!
Cape Hen/open, Del.
\\
\\
w
\ \
—; r
0 1645-24
Delaware
Maryland
1630-23
?A
Ocean City Inlet
10 15
20
NAUTICAL MILES
Depth Contours in Fathoms
0 D-F Site ^Average Wind Vector
39° 00' N
/ * / / /
'/ / / / „
/ / /.''r
38° 30
10.SO ,'6
1645*26
1120-27
04', c'6
It.' >0-?8
/
j :
;> \
TRACK DROPPEO
1500-29 l<
040-2^'
40
S
.9
" j N. Heyes Canyon
FIGURE 14.
Trace of Sea-Sonde No 008
Jon. 21 through Jan 29 1975
Current Trap Depth 10 ft.
» )
) S. Heyes, Canyon
,1000^.
\ S. Vries Canyon
! \->
75°bo'W
74° 30'
74° 00'
73° 30'
10
16-2
-------
EVALUATION OF WATER SAMPLES COLLECTED DURING LANDSAT-1 OVERPASSES
OF THE LOWER CHESAPEAKE BAY AREA
D. E. Bowker and W. G. Witte
NASA Langley Research Center
Hampton, Virginia 23665
Abstract
Water samples were collected on 18 days when the
LANDSAT-1 satellite was passing over the lower Chesa-
peake Bay area. A correlation between the various
water parameters has been performed for the more than
300 surface samples. Six days were sufficiently cloud
free that MSS digital tapes were useful for analysis.
Correlation of radiance values with the water param-
eters revealed a low correlation for chlorophyll and
good correlations with particles and sediment. The
relation of total particles to sediment was linear but
varied from day to day.
attenuation coefficient of the water during each
cruise. The locations of the sampling stations are
shown in Figure 2; all of the stations were not visited
each day, however.
Introduction
L-10299 1
I6-3
Figure 2. Location of sampling stations.
Table 1 lists the LANDSAT-1 overpass dates, the
number of samples collected, and the availabiliiy of
MSS tapes. Three of the LANDSAT-1 overpasses (Dec. 4,
Jan. 9, and July 26) were on days following the lower
Bay overpasses. Only the James River was visible on
these days. The number of helicopter samples varied
from 5 to 33 for each day, whereas the number of ship
samples varied from 2 to 9. This was often controlled
by local weather conditions. Two of the six Images
for which MSS tapes were analyzed were partially
degraded by cirrus clouds or ground fog. Fortunately,
both helicopter and ship samples were available on
3 of the 4 other days.
TABLE 1. RECORD OF SAMPLE COLLECTION AND ANALYSIS OF
MSS TAPES
LANDSAT-1 OVERPASS HELICOPTER
DATE SAMPLES
OCT. 10, 1972 18
DEC. 3.1972 18
DEC. 4,1972 14
JAN. 9, 1973 6
JAN. 26. 1973 33
FEB. 13, 1973 33
MAR. 3, 1973
APR. 8. 1973
MAY 14. 1973
JUNE 1. 1973
JUNE 19, 1973
JULY 7, 1973 7
JULY 25,1973 20
JULY 26, 1973 5
AUG. 30, 1973 13
SEPT. 17,1973 18
OCT. 5, 1973 18
OCT. 23, 1973 15
18 DAYS 218
SHIP LANDSAT-1 TAPES
SAMPLES ANALYZED
^ X
6 X
7
6
2
7
6
9 X (HAZY)
X (FOG)
The Langley Research Center and the Old Dominion
University have been engaged in an investigation of the
lower Chesapeake Bay area since October 1972. During
the first year's effort water samples were collected
every 18 days, weather permitting, when the LANDSAT-1
spacecraft was passing over the area. Figure 1 is a
band 6 image of the lower Bay area on October 5, 1973.
The James River enters the Bay from the lower left
region. MSS digital tapes were available for analysis
for six of 22 LANDSAT-1 passes. This report presents
the analysis of the water parameters monitored during
the investigation and the correlation between the water
parameters and the MSS radiance values.
Figure 1. LANDSAT-1 MSS band 6 image of lower Chesa-
peake Bay area, October 5, 1973,
Water samples were collected by helicopters and
by ship. Helicopter samples were primarily analyzed
for particle size, distribution, and chlorophylls.
Ship samples were primarily analyzed for total suspen-
sate weight, temperature and salinity; a transmissom-
eter on board the ship continuously monitored the
SAMPLES WERE
... ..u.,u ,„KEN ALONG THE JAMES RIVER BRIDGE
AND THE CHESAPEAKE BAY BRIDGE-TUNNEL ON 26 JANUARY 1973
AND 13 FEBRUARY 1973
OHELICOPTER SAMPLES
OSHIP SAMPLES
-------
Analysis of Water Data
The correlation between the parameters monitored
during the analysis of the water samples has been cal-
culated for the helicopter and ship data on a daily
basis and for the total. The correlation coefficients
are given in Tables 2 and 3, respectively, for the
total data. The correlation between the parameters
for the helicopter samples reveals several interesting
points. The correlation of total chlorophyll with
chlorophylls A, B, C, and carotenoids, columns 9 to 12
in the last row, is excellent (0.87 - 1.00). Thus,
chlorophyll A, B, or C could be measured and the total
chlorophyll data derived. The correlation of particles
with the chlorophyll data, columns 2 to 8 and rows 9 to
13, is low (0.02 - 0.22). This indicates that plankton
is not a significant contribution to the particle data
and is not associated with any specific particle size
environment.
TABLE 2.
CORRELATION COEFFICIENTS FOR HELICOPTER
SAMPLE VARIABLES
NO. VARIABLE
1
2 3 4 5 6 7 8
9 10 11 12 13
1 DEPTH
LOO
2 PART0-.5
-.13
1.00
3 PART .5-1
-.01
.73 1.00
4 PART 1-2
-.04
.65 .87 1.00
5 PART 2-4
-.14
.66 . 51 .69 1.00
6 PART 4-8
-.12
.57 . 36 . 53 . 92 1.00
7 PART 8-16
-.14
.31 .10 . 26 . 75 .81 1.00
8 PART TOT
-.09
.77 . 86 . 94 . 85 . 74 . 51 1.00,
9 CHLQR A
.10
'.10 l6~irT06~.0T"ir .14
1.00
10 CH10R B
.23
.10 . 21 .20 .03 ,02 ,04 .15
.91 1.00
11 CHL0R C
.20
.10 . 22 .21 ,05 . 03 .04 .17
.88 .98 1.00
12 CAR0T
.21
.09 .17 .15 . 02 . 03 . 05 .13
.83 .86 .86 1.00
13 CHL0R TOT
Jl
jo .a__.05 . i^
111 .99 1.00 .87 1.001
TABLE 3. CORRELATION COEFFICIENTS FOR SHIP
SAMPLE VARIABLES
NO.
VARIABLE
1
2
3
4
1
TIME
LOO
2
DEPTH
-.04
1.00
3
TEMP
-.19
.03
1.00
4
SALINITY
-.32
.05
.37
1.00
5
ATTEN COEF
.23
.07
-.41
-.69
6
SEDIMENT
.12
-.03
.03
-.39
.67 1.00
The correlation of total particles with each
particle size range (0.51 to 0.94) suggests that total
particles are fairly tepresentative of any given size
range. The average particle size histogram for each
day peaked in the 1- to 2-urn size range (column 4) 50%
of the time, and was equally divided between the next
lower and higher size ranges the other days.
The last point to be made about Table 2 is the
correlation of depth with each parameter. Although
the correlation coefficients are low, there is a nega-
tive correlation of depth with particle count and a
positive correlation of depth with chlorophyll; that
is, when depth increases chlorophyll increases and
particle count decreases. On a daily basis, the
correlations with total particles varied from -0.03 to
-0,61 and with total chlorophyll from -0.46 to 0.38.
The positive correlations with total chlorophyll gener-
ally occurred on days when the chlorophyll level was
high.
The correlation between variables for the ship
data, Table 3, indicates a negligible correlation
(-0.03) between sediment and depth, consistent with
the correlation between total particles and depth.
However, the sediment versus depth correlation for ship
data varied from -0.77 to 0.73 on a daily basis. One
possible reason for the large variation is that the
ship stations were all located in relatively deep water
(average depth 30 feet) whereas the depth at the heli-
copter stations varied from 2 feet to 73 feet.
The attenuation coefficient had a correlation of
0.67 with sediment for a combined total of 13 days.
The linear regression curves for each day's data, with
the exception of one, were nearly parallel. With the
removal of that one day's data from the total, the
correlation improved to 0,80. Thus the attenuation
coefficient was useful for checking the sediment data
for anomalous values.
Since the ship and helicopter did not occupy the
same stations, it was not possible to establish a one-
to-one relationship between sediment and particle
count. An average sediment and particle distribution
was determined for each day, however, and a comparison
was made for the 8 days when both sets of data were
available. The correlation between average sediment
and average total particles was 0.31. When total
particles were converted to total volume, by assuming
spherical shapes with a mean diameter equal to the mid
value of each particle size range, the correlation
improved to 0.40. The most meaningful relationship was
the correlation of sediment with the average histogram
mean (0.73). These data are plotted in Figure 3. It
can be seen that an increase in sediment indicates an
increase in the average particle size, not simply an
increase in particle number. A more direct relation
between sediment and particles will be established
later by using the radiance values at each station.
14
AVERAGE
SEDIMENT,
mg/ liter
REGRESSION EQUATION FOR LINE
AVG SED ¦ -5.93 + 4.0) AVG HIST MEAN
R - 0.73
.5124
AVERAGE HISTOGRAM MEAN, m
Figure 3. Average sediment versus
average histogram mean.
! x 10
,-<5
2
16-3
-------
Correlation ot MSS Radiances With Water Parameters
MSS radiance values for each sampling station were
extracted from the computer compatible tapes by locating
the station coordinates on a CCT generated gray level
map. An average of 48 pixels (six scan lines by eight
pixels per line) centered around the coordinate were
taken to minimize the effect of banding, caused by
unequal response of the six detectors for each band.
In the combined analysis of several days' data,
the radiance values were multiplied by 1/cos 0 to
correct for varying solar illumination.
Chlorophyll
The correlation coefficients for total chlorophyll
and total particles with each MSS band are given in
Tables k and 5, respectively, for the 6 days when MSS
tapes were available- The highest correlations
occurred on January 26. However, the correlations on
February 13 are all negative, indicating a significant
change from the January 26 date in the relationship
between chlorophyll and radiance. It has been shown*
that the variation in radiance values is due primarily
to total particles (or sediment), making the daily
correlations with total chlorophyll somewhat mislead-
ing. The linear regression equations given for the
four clear days indicate that the slope of the regres-
sion line varies over a wide range. When the data are
combined for a total analysis, the correlations are
greatly reduced, as expected.
TABLE 4. CORRELATION COEFFICIENTS AND LINEAR
REGRESSION EQUATIONS FOR TOTAL CHLOROPHYLL
6 A NO
OCT. 10
JAN. 26
FEB. 13
JULY 7 AUG. 30
OCT. 5
TOTAL
4
.37
.85
-.70
.80 .02
-.06
.12
5
.36
.84
-.68
.72 -.26
-.04
.24
6
.42
.76
-.61
.65 -.17
-.08
.33
7
.45
.10
-.61
.67 -.08
-.02
.16
LINEAR REGRESSION EQUATIONS
OCT. 10 BAND 4 = 0.469 + 3.01 x 10"4 (CHLOROPHYLL)
JAN. 26 BAND 4 = 0.295 + 4.53x 10"3 (CHLOROPHYLL!
FEB. 13 BAND 4 = 0.494 - 3.5? x 10"3 (CHLOROPHYLL!
OCT. 5 BAND 4 = 0.511 - 1.49 x if)"3 (CHLOROPHYLLI
Particles
A combination of bands (5-6) gives the most con-
sistent correlation with particles (0.49 to 0.87).
However,, the linear regression equations show a large
change in slope from day to day. Combining the data
results in a Low correlation, just as with chlorophyll.
A plot of batids <5-6) radiance versus total particles is
presented in Figure 4 for the October 5 data. A simi-
lar plot for January 26 and February 13 can be found
in Reference 1.
TABLE 5. CORRELATION COEFFICIENTS AND LINEAR
REGRESSION EQUATIONS FOR TOTAL PARTICLES
BAND
OCT. 10
JAN. 26
FEB. 13
JULY 7
AUG. 30
OCT. 5
TOTAL
4
.53
.73
.88
.50
.45
•31 ,
.15
5
.56
.74
.88
.57
.52
.38
.34
6
.53
.68
.85
.61
.42
.13
.43
7
.51
.13
.77
.57
.37
-.04
.14
5"6
.58
.74
.87
.49
.55
.58
.22
LINEAR REGRESSION EQUATIONS
OCT. 10 BANDS (5-61 = 0.119 + L03* Ifl"10 (TOTAL PARTICLES]
m. 26 BMiOS <5-6! = 0.058 + 3.54* l(f10 (TOTAL PARTICLES!
FEB. 13 BANDS (5-6) = 0.138 + 8.14 x 10"10 (TOTAL PARTICLES)
OCT. 5 BANDS (5-6) = 0.113 + 1.02X N?"10 (TOTAL PARTICLES)
REGRESSION EQUATION FOR LINE
BAND (5-6! • 0.U3 + 1.Q2E - 10 (PARTICLES!
R ¦ 0.58
BAND (5-6),
1
mtv/cm sr
.10 H
TOTAL PARTICLES/liter
Figure 4. Bands (5-6) radiance versus
total particles far October 5.
Sediment
The correlation between sediment and the various
MSS bands is presented in Table 6. The data for July 7
are not presented because the stations were visibly
covered by cirrus clouds which resulted in a poor
correlation. The correlation with band 5 is the most
consistent (0.78 to 0.93). Even when the data are
combined, the correlation is reasonably good (0.72).
It is possible to improve the total correlation,
however, by making an adjustment in the radiance levels
for October 5. labia 7 gives the gray scale histogram
maximum for each MSS band for all 6 days. Only the
water pixels in the lower Bay area are included (band 7
gray scale £ 3) and the gray levels have been corrected
for Sun angle by multiplying by 1/cos 6. Since many of
the water pixels have been eliminated, such aa those
displaying clouds or fog, the gray scale maximums for
August 30 are not too different from October 5. It is
3
16-3
-------
TABLE 6. CORRELATION COEFFICIENTS AND LINEAR
REGRESSION EQUATIONS FOR SEDIMENT
BAND
JAN 26
FEB 13
OCT 5
TOTAL
4
.93
.59
.85
.80
5
.86
.93
.78
.72
6
.79
.75
.45
-.08
7
-.11
.47
.85
.22
JAN 26 BAND 5 ¦ 0.080 + (X011 (SEDIMENT)
FEB 13 BAND 5» 0.142 + 0.0035 (SEDIMENT)
OCT 5 BAND 5 ¦ 0.202 + 0.0013 (SEDIMENT)
TABLE 7. CORRECTED GRAY SCALE HISTOGRAM MAXIMUMS
Discussion
The analysis of the chlorophyll and radiance data .
Indicates that chlorophyll at less than bloom condi-
tions has, at best, only a minor influence on the
radiance values. The variations in radiance are pri-
marily caused by particles (or sediment). This point
can be made more evident by examining the errors in the
radiance values associated with the linear regression
of radiance with particles. Figures 6 and 7 show a
plot of the differences between the observed and pre-
dicted radiances for bands (5-6) as a function of the
chlorophyll A values at each station for October 10 and
January 26, respectively. These 2 days had a good
spread in chlorophyll values. It is seen that the
errors appear to be randomly distributed with positive
and negative values; thus, there is no proof that
chlorophyll is influencing radiance for this band
combination.
BAND
OCT. 10
JAN. 26
FEB. 13
JULY 7
AUG. 30
OCT. 5
.04
4
33
35
39
31
36
37
RA0IANCE ERROR,
5
17
15
19
16
20
21
rrW cm2 sr 0
6
8
19
8
9
13
11
-.04
7
2
1
2
2
3
2
noticed that the gray scale maximums for band 5 on
January 26, February 13, and October 5 are 15, 19, and
21, respectively. The differences between January 26
and February 13 are presumably due to the average sedi-
(nent difference (highest on Feb. 13). However, the
sediment level on October 5 was similar to January 26,
thus the gray levels should be about the same. Reduc-
ing the average radiance level on October 5 by six gray
values increases the total correlation between sediment
and band 5 to 0.92. The increased radiance on October 5
Is probably due to the atmosphere, since there are a
large number of cirrus clouds to the northwest. It is
possible that sea state could be influencing the data
also.
A plot of the combined sediment versus band 5 is
given in Figure 5.
.40 r
-.12
O O
o
CHLOROPHYLL A, mgl m
Figure 6. Particle radiance error versus
chlorophyll for January 26.
.12
.04
RADIANCE ERROR,
2
mw! cm sr
.36
.32
BAND 5, mW/cm? sr -28
.24
.20
.16
-.04
-.1?
O
O
o
o
o
-o-
o
o ° o
o
o
o
REGRESSION EQUATION FOR LINE
BAND 5 ¦ 0.210 + 0.0139 (SEDIMENT)
.6 R . 0.92
2 i 6 8 10 12 14
SEDIMENT, mg/liter
8 12 16 2
CHLOROPHYLL A, m9/ m
24
28
Figure 5. Band 5 versus sediment for January 26,
February 13, and October 5.
Figure 7. Particle radiance error versus
chlorophyll for October 10.
The correlations of total particles with bands
(5-6) were reasonably consistent (0,58 to 0.87) for the
4 clear days. Particle counts generally tend to show
considerable scatter, even when the samples are taken
from the same station. Under these conditions, it is .
felt that the correlation with radiance values are very
good.
4
16-3
-------
Whereas the slope of the linear regression equa-
tions for particles versus bands (5-6) varied by a
factor of 8, the slope for sediment versus band 5 only
varied by a factor of 3. Combining the three sets of
sediment data, with an appropriate gray scale correc-
tion for October 5, did not degrade the correlation.
It is clear, then, that sediment (total suspensate
weight) is the most reliable water parameter for
remote sensing in the lower Chesapeake Bay area.
It was found that the average particle size
increased when the sediment level increased, but the
relationship between sediment and total particles was
not firmly established. The linear relation between
the two parameters and radiance can be used to make a
direct comparison. Using the appropriate regression
equation, total particle count was predicted at the
ship stations and sediment was predicted at the heli-
copter stations. The equations for the linear relation
between particles and sediment under those conditions
are given in Table 8 for the 3 days when helicopter,
ship, and radiance data were available. The equations
are shown plotted in Figure 8 along with the data for
February 13. It is apparent that an increase in sedi-
ment between days does not necessarily indicate an
increase in particle count. Actually, there was a
significant decrease in total particles on February 13
when the sediment level was relatively high; the
increased particle size was the more important param-
eter on this day.
TABLE 8. LINEAR REGRESSION EQUATIONS FOR
SEDIMENT VERSUS TOTAL PARTICLES FOR GIVEN DATES
JAN 26 SEDIMENT ¦ 0.949 +1.211 x lo'7 (PARTI CLES) R - 0.88
FEB 13 SEDIMENT - 2.001 +4.511 x 10'7 (PARTICLES) R ¦ 0,87
OCT 5 SEDIMENT * 0.507 + 2,084 x JO"7 (PARTICLES) R - 0.45
SEDIMENT, mg/liter
O HELICOPTER DATA
9 O D A SHIP DATA
FEB.13,1973
RB0.87
JAN. 26,1973
R-0.8
0Cr,5,1973
R-0.45
-L
4 6 8 10
TOTAL PARTICLES, liter
11 14x10
Figure 8. Linear regression plots for sediment versus
particles for given dates including data for February 13.
Concluding Remarks
An analysis of the radiance values from the MSS
digital tapes has revealed a low correlation with
chlorophyll and a very good correlation with total par-
ticles and sediment. The most consistent correlations
were established with sediment and, therefore, this
variable is considered to be the most important water
parameter for remote sensing surveys with LANDSAT-1.
A linear relationship was established between sedi-
ment and particles, but the relation varied widely from
day to day. Thus, it was not possible to infer the
sediment level from the total particle count without a
calibration curve, k further analysis of the environ-
mental parameters will be essential if a better under-
standing of the daily variations in the water quality
is sought.
Reference
"*"Bowker, D. E., and Witte, W. G,, "Evaluation of
ERTS MSS Digital Data for Monitoring Water in the Lower
Chesapeake Bay Area," Proceedings of the Fourth Annual
Remote Sensing of Earth Resources Conference.
March 24-26, 1975. To be published.
5
16-3
-------
LONG PATH INFRARED AMBIENT AND LABORATORY MEASUREMENTS,
CHEMICAL REACTIONS AND FREON STUDIES
B.W. Gay, Jr., R.C. Noonan, J.J. Bufalini, P.L. Hanat
Environmental Protection Agency
Atmospheric Chemistry and Physics Division
Environmental Sciences Research Laboratory
Research Triangle Park, N.C. 27711
Abstract
Long path infrared (LPIR) sensing methods are
used in ambient air measurements and laboratory
studies of atmospheric pollutants. Early studies
of photochemical smog using infrared techniques
led to the discovery of peroxyacetyl nitrate (PAN),
an important pollutant formed in the atmosphere,
and showed ozone to be the major oxidant. More
recently, LPIR studies of the Los Angeles atmos-
phere showed that PAN and formic acid are major
products of the well-reacted polluted air mass.
Vinyl chloride was measured three years ago by
LPIR in an air sample from, the Houston, industrial
area. LPIR is being used in the laboratory to
measure reactants and products,to study the mech-
anisms and kinetics of dark reactions such as
ozonolysis reactions, photolysis reactions and
photooxidation reactions. With the aid of cold-
trapping techniques, LPIR is being used to measure
Freons and halogenated hydrocarbons that exist
in the atmosphere at concentrations down to 50
parts-per-trillion. In the laboratory Freon and
halogenated hydrocarbon reactions are being in-
vestigated with respect to possible depletion of
the earth's upper ozone layer.
Introduction
Infrared absorption spectra exhibited by most
compounds have long been a valuable tool in the
identification of organic and inorganic compounds.
Infrared spectroscopy has been used with great
success in the study of the polluted atmosphere in
ambient and laboratory conditions. The low concen-
trations of ambient pollutants necessitates the
use of long optical absorption path lengths to
achieve low limits of sensitivity. Very long
optical path lengths can be obtained by using the
sun as a source when it is low on the horizon. This
gives a long slanted path through the atmosphere,
but atmospheric water vapor and carbon dioxide
(completing absorbing certain regions of the in-
frared spectrum) greatly reducing the information
that might otherwise be obtained. Studies in the
upper atmosphere using aircraft or balloons eliminate
much of the water vapor and carbon dioxide inter-
ference but also eliminate most of the chemical
pollutants of the lower troposphere. It is in the
lower troposphere that most pollutants of interest
reside and react photochemically.
In. the study of pollutants in the lower atmos-
phere, a long folded optical path through an
enclosed system haa several advantages over a path
through the atmosphere using the sun as a source.
One advantage is tha,t the infrared signal does not
pass through any unnecessary unpolluted air which
would contribute to water vapor and carbon dioxide
absorption. The folded path can be optimized for
greatest sensitivity and this length can be precisely
known and used for quantitative measurements. The
partial pressures of pollutants as well as the total
cell pressure can be varied to enhance the detecta-
bllity of certain pollutants. And the cell itself,
if made of proper material, can be used as a photo-
chemical reactor to study the transformation of
pollutants under the Influence of ultraviolet light.
Using a folded optical path system enclosed in
a cell, peroxyacetyl nitrate (PAN) was discovered by
its distinct infrared absorption bands when auto
exhaust was irradiated by early Investigators of
photochemical smog.^ PAN, a phytotoxicant and eye
irritant, is formed in the urban atmosphere during
photochemical smog formation. Ozone was also shown
by these investigators to be a major oxidant in
photochemical smog. Present-day instrumentation has
extended infrared detection to a range of pollutant
concentrations about 100 times lower than was obtained
by these earlier experimenters.
The Infrared System
A rapid-scan Fourier Transform spectrometer
using a Michelson interferometer and cryogenically
cooled detectors covers the spectral range of 700
to 3400 cm . A cylindrical glass chamber 10 meters
in length and 0.31 meters in diameter houses an
eight mirror optical system. Around the cell are
light banks which provide the ultraviolet energy
necessary for photochemical work. A more detailed
discussion of the optical system, interferometer and
2
spectral data techniques are given elsewhere.
Measurements of Ambient Pollutants
Ambient air samples can be collected in large
inert plastic bags and returned to the laboratory
for analysis but reactive and unstable compounds are
lost during transport. This method was used to
investigate the hydrocarbons in the Houston area,
and one such sample contained 37 pphm vinyl chloride.
Class sample vessels have also been used with some
success in transporting air samples.
By placing the instrument in the immediate area
of the polluted air mass, air samples can be drawn
directly into the absorption cell and measurements can
be obtained for reactive and unstable compounds.
Measurements of this type have been made on top of
a building at Pasadena, California during days of
3
low and high photochemical smog formation. On non-
smoggy days of good visibility, morning and evening
samples showed an. accumulation of pollutants such as
carbon monoxide, methane, ethylene, acetylene, and
non-methane paraffinic hydrocarbons. A lack of photo-
chemical products such as ozone, formic acid and PAN
was also observed.
During days of photochemical smog formation,
ozone, PAN and formic acid concentrations Increased
up to about 2:00 pm after which they decreased. These
LPIR measurements indicate that in the well reacted
air mass during high photochemical activity the major
photochemically produced carboa-coivtalTiing substances
are PAN and formic acid.
1
16-5
-------
Measurements In the Laboratory
In the laboratory LPIR systems are used to study
numerous photochemical reactions as well as dark
reactions. In irradiations of hydrocarbons and oxides
of nitrogen (typical auto exhaust reactants) reaction
products and rates are obtained directly from the
spectral data. Data obtained are used to determine
reaction mechanisms and kinetic rates. Products,
mechanisms and rates of dark reactions (ozonolysis
reactions) have been studied for various ozone-olefin
systems. The photolysis of alkylnitrites and benzyl
nitrite have also been investigated, and the methyl
nitrite system was computer-modeled. A study of the
photooxidation of chlorinated ethylenes has been
conducted and computer modeling of these systems is
underway. Recently a new method of using chlorine
photolysis to create compounds of the PAN family was
developed with the laboratory long path infrared
4
system.
Freon-Halocarbon Field and Laboratory Studies
With the present concern for ozone depletion
in the upper atmosphere due to Freons and other
halogenated compounds, LPIR studies are being con-
ducted in the laboratory and in the field to determine
the fate of these compounds and assess the threat
they pose to the stratospheric ozone layer. Freons
and halogenated hydrocarbons are being measured in
the field to determine background concentrations.
Cryogenic concentration procedures are used to raise
the sensitivity of the infrared analysis.5 Labora-
tory studies are directed toward determining
reaction products, mechanisms and rates of these
reactions in the lower and upper atmosphere.
Other LPIR Studies
A LPIR chamber is being constructed which will
simulate stratospheric conditions so that photolysis
of Freons and other compounds and their effect on
ozone can be studied. A four kilometer optical
path system is also being constructed which will
be used for ambient smog studies of the Los Angeles
atmosphere.
Literature Cited
1. Stevens, E.R., Hanst, P.L., Doerr, R.C., Proc.
A.P.I. 37, (III), 171 (1957).
2. Hanst, P.L., Lefohn, A.S., Gay Jr., B.W.,
Applied Spec., vol. 27, #3, 188 (1973).
3. Hanst, P.L., Wilson, W.E., Patterson, R.K.,
Gay Jr., B.W., Chaney, L.W., Burton, S.C., EPA report
0 EPA-650/4-75-006, Feb. 1975.
4. Gay, B.W. Jr., Noonan, R.C., Bufalini, J.J., and
Hanst, P.L., "Photochemical Synthesis of Peroxyacyl
Nitrates in the Gas Phase via Chlorine-Aldehyde
Reaction," to be published in Environmental Science
and Technology.
5. Hanst, P.L., Spiller, L.L., and Watts, D.D.,
"Infrared Measurements of Halogenated Pollutants and
Other Atmospheric Trace Gases," submitted to Journal
of the Air Pollution Control Association.
2
16.5
-------
REMOTE SENSING OF TRACE CONSTITUENTS
FROM ATMOSPHERIC INFRARED EMISSION AND ABSORPTION SPECTRA
D. B. Barker, J.N. Brooks, A. Goldman, J.J. Kosters,
D. G. Murcray, F. H. Murc.ray, J. Van Allen and W, J. Williams
Department of Physics and Astronomy
University of Denver
Denver, Colorado 80210
Summary
Atmospheric infrared emission and absorption
spectra obtained from aircraft and balloon-borne spec-
trometers are presented. From such spectra, mixing
ratio vs altitude profiles are derived for several minor
constituents. Our recent results on HNO-j, CF2CI2,
CFClj and HF are presented. In addition the feasibil-
ity of infrared detection of other trace constituents
such as HC1, HF, NHj, NO and SO2 against the rest of
the atmospheric background is studied. From this
study, made on a line-by-line basis for "state of the
art" airborne spectrometers, potential spectral fea-
tures for detection of the trace constituents are isola-
ted.
Introduc tion
Infrared spectroscopy and radiometry are very
powerful techniques for remote sensing of trace atmo-
spheric gases. This is due to the unique, identifiable,
patterns in the fine structure of the molecular spectral
absorption lines and to the quantitative information de-
rivable from these absorptions. The stratospheric
concentrations of a number of these trace constituents
is currently of interest scientifically and environmen-
tally. Both the total column density and the altitude
density profile are of interest,along with the diurnal,
seasonal and latitudinal variation of these densities.
Measurements of the concentration of a constitu-
ent using infrared techniques requires a knowledge of
theoretical molecular modeling with substantiating
laboratory data,as well as knowledge of the atmospher-
ic infrared background the constituent is to be ob-
served against. Different radiometric and spectro-
scopic techniques may be optimum for particular con-
stituents, depending on the spectral contrast to their
environmental background. There are three interde-
pendent parameters which generally determine the
choice of technique and instrumentation. These are:
required spectral resolution, choice of optical path,
and nature of infrared radiance source.
The resolution required for absolute identifica-
tion of a constituent is often greater than that required
for quantitative monitoring of it. Also different con-
stituents may require significantly different resolutions
for identification depending on their individual spectral
structure, the contrast of this structure to the atmo-
spheric background and the optical path employed,
Usually the greater the resolution employed the lesser
the energy available at the detector for a given source,
causing the need for sensitive instruments in observing
weak sources.
For maximum sensitivity for trace constituents,
the optical path should contain a maximum number of
constituent molecules and a minimum number of opti-
cally interfering molecules. Since the major inter-
ference is generally from H2O lines ,this is achieved
with nearly horizontal paths in the stratosphere. How-
ever, there are many examples where horizontal paths
are not required, which avoids many experimental
1
difficulties. (For tropospheric paths the interference
from H2O lines may limit the path that can be used.)
Two infrared radiance sources have been used in
field measurements to date: the sun at about 5000K
and the thermal emission from the atmospheric mole-
cules themselves at about 220K, The development of
tunable lasers will provide an additional source for
future experiments. The sun is an intense source of
infrared radiation. However, in the near infrared it
has a large number of absorption lines due to its own
atmosphere. It also requires some form of heliostat
and can only be used for long, horizontal path studies
for 15 to 20 minutes twice a day. Because of the high
temperature of the sun it is convenient to use this
source for the initial high resolution identification of
the constituent. Time and altitude profile studies can
then often be better accomplished using the thermal
emission of the constituent as the radiometric source.
Our group at the University of Denver has been
using these techniques in remote sensing of atmospher-
ic trace constituents for a number of years. Data per-
tinent to these studies have been obtained with aircraft
and balloon-borne spectrometers covering the spectral
region from 1 to30jxm. The spectral resolution obtained
in the solar spectra and in the atmospheric emission
spectra is 0.3 cm"* and Js2 cm"' respectively.
The derivation of the vertical distribution of the
minor constituents is done by fitting the observed inte-
grated absorption or integrated radiance to the corre-
sponding theoretical integrals-.- Two methods are used:
the line-by-line method and the band model method.
The necessary parameters (frequencies, intensities,
halfwidths and energy levels for the line-by-line meth-
od and average intensities and average halfwidths for
the band model method) are available for a number of
atmospheric molecules of interest.
The instrumentation, methods of data analysis
and results of stratospheric mixing ratio profiles for
a number of molecules: H2O, CH^, ^O, CO, HNOj
and NO,, have been described previously*' ^ (where
detailed^ references we re also given). In this work
summaries of our recent results for HNO^, CF2CI2.
CFClj and HF are presented, as well as sample cal-
culations which can be used to determine the optimum
parameters for additional constituents.
HNO^ Mixing Ratios
The major atmospheric emission in the 11. 3^m
region in the lower stratosphere is due to HNO-j.
Typical spectra obtained in this region from various
altitudes are shown in Figs. 1 and 2 (both obtained
from Holloman AFB, New Mexico), The 11. 3)jro HNOj
band has been used to determine the HNO, mixing ratio
vs altitude profiles by using a sensitive filter radio-
meter with a filter passing radiation from 10. 6 to
11.6^m, The instrumentation and data analysis have
been described previously. Such emission data have
advantages over the spectral data in that they are ob-
tained more rapidly and that the instrument has greater
16-6
-------
Fig. 1
a 10 II 12 13 14 15
WAVELENGTH (microns)
Typical spectral atmospheric radiance mea-
sured on 10 Dec 1 969 at altitudes of 3. 4, 10. 0,j
20. 1 and 24. 7 km.
for all records.
The zenith angle was 45
10'
'4.
$_io'
LU
O
z
<
o
. I4mb-cm-|
i/y N/"" B 18 NOV 1970 U«.24mb-cm
C 23 SEPT I9TO U«.28mb-cro-|
! D 13 JUNE 1970 U«.20tab-cm
»« '' I i i i i I i i i i I i i i i I i i i i
0.0 05 10 1.5 2.0 23
MIXING RATIO x 10® (gm HNOj/qm AIR)
3 Mixing ratio profiles of HNO^ as derived from,
balloon flights made from Holloman AFB, New
Mexico.
—i—i—i—i—i—|—i—r-
A 12 5CPT 1072 U" .ZBBtrt-om
8 ISSCPTOTi
C >2 SEPT 197) U«
o «sentm
¦ I.I I.... J ,i I I..
1,11 ids11*1"
fj
1.0 I.I
MIXING RATIO * 10" (gm HNO, /jrn AM)
Fig. 4. Mixing ratio profiles of HNO« and total HN0»
amounts as derived from balloon flight* maeti
from Fairbanks, Alaska.
16-S
-------
Recently, a sensitive spectral radiometer system
has been used on the WB57F aircraft as part of the
Climatic Impact Assessment Program (CLAP) of the
Department of Transportation. With this system it
became possible to obtain data over a wide geographic
area in a short time. The instrumentation and data
analysis have been described recently. ^ Data con-
cerning the total column density of above flight
altitude have been derived from the measurements
made during a number of flights and are presented in
Figs. 5 and 6, These show a definite dependence on
latitude and possible trends of seasonal dependence.
It is evident that the HNO-j column density increases
from the equator to high latitudes. The latitudinal
variation in the northern hemisphere differs from that
in the southern hemisphere. This difference appears
to indicate higher HNO-j values in the winter than in the
summer. However, more data are needed before a
definite seasonal dependence can be concluded.
The present results for total HNO^ amounts
agree with the conclusions reached recently from a
number of experimental investigations by Lazrus and
Gandrud, ® Fontanella et al. , ^ and Harries et al. ^ The
latitudinal and seasonal variations appear to show sim-
ilar trends to those predicted theoretically by Rao-
Vupputuri. 8
Fluorocarbons
In view of the recent interest in certain fluoro-
carbons and the associated photochemical derivatives
which may occur in the stratosphere, we have re-
examined earlier balloon flight data to determine what
pertinent information could be obtained. Figure 7
shows laboratory spectra of a number of fluorocarbons
with resolution similar to that obtained during previous
balloon flights. In particular CF^Cl^ ("Freon 1 2")
shows two sharp spectral features at 921, 9 cm"' and
923. 2 cm-' in addition to the broad band absorption.
The CF2C12 band occurs in a stratospheric "window''
between minor absorptions by HNO^ and weak C
bands. Figure 8 shows the solar spectrum obtained
from Holloman AFB on August 1 2, 1 968 from 30. 5 km
at a zenith angle of 93. 5°. This figure shows spectral
features which can be attributed to CF2CI2 and possibly
CFCl-j. Justification of these identifications has been
published recently. 9 By comparing the balloon data
with laboratory data, a volume mixing ratio of 5x 10"'*
was derived for CF^C^ and a probable volume mixing
ratio of 2x10"'' was estimated for CFCl^. These
values are consistent with recent calculations by Crut-
zen, Cicerone et al. ** and Rowland and Molina.
' 1
Sept
r ' i - r ' i
r ' i ' i '
r • i T' "
- *71
-12-46 Km
A10 Km
Q« (€ Jan 74
*22 Jon 74
o 24Jan 74
a 26Jon 74
A A 27 Jan 74
a Sept 70
• **
* AA
Nov70*
A
o • 26 Jon 74
x Sept72
x Balloon Data
&
June 70 x A
A
May 70 x 4
r
L
so r
60^- RESOLUTION »04crf(-<
950
WAVENUMBER (cm-')
Fig. 7 Laboratory absorption spectra of selected halo-
genated hydrocarbons in the 780-950 cm"'
region.
KH 1
s«p»
*71
*S«pf 72
?Q*
Nov 70* ' 4
June 70*
* £
May 70x 4*46
vlt Apr-*
, t5 Km
/ f /a Km
• l2.Apr74
o* l7Apr74
o« jdApr 74
A A 21 Apr 74
x Balloon Data
60N
•• 21 Apr -—4
.a,,,. 1
19 Apr
20N 0~
LATITUDE
- t7Apr-
—1—(—
40S
Fig. 6 Variation of HNO^ column density with latitude
from April Airstream series.
cr,a,
0.925
P4o no m no{
lyf CO
UNIVERSITY OF DENVER
BALLOON FLIGHT
AUGUST 12,
WAVENUMSER (cm")
Fig, 8 Solar spectrum observed at a height of 30.5 km
and a zenith angle of 93. 5°. The height of the
tangent path is 18. 4 km and the optical path is
6, 0 air masses. The approximate height of the
tropopause is 14 km. The volume mixing ratio
for CFis 5xl0""v/v and for CFCl- is 2x
10
- " v?v
3
16-6
-------
WAVELENGTH (microns)
3 40
'infinite" resolution
ALTITUDE 5.5km
100 METER PATH H37C|
10"
\WITH 0. Ippmv HCy V/
WITH 0.Ippmv HCI A ^
0 2
,A_ A
loo
WITHOUT HCI
0.2
WITHOUT HCI
2950
WAVENUMBER (cirrO
'0.0
2945
2940
2955
2960
2965
FiS" 9 withou^HG?3°lUti°n at™sPheric absorption calculations for a 100 meter path at 5. 5 km with and
In addition, we have placed an upper limit on the
amount of HF (a possible product of the fluorocarbon
photochemistry) present in the stratosphere by com-
paring the detection limit of a previous balloon flight
with calculated atmospheric spectra containing trace
amounts of HF (see below). In a balloon flight made
from Chico, California on September 1965, the spec-
tral regions near 4001. 2 cm"-1 and near 4039, 1 cm"1
(where minimal atmospheric interference occurs)
show no absorption (<1%) at altitudes of observation
from 7 to 30 km. From these data an upper limit for
the mixing ratio of 3x10"'® v/v was set throughout the
lower stratosphere to 30 km. However, during this
flight no data were taken into the sunset. Sunset ob-
servations would provide at least a tenfold improve-
ment in the minimum detection limit.
Feasibility studies
of remote sensing of trace constituents
The line-by-line techniques have been used not
only for the analysis of observational data, but also for
feasibility studies of infrared measurements of trace
constituents. To properly measure a small amount of
gas from weak spectral features, accurate knowledge
is required of the fine details of both the experiment
and theory. In particular, the effects of interference
between various molecular species that might be pres-
ent in the spectrum requires special attention. The
line-by-line techniques are very powerful in such
cases, as is demonstrated in a recent study of the
solar spectrum during a search for stratospheric NO.*^
It was shown that a detailed line-by-line analysis of the
interfering molecules is a basic requirement for a
meaningful determination of minor constituents from
weak spectral features.
In a recent study14 we have applied the line-by-
line techniques to the problem of finding detectable
limits for a number of trace constituents against the
rest of the atmospheric background. In particular we
have studied the effects of interferences on the detecta-
bility of small amounts of HCI, HF, NH3 and SO?under
conditions corresponding to selected practical atmo-
spheric experiments with absorption and emission in-
frared spectrometers. Selected results for these
species are presented below.
"Infinite" resolution atmospheric absorption cal-
culations for a 100 meter path at 5. 5 km altitude are
shown in Fig, 9. Assuming ~1% measurable absorp-
tivity allows the measurement of ~10 ppbv HCI. It is
seen that there is practically no interference from at-
mospheric lines for a 100 meter path.
An absorption spectrum as would be expected in
Bolar spectrum measurements from 25 km at an 85°
zenith angle and 0. 08 cm"' resolution is shown in Fig.
10. It is seen that the HCI isotopic lines near 2945
cm"' are well separated from the atmospheric lines.
Absorptivity of ~1% allows the measurement of ~0. 1
pphv HCI.
W4VEL£M}TH (mcrofm)
3«0Q
h RESOLUTION 008cm"
ALTITUDE 25tm
II ZENITH 85"
101
0
001
S O oooi
WITHOUT HCI
"8940"
2910 "
WAVENUMBER (cm-'l
Fig. 10 Calculated absorption spectra with and without
HCI with a resolution oi 0. 08 cm-1 for an alti-
tude of 25 km and an 85° zenith angle.
"Infinite" resolution absorption spectra in the HF
region are calculated for a 100 meter path at 5. 5 km
altitude in Fig. 11, Scaling the figure to ~1% emissiv-
ity shows that ~3 ppbv HF should be measurable in the
4039 cm"' region, where the HF line is significantly
stronger than the background atmospheric lines, but is
not completely isolated.
Atmospheric absorptions with and without HF are
calculated for solar spectrum measurements from an
altitude of 25 km, at an 85° zenith angle and 0.08 cm"'
resolution in Fig. 12. The HF line near 4039 cm"' is
well above the atmospheric lines and ~1% measurable
absorptivity permits the measurement of ~0, 02 ppbv
HF.
4
16-6
-------
2480
WAVELENGTH (microns)
2 475
"INFINITE" RESOLUTION
ALTITUDE 5 5km
100 METER PATH
WITH 20ppbv HF
WITH 20ppbv HF
| [\ WITHOUT HF
WITHOUT HF
4035
4040
WAVENUMBER (cm*')
Fig. 11
"Infinite" resolution atmospheric absorption
calculations for a 100 meter path at 5. 5 km
with and without HF.
IO"3
icr4
icr5
io*6
l* IO"7
w A
'£ icr®
o
u icr9
WAVENUMBER (cm*")
900
I 0
10"
IO"2
IO"3
IO-5
0.15
) 0.10
(O
2
Ui
0.05
0.00
BLACK BODY AT 216° K
RESOLUTION I cm-'
ALTITUDE 15 Km
ZENITH 80°
WAVELENGTH (microns)
2 4 80 2.475
RESOLUTION 0.08 cm-'
ALTITUDE 25 km
2.470
ZENITH 85°
WITH 0 I ppbv HF
WJTHOUT. HF
1
WITH 0 I ppbv HF
WITHOUT HF
4030
015
0 10
0.05
0 00
4050
4040
WAVENUMBER (cm"')
Fig. 12 Calculated absorption spectra with and without
HF with a resolution of 0. 08 cm"' for an alti-
tude of 25 km and an 85° zenith angle.
Atmospheric spectral radiance calculated with
and without NH, in the 10 to 1 2um region for 15 km
altitude, 80° zer
2 in the 10 to 1 2^m region f
^nith angle and 1 cm"' reso
lution is
10.5
12.0
I0"!
Fig.
11.0 11.5
WAVELENGTH (microns)
13 Calculated spectral radiance with and without
NH, with a resolution of 1 cm"' for an altitude
of 15 km and an 80° zenith angle.
shown in Fig. 13. A number of measurable NHj fea-
tures can be seen with 0. 1 ppbv NH^. The radiance
levels are above 10"? w cm"2sr~*, well within the
capability of current spectral radiometers.
An absorption spectrum calculated for solar
spectrum measurements from an altitude of 25 km, ar
85° zenith angle and 0. 08 cm"' resolution is shown in
Fig. 14. Here a number of isolated NH^ peaks make
it feasible to measure ~0. 05 ppbv NH^.
1.0
109
10
I0"2
I0"3
IO4
-5
10'
S IO"1
i IO"2
UJ
I0"3
10"'
10"
WAVELENGTH (microns)
IQ.8
10,7
RESOLUTION 0.08 cm*'
ALTITUDE 25km
ZENITH 85°
106
i WITH 0 05 ppbv NH3
M 'l
WITH 0 05ppbv NH3
~jUL_JJL
WITHOUT NH3
WITHOUT NH3
J
910
_L
J 1 , 1 ...1,
940
015
0 10
005
0.00
015
0 10
0.05
000
920 930
WAVENUMBER (cm*')
Fig. 14 Calculated absorption spectra with and without NH with a resolution of 0, 08 cm"' for an altitude
of 25 km and an 85° zenith angle.
5
16-6
-------
The atmospheric spectral radiance calculated
for a spectral radiometer at 15 km altitude, an 80°
zenith angle and 1 cm"1 resolution is shown in Fig. 15.
No individual, features of S02 can be isolated by the
addition of 5 ppbv SO., to the atmosphere, but there
are many potentially measurable changes of the atmo-
spheric features due to the added SO^.
An atmospheric absorption spectrum in a solar
spectrum measurement from an altitude of 25 km, an
85° zenith angle and 0. 08 cm"1 resolution is shown in
Fig. 16. It is evident that the strong SC>2 Q-branch
near 1260 cm"1 allows the measurement of ~0. 5 ppbv
SO^ from ~1 % absorptivity of isolated SO^ lines.
Future Plans
1380
WAVENUMBER (cm-1)
1360 1340
1340
„ «L ZENITH
2x10
-r- I»I0
'£
O
*
m ArK bow AT 2I6«K
RESOLUTION Icm"
ALTITUDE 15km
80
WITH 5ppbv S02
WITH 5ppbv SO2
WITHOUT S02
WITHOUT S02
(RESOLUTION lcm-i
ALTITUDE 15km
ZENITH 80'
7.4 7.5 72
WAVELENGTH (microns)
Fig. 15 Calculated spectral radiance with and without
SO, with a resolution of 1 cm"1 for an altitude
of 15 km and an 80° zenith angle.
WAVELENGTH (microns)
7.39 7.37 7 3 5 7.33
2 0.001
RESOLUTION 0.08cm"
ALTITUDE 25km
ZENITH 85
WITH 0 5ppbv S02
WITHOUT SO?
Fig. 16
1355 1360 1365
WAVENUMBER (cm-')
Calculated absorption spectra with and without
SOz with a resolution of 0. 08 cm"' for an alti-
tude of 25 km and an 85° zenith angle.
Currently we are in the process of building
ground-based and balloon-borne infrared spectrome-
ters capable of better than 0. 08 cm"' resolution to be
used in the 1 to 30jjm region. Solar spectra to be ob-
tained with these spectrometers are expected to yield
new information on atmospheric trace gases at various
altitudes.
Acknowledgments
This research was supported in part by the Air
Force Cambridge Research Laboratories, by the
National Science Foundation, by the DOT CIAP pro-
gram and by NASA Ames Research Center. Acknowl-
edgment is made to the National Center for Atmospher-
ic Research, which is sponsored by the National Sci-
ence Foundation, for computer time used in this re-
search. The figures were carefully prepared by C.
Bauer.
References
D. G. Murcray, A. Goldman, F. H. Murcray, W.J.
Williams, J.N. Brooks and D. B. Barker, "Verti-
cal Distribution of Minor Atmospheric Constituents
as Derived from Air-Borne Measurements of Atmo-
spheric Emission and Absorption Spectra" AIAA
Paper No. 73-103, AIAA 1 1 th Aerospace Sciences
Meeting, Washington, D. C. , Jan 10-12, 1973.
D. G. Murcray, A. Goldman, W.J. Williams, F. H.
Murcray, J.N. Brooks, J. Van Allen, R. N. Stocker,
J. J. Kosters and D. B. Barker, "Recent Results of
Stratospheric Trace Gas Measurements from Bal-
loon-Borne Spectrometers" Proceedings of the
Third Conference on the Climatic Impact Assess-
ment Program, Edited by A. J. Broderick and T. M.
Hard, pp. 184-192, Department of Transportation,
DOT -TSC-OST -74-15, 1974.
D. G. Murcray, A. Goldman, A. Csoeke-Poeckh,
F. H. Murcray, W.J. Williams and R.N. Stocker,
"Nitric Acid Distribution in the Stratosphere" J.
Geophys. Res., vol. 78, pp. 7033-7038, 1973.
D. G. Murcray, D. B. Barker, J.N. Brooks, A.
Goldman and W.J. Williams, "Seasonal and Latitu-
3.
4.
5.
7.
dinal Variation of the Stratospheric Concentration
of HNO " Geophys. Res. Lett. , vol. 2, pp. 223-
225, 1975.
A. L. Lazrus and B. Gandrud, "Distribution of
Stratospheric Nitric Acid Vapor" J. Atm. Sci. ,
vol. 31, pp. 1102-1108, 1974.
J. C. Fontanella, A. Girrard, L. Gramont and N.
Louisnard, "Vertical Distribution of NO, NO^ and
HNO3 as Derived from Stratospheric Absorption
Infrared Spectra" Appl. Opt. , vol. 14, pp. 825-
839, 1975.
J. E. Harries, J. R. Birch, J. W. Fleming, N. W. B.
Stone, D. G. Moss, N. R. W. Swann and G.F. Neill,
"Studies of Stratospheric H^O, HNO
N20 and NOz
Proceedings of the Tfiird Conference
from Aircraft
on the Climatic Impact Assessment Program, Edi-
ted by A. J. Broderick and T. M. Hard, pp. 197-
212, Department of Transportation,DOT-TSC-OST-
74-15, 1974.
8. R.K. Rao-Vupputuri, "Seasonal and Latitudinal
Variations of N20 and NOx in the Stratosphere"
J. Geophys. Res. . vol. 80, pp. 1125-1132, 1975.
9. D. G. Murcray, F. S. Bonomo, J.N. Brooks, A.
Goldman, F.H. Murcray and W. J. Williams,
16-6
-------
"Detection of Fluorocarbons in the Stratosphere"
Geophys. Res. Lett., vol. 2, pp. 109-112, 1975,
10. P. Crutzen, 1 Estimates of Possible Future Ozone
Reductions from Continued use of Fluoro-Chloro-
Methanes (CF2C12> CFCl.)" Geophys. Res. Lett.,
vol. 1, pp. 20S>-2o8, 1974.
11. R.J. Cicerone, R. S. Stolarsky and S, Walters,
"Stratospheric Ozone Distribution by Man-Made
Chlorofluoromethanes" Science, vol. 185, pp.1165-
1167, 1974.
12. F. S. Rowland and M. J. Molina, "Chlorofluoro-
methanes in the Environment" Revs. Geophys.
Space Phys. . vol. 13, pp. 1-35, 1975.
13. D, G. Murcray, A. Goldman, W.J. Willaims, F.H.
Murcray, J. Van Allen and S. C. Schmidt, "Obser-
vations of the Solar Spectrum in the 1800-2100 cm''
Region and the Search for NO Lines" Proceedings
of the Third Conference on the Climatic Impact
Assessment Program, Edited by A. J. Broderick
andT.M, Hard, pp. 246-253, Department of
Transportation, DOT-TSC-OST-74-15, 1974.
14. A. Goldman, W.J. Williams and D. G. Murcray,
"Measurements of Trace Constituents from Atmo-
spheric Infrared Emission and Absorption Spectra -
A Feasibility Study" Final Report, Sept. 1974,
Contract No. NAS2-8200 for NASA Ames Research
Center, by Department of Physics and Astronomy,
University of Denver, Denver, Colorado 80210.
7
16-6
-------
A NET ENERGY ANALYSIS OF THE USE OF NORTHERN GREAT PLAINS
SURFACE MINED COAL IN LOAD CENTER POWER PLANTS
Thomas Ballentine
Department of Environmental Engineering Sciences
University of Florida
Gainesville, Florida 32601
Introduction
As this nations richer, more easily obtained
fossil fuels have declined in abundance, increased
emphasis has been placed on the development of prev-
iously marginal domestic energy resources. Yet most
of these energy resources are inherently more dilute
than those which propelled our nation to prominence
as a world power. Greater energy expenditures will
be required to concentrate and deliver them in a use-
ful form to the final point of consumption. Develop-
ment of these marginal resources will necessarily
require that a greater portion of economic output
(which Is generated by the use of energy) be fed back
to energy finding, producing, transporting, and con-
version activities. Less net energy per unit raw
energy produced will remain for use by the general
economy. Thus inherent energy constraints on the U.S.
economy suggest that economic growth measured in
physical output may no longer be possible and that the
present economic downturn will continue.
Among these marginal energy resources may be the
surface mineable coal resources of the Northern Great
Plains. This is suggested by the coal's low heat
content, its great distance from consumption centers,
the necessity for its conversion to higher quality
forms before it will be useful, and the possibility
that regional energy patterns largely dependent on
free natural energies will be interrupted due to the
extraction of the coal. High energy costs will also
be incurred in constructing and supporting new commun-
ities in arid, isolated areas; and in sequestering
adequate water for coal development.
Simultaneous with increased emphasis on Northern
Great Plains coal development, considerable controver-
sy has arisen over whether and how such development
should occur. As a result, extensive studies have
been done to assess and forecast the effects of coal
development on the region. Complicating this assess-
ment, though, have been the current economic recession,
the increasing difficulty of obtaining investment
capital for coal development, the slowdown in the
growth rate of demand for electricity, the lack of
direction in national energy policy, and opposition to
development by various groups.
The urgency of Che present energy situation dic-
tates that an encompassing unity be found among these
difficulties and conflicting viewpoints, and that an
objective appraisal of Northern Great Plains coal as
an energy Bource to the nation be performed. A useful
methodology for such an appraisal is net energy anal-
ysis utilizing energy concepts, an energy systems
language, and an energy quality scale developed by
Dr. H.T. Odum of the University of Florida.
A net energy systems analysis can clarify what Is
energetically possible as compared with what is
desired by energy policy makers and the public.
Energy constraints imposed by the next larger system
as well as internal constraints will ultimately deter-
mine what will and will not occur. If surface mining
of Northern Great Plains coal is a large net energy
yielder (when all energy costs are included), its
development la assured, if not, then coal development
will ultimately prove unsuccessful; its failure delayed
only by the delay of system feedbacks.
Holistic Overview of Northern
Great Plains Coal Pevelopment
Utilizing Energy Systems Diagrams
farmers'
[vfGETAT!
DAMS AND I
DOWNSTREAM
WATER US£RS
I TOURIST
rfiucft
TOURIST^]
tourists!
our J
TSRd
Figure 1. Energy Issues Associated with
Northern Great Plains Coal Development,
Figure 1 is an energy diagram drawn to facilitate
holistic understanding of the most important issues
associated with Northern Great Plains coal development.
The diagram Indicates, on a macroscopic level, regional
energy patterns along with their connection and ulti-
mate dependence upon the U.S. economy and its main
energy sources. The region's main subsystems are ident-
ified along with the main energy, water, and material
flows (solid lines); and money flows (dashed lines)
between components. The diagram indicates the interde-
pendency of all subsystems, the importance of energy
subsidies to the region from the environment which are
not directly associated with money exchange, competition
for water, the effect of removing land from natural
production by surface mining, and inputs to the region
from the national economy.
Traditionally, the various economic sectors have
been concerned almost exclusively with their own well-
being, represented by the flows directly connecting
1
17-1
-------
them to the larger system. But as indicated here,
each sector is ultimately dependent upon all others.
For instance, if energy flowing into the national
economy decreases, feedbacks from the economy to the
region will Increase in price and decrease in avail-
ability. Such things as farming equipment and
supplies will become more expensive. Ultimately,
tourism will decline along with the standard of living.
In another case, if water is used within the
region for energy production, it will be unavailable
to dams and downstream users. Thus their outputs will
have to be produced with larger portions of purchased
energy fedback from the economy, rather than free
water, and their net contribution to the economy will
decline. Examples of this are the production of
electricity with fossil fuel rather than by hydro-
electric methods and irrigation by pumping with elec-
tricity rather than by gravity flow.
The vitality of the Northern Great Plains region
is underwritten by free environmental energies which,
although small compared with the energy contained in
the area's coal resources, have the ability, through
amplifying interactions, to release large potential
energies. For example, vegetation is directly con-
sumed by livestock, aids in decreasing runoff and in
the capture of rainfall for the system's use, contri-
butes aesthetic value which assists "pumping" in
tourists, and provides services to return land dis-
turbed by surface mining to a fully restored condition.
Competition for water will exist between vegeta-
tion, farmers and livestock, downstream water users,
the tourist industry, and the coal industry. Water
used downstream may generate more energy value than
has yet been determined, for through multitudinous
interactions the area's waters generate benefits all
the way to the Gulf of Mexico. Among these benefits
are: the generation of hydroelectricity, navigation,
the maintenance of water quality which allows irriga-
tion and minimal water treatment by all users, and
the maintenance of aquatic and wetland ecosystems.
From Figure 1, it can be seen that the cost of
coal development in the semi-arid Northern Great
Plains may be higher than anticipated due both to the
loss of benefits currently generated within the system
and elsewhere, and to the high energy cost of coal
producing and transporting systems. One of the
prime advantages offered by diagrams such as Figure 1
is that they facilitate holistic understanding of the
consequences of various policy decisions before irre-
versible commitments are made. They also offer
considerable insight into avoiding conflicts and
difficulties during the implementation of policy deci-
sions. Possibly most important, they can suggest ways
by which conflicts and competition for limited resour-
ces can be altered by system modifications into rela-
tionships of symbiotic cooperation so as to maximize
benefits to the total system of man and nature.
Introduction to Energy Systems
Theory and Methodology
Energy as a Common Denominator and Measure of Value.
Energy is the common basis for all processes.
Its use is necessary for the creation, survival, and
organization of all systems. Without energy, systems
cease to exist and all order (structure, information,
organization, etc.) degenerates to disorder. Since
energy use by systems is universal, a good measure of
the vitality of any system is the rate at which
energy of all types is being captured and processed
by that system} A measure of the stored value of the
system itself is the amount of energy required to
create the system.
Energy Quality.
All energy does not have the same ability to do
work. As required by the second law of thermodynamics,
higher quality more ordered energies can do more work
during their degradation to a disordered state. A
kilocalorie of sun, wood, and coal have different abil-
ities to do work due to their different energy qualities,
but they all have the same dispersed heat value. When
comparing energies then, a quantitative scale of
ability-to-do-work is needed so that different kinds of
energy can be compared on a common basis and their
relative contributions to overall system vitality can
be determined.
Type of Energy Work Potential of 1 Kcal of this
Energy In Kcal of Fossil Fuel
Equivalent (FFE)
Sun
0.0005
Sugar
0.05
Wood
0.5
Coal
1.0
Electricity
3.6
Table 1. Energy Quality Scale.
Shown in Table 1 is an energy quality scale
developed by Odum^ which indicates the work potential
of 1 Kcal of various types of energy In terms of the
work potential of fossil fuel. Energy quality increases
from sun to electricity because in each case one Kcal
of the higher quality energy form, when properly used,
can accomplish a more valuable work service for a
system than the lower quality form. As shown, it would
require 2000 units of sunlight or 0.278 (1/3.6) units
of electricity to do the work of one unit of fossil
fuel equivalent.
The figures in the table are subject to refinement
as techniques for their evaluation are further devel-
oped and studies in progress are completed. However,
none is expected to change substantially. Considerable
effort by Odum and colleagues is presently being
devoted toward both extending this concept and quanti-
fying other energy quality factors.
Lotka's Maximum Power Principle
The criteria for the proper use of energy by any
system is set forth by Lotka's Principle.It states
that surviving systems maximize their use of power from
available energy sources and distribute this power so
as to assure their long term survival.
According to Odum,^'^ as systems process energy
they continually upgrade lower quality fuel energies
into higher quality energy forms. The justification
for this upgrading is found in the ability of these
higher quality energy forms to accomplish feedback
amplification to facilitate the capture of yet more low
quality fuel energies and thus assure system survival.
Any system that pays the energy cost of upgrading
lower quality fuels to higher quality forms and then
uses the higher quality forms as a fuel rather than as
an amplifier cannot long remain competitive in the real
world.1»3>4 The implication here too is that systems
do not and cannot manufacture more high quality forms
than needed to "pump in" the low quality energies avail-
able to them. An example of this principle at-work may
be the present difficulties being encountered by the
electrical utility industry in attempting to expand
17-1
-------
generating capacity when fossil fuel availability is
already declining. At a larger scale, the difficulty
in stimulating capital expansion of U.S. industry may
have a similar origin.
Odum1 has further suggested that the following
corrolaries emerge from Lotka's Principle. First,
growth is successful during periods of abundant energy
because those systems shich grow most rapidly preempt
energy usage by competing systems. Second, when
excess energies have all been tapped, a premium is
placed on mutual cooperation among systems to utilize
remaining energies more efficiently. In this case,
maintenance of existing structure will pre-empt the
construction of new structure. Only worn-out compon-
ents will be replaced. And third, a system's total
activity and structure will decline as it begins to
use significant percentages of energies with lower
net yields.
Set Energy
Whereas the viability of a system depends on the
total energy processed, the success of an energy
extraction alternative depends on the net energy
yielded to the larger system.1 Net energy is the
energy available after all the energy costs assoc-
iated with finding, producing, concentrating and
upgrading, and delivering the energy to its final
use site in a useable form have been paid. Deducted
also must be the loss of natural energies due to
environmental stresses imposed by the extraction and
use of the energy source.
A useful number for appraising the net energy
available from various alternatives is the net energy
ratio, defined as the ratio of yielded energy to the
energy cost of developing the yield expressed in
energies of the same quality.
Enhancement of Overview by Modeling the Next Larger
System.
A practice advocated by Odum to better under-
stand a system is to model the next larger system in
order to investigate and clarify the constraints it
imposes on the system of interest. Shown in Figure 2
is an energy diagram of the U.S. economy. It will be
used to clarify sane of the principles already set
forth, to introduce some new principles, and to
investigate the constraints which will be imposed on
coal development in the Northern Great Plains.
1UTJES
fUCUL CtC
WARGIhAL
OO'-tESTiC
nsiouHS*
ACKlCULTUNAi
SYSVWS
usfe
*£TA90U«M
or cp>ew*
Figure 2. Aggregate Energy Diagram for
the U.S. Economy.
The most important point of Figure 2 is that the
viability of the U.S. economy (metabolism of the
general economy) depends upon the continued inflow of
of all its ehergy sources, because all are contributing
to the ability to do work. For instance, if the inflow
of foreign oil is decreased, the ability to feedback
high quality energies (drilling rigs, draglines, tech-
nological services, etc.) tt> amplify domestic energy
production will also decrease. There is some evidence'
to suggest that domestic energy production has actually
been subsidized by Inexpensive imported oil since the
early 1960's. If this is so, an upturn in.energy
imports may have to occur before domestic energy
production can be increased.
Analysis of Figure 2; however, indicates that
increasing domestic production from marginal resources
may not be the wisest course of action. For if
production from these resources is to displace foreign
oil, a greater portion of gross energy production will
have to be used to construct and operate new energy
production facilities. Considerable upgrading, at a
high energy cost, may be necessary before these energies
are in a useful form. Thus the net effect of a policy
decision to increase domestic production may be smaller
net yields to the general economy accompanied by a
conversion of much of the economy to energy
getting endeavors. Such a transition would probably
cause widespread unemployment in those sectors of the
economy which are ill-equipped to contribute toward
increasing energy production.
Current U.S. trade surpluses plus further analysis
of Figure 2 suggest that the economy may choose increas-
ing oil imports as its best energy strategy. Since
smaller amounts of energy company structure are involved
in importing oil on a per barrel basis, more of this
energy is available to the general economy. For the
energy received, a diversity of goods and services is
currently being exported. Thus an increase in oil
imports will probably not necessitate drastic changes
within the economy. The primary change will be that
more U.S. production will be consumed abroad rather
than at home.
Figure 2 also illustrates the great importance of
natural and agricultural systems. If they are stressed,
then their net energetic contributions to the economy
will decline; Two examples will serve to illustrate
this principle. In the case of a natural system, if a
highly productive estuarine fishery is destroyed as a
result of dredge and fill operations, its food
production may have to be replaced by a much more
energy intensive agricultural operation elsewhere. In
the case of an agricultural system, if productive low-
lands are flooded by a reservoir, energy intensive
farming on marginal uplands may become necessary. Such
actions decrease the amount of fossil fuel available
for other uses, including energy development, and
eliminate sustainable natural energy sources, From
the above it can be easily concluded that protection of
natural systems has its place among energy production
strategies.
The Flow of Money Countercurrent to Energy.
Indicated by dashed lines in Figure 2 is the
countercurrent of money associated with energy. Not
all energy flows are directly associated with money,
bijt all do ultimately contribute value to those flows
upon which man places monetary value.^ When energy
is received money is exchanged which can then be
exchanged by the holder for other work services. Energy
then, and not money, is the true basis of value. Money
is simply a mechanism which aids the efficient
allocation of energy and work services and therefore
aids economic systems in selecting for maximum power.
Maximization of money flows does not necessarily maxi-
mize system vitality, especially if free environmental
subsidies to the economy are foregone in the process.
17-1
-------
Table 2. Ratio of U.S. Energy Flows Co Gross National Product,
Year
Fossil
Fuels
10l5Xcal
/yr
Fossil Fuels
Plus Natural
10 Kcal/yr
10H
Total Energy
par
lO^Kcal FFE/S
Relative
Energy
, per
Dollar
PurchaJlrig
Power
of the
Dollar
(C?l)
1967
14.68
21.42
793.9
27.0
1.000
1.000
1968
15.56
22.30
864.2
25.8
0.956
0.960
1969
16.37
23.11
930.3
24.8
0.919
0.911
1970
16.94
23.68
976.4
24.3
0.900
0.860
1971
17.33
24.07
1050.4
22.9
C.34B
0.824
1972
18.17
24.91
1151.8
21.6
0.800
0.799
1973
18.87
25.61
1294.9
19.8
0.732
0.7S2
1974
18.46
25.20
1397.4
18.03
0.668
-
Energy to Dollar Conversions.
Shown in Table 2 are energy-to-dollar ratios for
the U.S. for the years 1967 through 1974 taken largely
from Klystra but updated by the author for recent
years. The "fossil fuel" column represents the total
amount of fossil fuel used within the U.S. economy
plus the fossil fuel equivalent that would have been
required to generate the electricity produced by
nuclear and hydroelectric methods. The "fossil fuels
plus natural" column is the sum of the first column
plus 6.74 X 10" Kcal FFE, which is the fossil fuel
equivalent of the total incident solar radiation on
the U.S. land mass in one year.
The ratios in this table have been used to convert
dollar costs to an energy cost equivalent, especially
when the energy required to produce the good or service
is consumed in diverse ways and locations.
Since 1967 the energy (or equivalent work) that
can be purchased with a current dollar has decreased.
It is notable that this decrease corresponds very
closely with the purchasing power of the dollar as
measured by the Consumer Price Index. This suggests
that money's value is in fact based upon energy and
that declining net energy ratios are one of the prime
driving forces behind world inflation.
Inflationary Effect of Current Monetary Policies.
Possibly the most severe constraint on Northern
Great Plains coal development has been inflation. The
inflated cost of facilities themselves along with the
increased cost of capital have combined to make
investors extremely wary of commencing proposed
projects.® There is much to suggest that this imped-
iment to development is not about to dissipate.
Increasing the possibility of even more severe
inflation are the government's recent policies of
large budget deficits and increasing the money supply.
ha indicated in Figure 2, such pump priming activities
can only be successful when they can cause an increase
in the net energy flowing into the economy. Otherwise
the energy-to-dollar ratio will decrease resulting in
inflation. Unfortunately, the U.S. possesses few
untapped energies with high net energy ratios which
can rapidly be brought on stream to give value to the
increasing supply of money. Thus inflation will
continue to increase the dollfar cost of energy develop-
ment. In fact, inflation may increase even further
this winter when the economy simultaneously encounters
a severe natural gas shortage and is operating on a
reduced solar input.
Necessity for Minimal Disruption of Existing Energy
Patterns During the Establishment of flew Subsystems.
transportation
service.
Tourism
Atmos.
Oust j
Evaporation
Wildlife
V««.
Rain
ItfH-
trotion,
runotl
[Dam
runoff-tfrumf low
irvoii
'Soil
Umsnl]
Iter,
Figure 3. The Northern Great Plains
Without Coal Development,
Detp
iqvlfer
17-1
-------
The most severe constraints on coal development
after inflation have been internal constraints imposed
by the Northern Great Plains system, a system whose
energy patterns are largely dependent on sustainable
natural energy sources."'10'11 Shown in Figure 3 is a
macroscopic energy diagram of the Northern Great Plains
system as it existed without significant surface mining
activities. Components of the natural system and their
interactions are diagrammed on the left. These compon-
ents are driven by the sun and wind, rain (which at
times is unreliable), and soil formation processes
which proceed at a very slow rate. None of these
energies or the many services exchanged by the natural
system's components are paid for with money exchanges.
Only further downstream, when they enter the value
system of man, do they acquire economic value.
Diagrammed on the right are economic activities; each
of which requires an exchange of money (dotted lines)
for work services rendered between components.
Evident in this diagram is that the entire
system's viability is underwritten by the flow of
natural energies and stored energy value in the natural
system for which no money is exchanged. For instance,
the great value of the geologic - hydrologic regime,
and the vegetation which combine their efforts so
effectively to "pump" water underground for gradual use
by the system is largely unrecognized by the economic
sector.
As shown in Figure 3, the economic value of
vegetation is realized only when it is converted to
livestock protein and sold. Yet if vegetation is
destroyed as a result of mining, its service in aiding
recharge to the region's aquifers will be foregone.
Oust levels will consequently increase discouraging
tourism; and flooding and increased sedimentation will
occur downstream causing property dawage, decreased
reservoir lifespans, and increased stresses on aquatic
ecosystems.
Thus if coal development is to become established
in the Northern Great Plains, considerable effort will
have to be put forth not to disrupt the system's
natural energy sources.
Net Energy Analysis of load
Center Power Generation
After considering the multitude of constraints
imposed on Northern Great Plains coal development by
larger systems, due consideration will now be given to
the net energy ratio associated with the use of this
coal for load center, electrical power generation.
Energy Goat of Mining and Transport.
Shown in Figure 4 is an energy diagram for the
mining and transport of one-million tons of 8,500 BTU
/lb coal to a load-center power plant 1000 miles
(1200 rail miles) from mine mouth. Energy values
(solid lines) and dollar costs (dashed lines) are
labeled adjacent to their respective pathways. Energy
values are in units of 1010 BTD's Fossil Fuel Equiv-
alent (BTU FFE). These values are itemized in Table
3. The energy costs of mining were calculated using
data mainly from Bureau of Mines Information Circular
8661,12 the only detailed information available to
the author. They are estimated to be quite conservative
since the coal industry is currently negotiating
contracts for F.O.B. coal at twice the price quoted
in IC8661. Detailed calculations are of sufficient
length to preclude presentation, here, but can be
obtained from the author upon request,
General support consists of all the energy required
to support the workers' lifestyles, provide capital
equipment, operate and maintain the various systems;
and to produce, refine, and deliver the fuels used.
Money is not exchanged directly for the fuel, but
instead is exchanged in direct proportion to the
general support services necessary to provide the fuel
(9.31 x IOIObtu FFE). The actual energy value of the
fuel (52.93 x lO^BTU FFE) is provided free by the wort
of geologic cycles which have upgraded and concentrated
stored solar energy. The energy cost of general
support was determined by multiplying the dollar costs
of non-fuels by the ratio 99,000 BTU FFE/$, which is
the approximate energy-to-dollar ratio for the 1968-
1970 period.
RECLAMATION
general
£LECTHIClT*
S'JPPQRI
I08.9C
.411,000,000
1S44QQ0
'¦"I**® „„
energy cost of
mining and delivering coal is represented by this
input. Therefore if diesel fuel shortages develop,
western coal mining may rapidly become impossible. Yet
this vulnerability is largely unrecognized by economic
analyses because at the present time only about 5% of
the dollar cost of coal delivered to load center is
directly associated with fuel. Economic analyses,
therefore, may overemphasize the value of western coal
as an energy source and undetemphasiae its energy cost.
Energy Cost of Power Generation.
Table 4 lists the energy costs associated with
power generation from coal in a 1000 JWE power plant
exclusive of natural productivity foregone. Costs per
KWH sent out from the plant we're determined by assuming
that the plant would operate at 65% capacity over a 30
year life. 10 megawatts are used to operate cooling
towers so only 990 megawatts are sent out. The capital
cost of the plant was determined by converting the
dollar costs of energy intensive materials and machinery
to energy at four times the average energy-to-dollar
ratio for the year in question, The other costs were
converted at the average ratio listed in Table 2.
17-1
-------
Table 3. Energy and Dollar Costs Associated with
Mining and Transporting One Million Tons of
Northern Great Plains Surface Mined Coal to
a Load-Center Power Plant.
Description
Dollar Cost
Coal Energy Made Avail-
able at Minemouth
Energy Cost
(I010BTU FFE)
$2,660,000.
1700.00
100,000.
Cost of Mining and Delivery of Coal to Minemouth
Location12.
General Support 2,500,000.
Energy associated with
diesel fuel
General support to
acquire fuel
Energy In diesel fuel
Locally generated
electricity
Reclamation
Photosynthesis foregone
Natural System destroyed
Subtotal $2,730,000.
Transport of Coal to Load Center
General Support $8,400,000.^
(10% used to obtain fuel)
Diesel fuel —
Subtotal
24.75
60,000.
70,000.
.99
5.57
9.22
0,69
0.66
0.30
Total Invested and
Foregone Energies
$8.400.000.
$11,130,000.
42.18
83.16
47.36U
130.52
172.70
Table 4. Energy Cost of Electric Power Generation in a
1000 MWE Coal-Fired Power Plant Exclusive of
Natural Productivity Foregone!^15
Description
Energy Cost
(BTU's FFE)
Total Capital Per
Cost (1010) KWH
Sent
Capital Investment to Construct Power Plant
Heavy manufactured equipment 2891.1
Labor, Professional Services,
Other
Subtotal - Capital Investment
cost
Energy Investment to Operate and
Maintain Power Plant
965.2
3856.3
Subtotal - total energy
investment to operate power
plant
Coal Used in Power Generation
(990 MWE sent out)
Total
3856.3
171.96
57.07
229.03
83.18
312.21
^a
10.897.0
3856.3 11,209
(a) BTU's of Fossil Fuel Equivalent available at
the plant site exclusive of the energy cost
of production and transport.
Table 5. Annual Energy Cost of Railroad Transport
of Coal with Power Generation in a 1000
MWE, Load Center Power Plant.
Total Energy Cost of Power Generation
Table 5 itemizes the total energy costs of bulk
electric power at the plant generated by a load center,
1000 MWE coal fired power plant. The energy cost of
distribution is not included. Data for the associated
calculations were obtained from Tables 3 and 4, and
were based on a requirement of 3.613 million tons of
8,500 BTU/lb coal for operation of the plant at 657=
load factor.
Figure 5 is an energy diagram which used the
values in Table 5 to emphasize the net energy and
systems aspects of power production. Shown in the
hexagons, which are symbols for structures capable of
drawing energies to maintain themselves, are the
energy costs of these structures. At each heat sink
is shown the total energy consumed at this respective
stage. Heat losses include operation and maintenance,
and the prorated energy cost of structure used in
the process.
Not included, but possibly significant, are
energy losses associated with power production. For
example, acid leaching of soil nutrients could possibly
lower productivity over a wide region for a considerable
timespan. In urban areas, increased use of coal will
increase the maintenance cost of structure and decrease
its useful lifespan, and also increase energy
expenditures associated with health care.
IP*0 BTU FFE
Invested
and Lost
Energy
Starting Energy
Energy Out
Mining Purchases 148.93
Photosynthesis lost 3.47
Transport of Coal 471.57
Power Generation 175.99
Total 799.96
7677.63 6142.10
6142.10 6142.10
(5636.5 x
106KWH)
6142.10 6925.5
The net energy ratio associated with bulk
electricity at the power plant not including the above
losses is 8.66 units of yield per unit of investment
plus natural loss (6925.5/800.0). This Is a moderately
high ratio as compared with some other energy alter-
natives such as nuclear, oil shale, solar, and wind;
but difficulties discussed in earlier portions of this
paper will probably continue to impede coal development.
Foremost of these is that considerable energy will be
needed to initiate Northern Great Plains coal develop-
ment, yet the U.S. is a system already constrained
6
17-1
-------
POWER
TRAINS
riAHT
U
-------
TRACE CONTAMINANTS FROM COAL-FIRED POWER PLANTS*
R. C. Ragaini and
-------
Integrity of the analyses was Insured through the
use of NBS standard reference materials 1632 and
1633 which were analyzed along with samples from
this study.
Particle Size Characterization
Particle-size distribution parameters for
stack aerosols observed in this study were obtained
from SEM observation of particles collected on
membrane filters and on impactor stages. In the
former case number-size distributions were obtained
by counting particles in discrete size ranges, after
solvent suspension and sonic redispersion, with the
aid of an SEM-Quantimet system. Since the number
distributions observed were log normal, both mass
and surface median diameters (M™ and S__,
respectively) were calculated from the observed
number median diameter (Hcq) and geometric standard
deviation (a ). In the latter case, values of the
M__ and a f§r the aerosol were obtained from log
probability plots of the percent cumulative mass vs
Impactor Calibration
Fifty percent cut-off diameters (d,-n) for the
University of Washington Mark III cascacrer impactor
are available from the manufacturer. However,
accurate determination of stack aerosol distribution
parameters precluded use of the quoted d^0 values
for the following reasons: first, the actual value
of stage parameters are dependent on the
distribution parameters of the aerosol being
sampled; secondly, nonideal impactor behavior such
as particle bounce-off and reentrainment alters
values of the actual stage parameters. Therefore,
stage diameters used in determining stack aerosol
distribution parameters were determined empirically
by SEM observation. In addition, SEM observation of
the back-up filter providedan value where only an
upper unit could otherwise be reported.
A typical in-stack impactor sample, run for 2
hours at a flow rate of 0.38 cfin (corresponding to
102J5 isokineclty) was chosen for SEM examination.
Since the collection substrates were densely covered
with particles, portions of the substrates were
sonicated in trichloroethylene. The resulting
dispersions were centrifuged onto Formavar coated
200-mesh transmission electron microscope grids.
Particles were sized from negatives using a Zeiss
particle analyser.
Results and Discussion
Impactor CfcT Jbratlon Data
Impactor calibration data are shown in Table
1, in which observedH-0 values are tabulated along
with dj-- values obtained using calibration curves
provided by the manufacturer (quoted values). To
facilitate the comparison, quoted d_0 values were
adjusted to account for the densit/of fly ash —
assumed to be 2.2g/cm . Since the fly ash particles
observed were spherical, shape correction factors
were not applied.
Agreement between the N Q and quoted d_Q
values is quite good for stages H, 5, 6, and 7.
For stages 1,2, and 3, however, observed values are
much Bmaller than those quoted. On stage U (on
which the largest particles were collected) 95.5f
(i.e. 2a) of the particles observed have diameters
less than 5.7 um- In fact, no particles greater
than 3 Mm were observed on any of the impactor
stages. Deposition on stages 1, 2, and 3 is
attributed to the finite efficiency of an impactor
stage for any size particle.
In Fig. 2 is shown the number size
distribution observed for particles collected on the
back-up filter. An additional point (i.e. K-q
value) was obtained from particle counts done on the
back-up filter. The distribution observed is
clearly bimodal, having number median diameters of
O.058 and 0.80 ymm, and respective geometric
standard deviations of I.60 and 1.65. Since stage 7
has a d,- value of 0.32 um, the large particle
distribution on the back-up filter is clearly due to
particle bounce-off and/or reentrainment. On
inspection of the number distribution, the
bounce-off problem seems small. However, from the
corresponding cumulative mass distribution, shown in
Fig. 3, it is seen that approximately 98!? of the
particulate mass observed on the back-up filter is
accounted for by the large particle distribution.
Cumulative mass curves for a typical impactor
run are plotted in Fig. t using both the quoted d_Q
and observed N-Q values. In both cases a value or
0.065 »m was taken as the K.fl of the back-up filter
for reasons described below. For the quoted curve,
no reentrainment was assumed. For the observed case
only 2% of the observed back-up filter mass was
plotted vs the 0.065 Vim value.
Values of the M„Q determined from each of the
curves agree reasonably well, having values of 1.28
um and l.lU vim for the quoted and observed values
respectively. The dispersion of the distribution,
given by 0 , however, is shown to be less broad by
the observed curve (a * 2.03) than by the quoted
curve. From the figure, one can see that the quoted
curve underestimates the fraction of aass leas than
a stated size for particles of diameter greater than
1 um, and overestimates the corresponding fraction
less than 1 um, The discrepancy may be of
particular importance in assessing toxicological
effects since alveolar deposition appears to
increase with decreasing particle size .
IlSffiSBi&U'j Q valwa ag values of stack Aerosols
In Fig, 5 is shown the number-size
distribution observed for the whole-filter sample.
The distribution observed is clearly bimodal, and
can be resolved into two distinct distributions
having number median diameters at 0,065 vm and 0.565
Vim. Corresponding Mc_ values for the two
distributions are O.O08 um and 1.04 u»'» respective e
values are 1.37 and 1.57. These parameters 8
represent the distribution of total particulates.
Values of the M-0 for the large-particle
distribution can also be obtained from impactor
data. Based on the back-up filter, the M__ of small
particle distributions is 0,103 um, with 0 of I.56.
These values are in good agreement with M ® of 0.088
and 0 of 1,37 determined for the whole filter
sample. Due to the possibility of distortion of the
back-up filter distribution such as by collection of
some of the small particles on impactor stages, the
whole filter and a are considered the more
accurate parameters describing the total mass
distribution.
Elemental distributions may not follow the
total mass distribution. Davison et_ al., and
Kaakinen et al., have suggested that surface
adsorption results in the concentration of volatile
elements on small particles. If this is so, the M„
of the adsorbed element would be equal to the
2
17-2
-------
surface median diameter (S,-n) of the small particle
distribution. The small particle S_g observed on
the whole filter is 0.079 pm. It 1b uncertain,
however, whether or not the entire mass of the
volatile element is surface adsorbed. Some of the
mass may in fact be contained within the volume of
the particle. If this is the case, the actual M_0
for a volatile element may lie between the mass and
surface median diameters (i.e. 0.079 ym < M,-. <
0.088 ua). >0
In Fig. 6 percent cumulative mass observed on
impactor stages is plotted vs observed impactor
U_Q values for As and Al. Stages with similar
observed H Q values were combined. In each case,
the mass of the element associated with the large
particles found on the back-up filter was subtrated
from the total mass observed, leaving only the
elemental mass associated with the small particle
distribution. From the figure, it is evident that <
2% of the Al is associated with the small particle
distribution. However, about 50? of the As is in
fact associated with small particles.
In table 2 are listed the relative amounts of
elements observed on small and large particle
distributions in stack fly ash, along with
distribution parameters determined for the large
particle mode. In Group I and Group II are listed
elements with significant small-particle
association. Elements with little or no
small-particle association are listed in Group III,
Elements within each of the groups are well
correlated as a function of impactor stage; having
correlation coefficients (r) i. 0.9^. In Group I all
of the elements except V have inorganic species with
boiling point or sublimation temperature at or below
the combustion temperature of a coal-fired power
plant; typically 1500-1600 C. In Group II, all of
the elements listed have no reasonable inorganic
species boiling or subliming at temperatures s
1550 C. At combustion temperatures, inorganic forms
(except for some volatile chlorides) of U, V, Co,
Cr, Fe, Mn, and Ti are essentially non-volatile.
However, significant small-particle association is
observed. Manganese, Fe, and Cr are major
components of stainless steel which is used in many
parts in contact with the gas stream. Abrasion may
be responsible for some of the small-particle
component observed for these elements. Elements in
Group III (i.e. rare earths, Sc, Al, Mg, and Na) are
associated with nonvolatile mineral phases found in
coal, such as clay minerals. Particle size
distributions observed for these elements, shown in
Fig. 7, are similar to the total mass distributions.
Parameters characterizing large-particle
distributions for elements listed are also reported
in table 2. Mass median diameters for elements in
the three groups appear to decrease in the order
M_Q-Group I < M_0~Group II < M^-Group III: the
averages for the three groups being 0.98 ± 0.07,
1.10 ± 0.03, and 1.17 ± 0.05, respectively. Given
the uncertainty in individual determinations
(estimated to be 5-10?), these differences are
probably insignificant. As noted above a smaller
M5q for the volatile element would result if the
element were surface adsorbed. For group III
elements the corresponding S.Q is 0.9l nm. The
observed !¦'„ for Group I elements lies midway
between the S^q and M,-0 values for Group III
elements. Therefore the Group I elements may in
part be adsorbed on the surface of large particles.
The differences in observed M5Q values for the three
grov®s may well be a distortion resulting from
collection of small, particles (which should be
collected on the back-up filter) on the last few
impactor stages.
Enrichment Factors for Stack Aerosols
Comparison of emission data is facilitated by
computation of an enrictar,erA factor relative to
input coal (EF). In this work EF is defined as the
ratio of the concentrations of an element and Sc in
stack fly ash, divided by the corresponding ratio in
coal. Presented in this manner the stack emission
data are independent of concentration variations in
coal.
Enrichment factors relative to coal for the
plant., described above and for the Chalk Point
Plant for 20 elements are shown in Fig. 8. Values
reported for both plants represent averages of
enrichment factors computed for individual samples.
For many of the elements, enrichment factors of < 1
are observed. These elements, including Al, Mn, Ti,
Sm, Na, Th, Ta, and La are the nonvolatile elements
found in groups II and III of table 2. Many of the
group I and II elements are significantly enriched
(i.e. > l). The enrichments observed are due
primarily to the fractionation of elements as a
function of particle size, and the greater
collection efficiency of the electrostic
precipitator for the large size particles. Elements
with EF < 1 are apparently on particles which are
more efficiently trapped by the electrostatic
precipitator than those bearing Sc. From Fig. 8 it
is evident that for several elements, including V,
Ga, Ba, Se, and Cr, large differences (as large as a
factor of H) exist between EF values observed at the
two plants. The largest differences are observed
for the volatile elements which also tend to be the
most toxic.
Conclusions
It has been shown that significant errors may
result from the use of manufacturer calibration data
for cascade impactor samplers used in stack
sampling. Possible sources of error are bounce-off
and reentrainment, significant stage collection
efficiencies for particles outside the desired
collection range, and resultant overestimation of
large particle fractions. Errors associated with
these problems are reduced by observation of actual
stack samples using SEM techniques. Since accurate
knowledge of particle size distribution is necessary
for the evaluation of lung deposition, and since
significant fractions of volatile elements were
observed on small particles collected on the back-up
filter, we conclude that SEM particle sizing is
essential to source emission studies. Enrichment
factors for two coal-fired power plants are
significantly different, particularly for volatile
elements. These differences are attributed to: 1.
differences in the chemical composition of the
coals, 2. differences in combustion and stack
temperatures, and 3. differences in efficiencies of
particulate control equipment. More studies are
needed in each of the three areas in order to
adequately assess the impact of trace element
emissions from coal-fired power plants.
REFFEBEHCES
1. Bertine, K.K., Goldberg, E.D., "Fossil
Fuel Combustion and the Major
Sedimentary Cycle," Science 173, 233
(1971).
3
17-2
-------
2. Gordon, G.E., Zoller, W.H. and Gladney,
E.S., Abnormally Enriched Trace Elements
in the Atmosphere, Trace Substances in
Environmental Health-VII. D.D. Hemphill,
ed., (Univ. of Missouri, Columbia, 1973)
pp. l6l-nk.
3. Harrington, R.E., Fine Particulates -
The Misunderstood Air Pollutant, J. Air
Poll. Cont. As so., 2h, No. 10, (19?lt),
h. Ragaini, R.C., Kalston, R. , and Garvis,
D., Trace Element Analysis at the
Livenaore Pool-Type Reactor Using
Neutron Activation Techniques,
University of California, Lawrence
Livermore Laboratory report, UCRL-51855,
(19T5).
5. Gunnink, R., and Niday, J.B. , "The
Gamanal Program," University of
California, Lawrence Livermore
Laboratory Report UCRL-51Q61, Vol. 1-111
(1973)•
6. Ondov, J.M,, Zoller, W.H., Olmez, I.,
Aras, N.K., Gordon, G.E., Rancitelli,
L.A., Able, K.H., Pilby, R.H., Shah,
K.R. , and Ragaini, R.C, , Elemental
Concentrations in the national Bureau of
Standards' Environmental Coal and Fly
Ash Standard Reference Materials, Anal.
Chem., itX, No- (1975).
7. Herdman, G., Small Particle Statistics.
Academic Press, Inc., Buttervorth & Co.,
Limited (i960), p. 85.
8. Davison, R.L. , Natusch, D.F.S. , Wallace,
J.R., and Evans, Jr., C.A. , Trace
Elements in Fly Ash, Dependence of
Concentration on Particle Size, Environ.
Sci. and Technol., 8, No. 13 (197^).
9. Kaakinen, J.W., Jorden, R.M., Lawasani,
M.H., and West, R.E., Trace Element
Behavior in a Coal-Fired Power Plant,
paper presented before the Division of
Environmental Chemistry, American
Chemical Society, Philadelphia, (1975),
10. Gladney, E.S., Trace Element Emissions
of Coal-Fired Pover Plants: A Study of
the Chalk Point Electric Generating
Station, Ph.D. Thesis, University of
Maryland, (1975).
"y
ash
Hopper
»sh
Bottom
ash
Stack
Furnace
Ambient
sir
Emission
control
system I
Input
coor
[f©
Transport
models
Aerosol
physics
and
chemistry
F1g, 1. Inhalation pathway for dose-to-man
assessment of trace element emissions
from coal-fired power plants.
k
1000
mo
200
100
s»
4)
"i
3
Z
Diameter (pm)
Fig. 2. Number-size distribution of particles
collected on a back-up filter of a 2 hour
Impactor sample (Run #8).
0.2 0.5 1.0 4.0
Diameter Urn)
Cumulative mass distribution of particles
collected on a back-up filter of a 2 hour
•Impactor sample (Run #8).
17-2
-------
Table I. Comparison of stage constants observed for impactor samples collected
in-stack from a modern coal-fired power plant with those obtained using
manufacturer provided data.
Observed3 Quotedb
Stage N50 o£ d50
1
1.23
1.35
26
2
1.26
1.37
11
3
1.99
1.30
5.3
4
2.13
1.34
2.2
5
1.30
1.42
1.2
6
0.72
1.39
0.63
7
0.32
1.33
0.35
aNumber median diameters (d50) determined by SEM observation on an impactor sample
collected in-stack at a flow rate 0.38 cfm (102% isokinecity) for a 2 hr period.
b5W cut-off diameters determined from calibration curves provided by the
manufacturer, after correction to an assumed particle density of 2.2 g/cnw.
li 1 111
0.1
q 1 1—1 1 1 1 1 11 r-
Observed
Quoted ~
A Quoted curve 1.28 3.54
O Observed curve 1..14 2.03 -
1 1 1 1111
_i 1 L.
-LLiJ L.
1.0
Diameter (pm)
Fig. 4. Cumulative mass distribution of stack
particulates as a function of observed
and quoted impactor stage parameters.
10 20
100,000
10,000 -
1000 —
0.02 0.05 0.1 0.2 0.5 1.0 2.0
Diameter (//m)
Fig. 5. Number-size distribution of particles
collected on a whole filter sample.
5
17-2
-------
Table 2. Relative amounts of elements observed on small and large particle
distributions in-stack fly ash.
Small Distribution Large Distribution
(0.08 vim < m50 < 0.09 )im
Element Fraction (%)±oa Fraction (%) ±a Mun (um) eg
•-.-wi *r—/
Group I
I
62 ±
8
38
+
12
0.93
2.00
Ba
55 t
9
45
+
11
0.92
1.93
Br
62 ±
29
38
±
12
0.91
2.46
As
57 ±
7
43
±
12
0.93
1.93
Sb
49 ±
2
51
±
5
1.05
1.98
Mo
44 ±
8
56
±
9
1.03
2.03
U
43 ±
7
57
±
6
ND
ND
Hg
40 ±
9
60
+
5
1.01
1.94
V
31 ±
6
69
±
9
0.95
1.96
Ga
29 ±
7
71
+
15
1.00
1.85
Se
27 ±
3
73
±
9
1.11
1.89
Zn
18 ±
2
82
+
10
1.01
2.03
CI
11 ±
9
89
+
45
0.85
2.47
Group II
Co
20 ±
3
80
±
9
1.13
1.87
Cr
20 ±
2
80
+
6
1.10
1.84
Fe
10 ±
3
90
+
6
1.06
1.86
Mn
8 ±
4
92
±
12
1.09
1.87
T1
6 ±
3
94
±
5
1.11
1.87
Group III
Sm
< 9
91
+
15
1.20
1.83
Ta
< 6
94
±
8
ND
ND
Th
< 6
94
±
8
1.17
1.88
Sc
< 5
95
±
5
1.10
1.79
Mg
< 14
86
+
26
ND
ND
Na
< 5
95
+
11
1.15
1.77
In
< 6
94
+
24
1.13
1.78
A1
< 6
94
+
12
1.20
1.93
La
< 6
94
+
9
1.24
1.87
a0 represents the dispersion of Individual measurements or analytical
uncertainty, whichever is larger.
6
17-2
-------
Table 3. Enrichment factors for elements on suspended particles with respect
to input coal at two coal-fed power plants.
Element
This
Work3
Chalk
Point''
Ba
2.6
0.9
0.92
0.08
As
7.3
3.2
6.3
1.7
Sb
6.7
1.6
4.0
0.9
V
3.8
1.0
0.72
0.06
Gd
3.1
0.7
1.0
0.2
Se
2.5
1.1
5.7
1.3
Zn
6.4
4.4
2.9
0.6
Co
2.2
0.3
1.0
0.2
Cr
2.0
0.3
0.92
0.08
Fe
1.2
0.1
0.83
0.33
Mn
0.66
0.08
1.3
0.2
Ti
0.54
0.14
0.78
0.14
Sm
0.83
0.16
0.69
0.07
Ta
0.95
0.40
0.66
0.02
Th
0.91
0.10
0.80
0.03
Sc
= 1.0
; 1.0
Mg
0.42
0.26
1.2
0.3
Na
0.85
0.17
0.92
0.17
A1
0.41
0.06
0.83
0.08
La
0.98
0.09
0.71
0.05
Values reported are averages of individual runs. Uncertainties are the
standard deviation of replicate determinations or the larger analytical
uncertainty of individual measurements, whichever is larger.
^Values reported here were obtained from Gladney^ and renormalized to Sc.
7
17-2
-------
*i 50
" 0.1
+j
u
9.0
8.0
7.0
6.0
5.0
0.06
Fig. 6.
1000,
0.5 1.0 2.0 3.0
Diameter (urn)
Typical cumulative mass distributions
for As and A1 based on observed d50s.
ScX 10J
t-*-\
10b
C
0)
u
c
o
o
1 nI In. I 1 1—I 1.'¦ 11 L
f* 7 6 5 4 3 2
„4i ,i 1.1
01
§ 4.0
3.0
2.0
1.0
0
-i—i—i—r~!—r—i—r~i—i—i—i—r~i—r~i—i—(—i—r
• 4
A -J
• •
A
' 111111)' I11'l' l' 1' {¦' I11' 11111' i'j Y 1
A1 I Mn ) Na I Ta j Sc | Cr I Se | Ga | in | As
T1 Sm Th La Fe Co Ba V Sb
Elements
Fig. 8. Enrichment factors for elements observed
in stack aerosols at the power plant
described 1n this study and at the Chalk
Point Plant,
Impactor stage
F1g, 7. Impactor distributions for several
elements with large particle association.
-------
EVALUATION OF EFFECTS OF MULTIPLE POWER PLANTS
ON A RIVER ECOSYSTEM
G.C. Slawson, Jr.
B.C. Marcy, Jr.
Ecological Sciences Division
NUS Corporation
1910 Cochran Road
Pittsburgh, Pennsylvania 15220
Introduction and Summary
The evaluation of environmental impacts of power
plant development activities is not only a legal re-
quirement but also a logical necessity. This paper
discusses the impact of power plants on a river, eco-
system. The relative magnitude of these effects as
related to plant siting and to plant design factors is
presented. The biological components considered are
gro-ups of organisms selected to allow interpretation
as to the desirability of induced changes. A biologi-
cal data set is outlined which allows identification
of alternative site locations and designs to minimize
the effects of entrainment of organisms in cooling
water intakes. Alternative plant locations and opera-
tional schemes are further evaluated by projecting the
impact of waste heat discharges on the river biota.
These projections are in the form of impact profiles
of the power plant-river system. A methodology for
defining the overall impact of alternative basin power
plant systems is outlined. These assessment procedures
provide information useful for the siting of new plants
and for the delineation of cumulative impacts of exist-
ing and proposed power plant systems.
Perspective and Approach
The approach discussed herein addresses a region
or basin-wide evaluation of the cumulative effects of
multiple power plant operation. Implementation of the
evaluation framework will provide a biological basis
for evaluating alternative power plant sites and
designs. For alternative schemes, impact profiles
which estimate the magnitude and spatial extent of
power plant effects on aquatic biota are produced. An
example of such profiles is shown in Figure 1. Impact
descriptions of this sort provide information to en-
able the assessment of the impact imposed by the opera-
tion of a given plant on the basin as a whole and to
assess the cumulative impacts of the operation of
multiple power plants. In this manner, zones of
appreciable impact may be delineated within the river
system for various operational, siting and plant system
configurations.
Obvious uses for such a framework include plant
siting and the administration of regulations such as
Sections 316(a) and 316(b) of the Federal Water Pollu-
tion Act Ammendments of 1972. These regulations
require demonstration of the use of best available
technology in the placement, design and operation of
cooling systems and intake configurations. It must be
demonstrated that no appreciable harm to aquatic biota
will result. The methodology may also be used to
design operational schemes and schedules for power
plant networks to minimize operational effects during
biologically critical periods such as spawning and
developmental periods. By formalizing the environ-
mental assessment process, alternative systems can more
easily be considered and data voids which preclude
cognizant decision making would be readily identified.
Important Organisms
A key element in the design of a system for evaluating
1
probable environmental impacts is the definition of
the groups of important organisms to be considered.
The character of these ecosystem components dictates
the utility of the evaluation results. Commonly,
impact evaluations are based upon important groups of
organisms defined, for example, as (1) commercially
important species; (2) recreationally important
species; (3) rare and endangered species; and (4)
species whose role in the food web is important to the
welfare of species in groups 1, 2 or 3 and, in
general, the stability of the food web. The difficulty
in characterizing important biotic groups in this
manner is that virtually every species falls into one
of these categories. Further, impact projection and
interpretation at the species level is in many in-
stances not possible and may not even be required.
Although there is no clear cut answer to this defini-
tion problem, a basic criterion is offered: to con-
duct an impact assessment, important organism groups
should be defined for which interpretations as to
water quality or the health or the desirability of the
system may be made. In operational terms, one would
define a biological group as important if one could
attach a label of "deleterious" or "not detrimental"
to changes in that group of a given magnitude in a
given direction.
Such a criterion places no significant limits on
the taxonomic makeup of a particular group. For
recreationally or commercially important organisms, a
group may, and many times will, consist of a single
species. As a general rule, only some fish and some
benthic invertebrates (such as shellfish, crabs,
lobsters, and the like) will fall into thege single
species groupings. For other organisms, larger
taxonomic (or other type) groupings are possible and
are desirable.
Intake Effects as Siting Factors
The interface between power plant systems and
aquatic ecosystems is defined by the amount of waste
heat rejected to the receiving stream and the cooling
water flow rate. These characteristics determine (1)
the water temperature change through the plant,
(2) the probability of entraining organisms in the
plant cooling system and (3) the likelihood that en-
trained organisms are killed. Probably the most
important biological criteria in the siting of power
plants is the reproductive capacity of the river biota
of the system under consideration. The greatest
opportunity for minimizing environmental degradation
is in the selection of new power plant sites and in
the evaluation of alternative plant design and opera-
tion schemes. For achieving this minimization of
impact, the delineation of spawning, nursery, feeding,
wintering areas, or other areas of seasonally high
concentrations of individuals of important organism
groups is critical.
Non-screenable plankton or small fishes which
pass through the intake screens .intact are susceptible
to various and simultaneous stresses, often leading to
high in-plant mortality. Entrainment losses are
potentially more detrimental to fish populations than
17-3
-------
SYSTEM A
SYSTEM B
o, 40H
9 .
UJ
c
Z3
<
cc
UJ
0.
2
UJ
H
38
36
34'
32H
30
CO
ac
UJ
+100-
til
o
o
z
D
g
*50-
<
X
CL
0-
o
>-
3«
or
3
-50-
5
Q.
-100'
; >l
M»' - U-
O -50
• BLUE-GREENS
GREENS
DIATOMS
TOLERANT
ARTHROPODS
INTOLERANT
ARTHROPODS
W
25
x o -50 H
OTHER
INVERTEBRATES
ZOOPLANKTON
ROUGH FISH
SPORT FISH
FORAGE FISH
-25'
75
-100
DISTANCE (MILES) DISTANCE (MILES)
FIGURE X
WATER TEMPERATURE AND BIOLOGICAL IMPACT PROFILES SHOWING
ESTIMATED CHANGES IN AQUATIC BIOTA INDUCED BY
WASTE HEAT DISCHARGES FROM
HYPOTHETICAL POWER PLANT SYSTEMS A AND B
2
17-3
-------
to organisms of lower trophic levels. This is due to
the slow regeneration capacity of the fish community
relative to other organism groups. The faster the re-
generation capability, the more stress the community
can withstand. Regeneration times for fish commonly
range from 2 to 4 years for estuarine species and 4 to
6 years for freshwater species. However, phytoplank-
ton populations may be re-established within several
hours from the time of passage through a plant. Re-
generation time for crustacean zooplankton may be on
the order of a few days.
Normally, at the initial stages of a site evalua-
tion study, detailed information on spawning activity
in various river reaches is not available. However,
it is possible in many cases to use existing litera-
ture and perhaps preliminary field surveys to delin-
eate areas of potentially deleterious impact. A
primary task is the definition of the physical and
hydrologic characteristics of the river basin with re-
gard to critical existing or potential spawning
habitats such as shallow weedy areas, sand bars, back-
water areas and tributaries. Most fish spawning takes
place in tributaries and back areas off the main
stream of larger rivers. Intake and discharge struc-
tures should be located away from potential critical
spawning habitats and in areas where the early develop-
mental stages are less vulnerable to entrainment.
Entrainment potential is diminished at sites located
in nursery areas. Here, fish are of a larger size,
can more easily avoid entrainment and can be handled
by an adequate fish return system.
Another initial site selection consideration is
the compilation of a lift of important fish groups
found in the river system. This may be accomplished
from existing literature or an initial survey. Such
groups will usually be single species. Thus, for this
discussion, "important fish species" will be synony-
mous with "important fish group".
Data on reproduction characteristics provide in-
dications as to the potential for detrimental impacts
for alternative sites and plant systems. Information
required for this assessment include fecundity (number
of eggs carried by a female), type of spawning
(dimersal or in the water column), incubation and
development time, longevity, regeneration time and
other factors related to the vulnerability of
important fish species to induced stress. This type
of information is generally obtainable from existing
literature.
Fecundity estimates indicate the ability of a
population to absorb plant induced losses. If most of
the species present have a high fecundity, the poten-
tial for entrainment impact will be diminished. Such
populations will have a high rate of recruitment and
thus can endure a high spawning product mortality.
Plants should be located away from potential spawning
areas of important species with low fecundities.
Incubation and development times indicate the time
period of exposure to entrainment damage. If, for
example, the majority of species have a short egg to
larval development period in a particular area, their
vulnerability to entrainment is reduced and the prob-
ability of survival increased. The development time
can be critical in determining how far a developing
planktonic stage may travel and thus be exposed to
plant effects in the river basin. Such considerations
aid in the evaluation of the effect of locating plant
intakes at various distances downstream from spawning
zones.
The age at maturity and longevity of fish popula-
tions are also important factors in siting evaluations.
A community made up of slow growing species usually
indicates that they mature at an older age. This type
of community would be more vulnerable to long-term
stress induced by plant operation. Impacts will be
mitigated at sites located in areas of the river basin
where there is a high percentage of fast-growing (e.g.,
forage fish) species. Such areas exhibit rapid turn-
over rates and generally high fecundity.
Beyond the geographic placement of intakes,
alternative plant designs produce widely varying en-
trainment potentials. The number of organisms that are
entrained can be reduced by designs which reduce cool-
ing water intake flows. This can be accomplished by
either (1) increasing the At through the cooling system
or. (2) using cooling towers. The relationship between
cooling water flow, plant size, and condenser AT for
both fossil and nuclear plants is shown in Figure 2.
Two options exist for the first of these cooling system
alternatives. One option is to limit the cooling water
flow to the upper temperature tolerances of the
entrained speces. A useful guide for selecting levels
of entrainment mortality from lethal temperature data is
illustrated in Figure 3. In this mariner, one may seek
to minimize both intake volume (and thus, numbers en-
trained) and the probability of mortality of entrained
organisms. It has recently been shown that AT's up to
1 2
10 and 12°C ' were not detrimental to egg development
or hatching success of many fish species. Schubel^
has also indicated that even higher AT's are possible
without killing 50% of the entrained fish.
NUCLEAR, 33%,
IN-PLANT LOSSES = 5%
FOSSIL, T),*40%,
IN-PLANT AND STACK LOSSES -15%
2500
2000
t 1500
s
Z 1000
5
s! 500
0 1000 2000 3000
COOLING WATER FLOW (Q), cfs
AT = CONDENSER TEMP. RISE
« PLANT THERMAL EFFICIENCY
FIGURE 2
RELATIONSHIP BETWEEN AT AND COOLING WATER FLOW
FOR VARIOUS POWER PLANT TYPES AND SIZES
3
17-3
-------
10(86)
LEVEL
H0RTM.lt *
«J ZSfT7)
20166)
isT5o! isiiii lofgi zsTrn io««)
AMBlEKT timper*ture «c(°n
FIGURE 3
mortality of emtrained organisms as related
TO TEMPERATURE OF COOLING WATER FLOW
The second option, and perhaps the most viable
based upon recent data, is to assume 100% mortality of
entrained organisms in the cooling system. This has
been found to be generally true for cooling towers.
Numerous studies over the past two years have shown
that mechanical damage from abrasion and shear stress,
not heat, was the highest single cause of entrainmen!
mortality, resulting in nearly 100% mortality for
4 5 6
raost entrained ichthyopiankton. ' ' The low volume
concept is presently the best technology available for
minimizing the adverse effects upon entrained organ-
isms.
With the river basin descriptions, species repro-
duction data and plant operation information, the
entrainment effects of a given plant operation on a
given spawning zone may be evaluated as follows:
Proportional
Entrainment
Loss
Proportion
Present in
Vulnerable
Stages
Proportion
Being
Entrained
Proportion]
Killed
When
[Entrained
such as that proposed'by Lawler,' can be utilized to
estimate resultant losses to adult populations.
The swimming speeds of fish species is another
important siting criterion. If, for example, the
majority of species in the river basin community have
low swim speeds, they will be more vulnerable to im-
pingement on intake screens. Fish swim speed data aids
in 1) indicating what species might be vulnerable to
impingement and 2) define upper limits of intake
velocity. Detailed location and design criteria are
presented by the Lake Michigan Cooling Water Intake
Technical Committee,® USEPA,9'and the American
Nuclear Society
11
Thermal Discharge Effects on Biological Systems
Those considerations dismissed above represent a
guide for the identification of alternative site
locations and plant designs. The impact of waste heat
discharges of these alternatives is considered in the
following sections.
Primary Producers
Primary producers form the base of the aquatic
food web, providing materials for productivity at all -
other trophic levels. In addition, algae are" useful
indicators of water quality as shifts in algal com-
munity size and composition have been associated with
induced environmental alternation.
To enable an evaluation of algal community com-
position, the algae are classified into groups which
are relatable to the quality of the aquatic system.
12
Canale and Vogel summarize an extensive literature
review on temperature effects on algal growth using
four algal groups: diatoms, greens, blue-greens and
flagellates. The results of their review are shown in
Figure 4. These data were used to formulate a model
to estimate changes in algal growth due to changes in
water temperature. The specific growth rates shown
in Figure 4 were utilized in the following relation:
V [T] - y [TJ
u [Ta]
This relationship estimates the proportion of the
production of a' spawning area lost by entrainment
within a downstream plant. The proportion of organ-
isms present in vulnerable stages (i.e., fish eggs and
larvae in a planktonic stage of development) is a
function of the distance between the plant and an up-
stream spawning zone at a particular time of year. A
closer proximity results in a higher probability of
presence during the spawning season. The proportion
being entrained is estimated by the ratio of the plant
cooling water intake rate to the river discharge rate.
The proportion killed once entrained is related to
numerous plant operation characteristics (At, pressure
changes, abrasion, biocide vise, etc.). If available,
information on densities of eggs and larvae can be
used to estimate absolute entrainment losses. These
estimates, used in conjunction with population models
where Ay is the change in the growth rate of the algal
group, vi CTl is the growth rate at temperature T, and
T is the ambient or normal water temperature. The
Si
model form presented above is supported by the exten-
12
sive literature reviewed by Canale and Vogel and by
the results of numerous studies conducted at power
plants.
Primary Consumers
Aquatic invertebrates are the major link in the
aquatic food web between primary producers and higher
trophic levels such as fish. In addition, these groups
(and particularly nonmotile groups) have commonly been
proposed as indicators of water quality. For this
discussion, primary consumers will include invertebrate
herbivores as well as many primary carnivores.
Knowledge of the response of the intermediate food
web levels to thermal stress is not as well developed
as is that for the primary producers and for the fish
groups. Some generalizations can be made concerning
the thermal tolerance of macroinvertabr&tes. Parker
13
14
and Krenkel, " Bush et al,*"" and others have made the
observation that the heat tolerance of invertebrates
A
17-3
-------
DIATOMS
¦— GREENS
—•BLUE-GREENS
FLAGELLATES
i
>,
o
a
Ui
x
i-
3
tc
o
40
24
TEMPERATURE (°C)
FIGURE 4
ALGAL GROWTH AS A FUNCTION OF TEMPERATURE
is commonly above that of fish. Thus, using tempera-
ture relationships such as those proposed in the
following section for fish groups to describe the re-
sponse of the invertebrate groups grazed upon by the
fish should produce conservative estimates of inverte-
brate response. Such approximations can be improved
by using available information on dominant species or
other groupings identified for a given ecosystem.
Available literature reviews (such as references 15,
16, 17 and 18) provide source material for expressing
these relationships for invertebrates.
Fish
Many criteria have been proposed which address
the effects of thermal stress on fish populations.
Many of these define a single temperature level either
as permissible or as lethal. The methodology used
here combines information on "no effect" levels and
lethal levels to define a function for expressing
levels of reduction for various temperature states.
The reductions considered here are relative to a de-
fined optimum temperature.
Figure 5 illustrates the form of the model used
to evaluate reductions or the elimination of fish
groups due to thermal stress. A relationship of this
form is defined for each fish group being considered.
This relationship was formulated by defining (1) a
maximum temperature resulting in no deleterious effect
(an optimum temperature) and (2) a temperature at
which heat effects would eliminate the group of fish
(elimination temperature)
The proposed approach uses data from the litera-
ture (such as references 13, 17, 19, 20, 21 and 22)
to define optimum and elimination temperatures. For
19
example, National Academy of Sciences lists optimum
and ultimate incipient lethal temperatures for several
species of fish. The ultimate incipient lethal tem-
perature is defined as the "breaking point" between
the highest temperature to which the organism can be
acclimated and the lowest of the extreme temperature
that will kill the warm-acclimated organism. This
level may be defined as an elimination temperature by
inferring that, above this level, fish will be elimin-
ated due to either avoidance behavior or through
lethal effects.
The temperatures used as optimum and elimination
2 I00-)
SMALLMOUTH BASS-^
LARGEMOUTH BASS /
80-
60-
40-
CHANNEL CATFISH
40
30 35
TEMPERATURE (°CJ
FIGURE 5
TEMPERATURE RESPONSE FUNCTIONS FOR THREE
WARMWATER SPORT PISH
temperatures will, of course, be a function of season.
Various life cycle activities such as reproduction and
development are seasonal and have different tempera-
ture requirements than for growth or survival of
adults, Thus the form of a function as described in
Figure 5 will vary with the season and the effects
imposed by a given plant-river system configuration
will vary with seasonal ambient water temperatures.
During a spring spawning period, for example, optimum
and elimination temperatures would indicate the range
of temperatures for successful spawning and early
development.
Data on fish swim speeds can also provide guidance
for discharge design through consideration of velocity
- AT relationships in the mixing zone. Determining
what species and life stage may penetrate the heated
effluent and for what period of time it may reside in
a particular isotherm can be of critical importance.
Generally, the greater numbers of fish that can main-
tain themselves at a higher AT isotherm for extended
periods of time, the greater the potential impact.
Food Web Model
The preceding sections have addressed the direct
response of aquatic biota at various trophic levels to
induced temperature stresses. Changes in the magnitude
and composition of food material produced at one
trophic level may result in changes at higher trophic
levels. The biological groups defined as important
are components of a model of food web relationships.
Interconnections among these components are based upon
conversion efficiencies between trophic levels and
variation in preference for food materials. The con-
cept of preference encompasses grazing selectivity and
grazing success (in the sense of satisfying nutritional
requirements).
The efficiency of conversion between trophic
levels is widely discussed in the literature. Though
the actual amount of energy and material transferred
between trophic levels varies greatly, conversion
efficiencies of 10 to 20 percent are commonly given in
the literature.
Preference factors are used in conjunction with
estimates of conversion efficiency to estimate the
relative amount of change at one trophic level induced
5
17-3
-------
by changes at lower trophic levels. Preference
factors weight the feeding activity of grazers among
various food materials.
For illustration, consider cladocerans as a graz-
ing group and suppose the available food groups are
diatoms, green algae, and blue-green algae. Consider
the following preference factor assignments:
Food Group Preference Factor
diatoms .3
green algae .5
blue-green algae .2
These factor assignments would indicate highest pre-
ference for green algae and lower desirability (at
least for cladocerans) of diatoms and blue-green algae
as food materials. Given relative changes of -20%,
-15% and +15% for diatoms, green algae and blue-green
algae respectively, the resultant change (Ax) in
cladocerans may be estimated as follows:
Ax = [(-.20)(.3) + (-.15>(.5) + (.15) (.2)] (.15) =-0.02
where the last 0.15 term is the estimated conversion
efficiency from algae to cladocerans. This result in-
dicates a 2% reduction in cladocerans due to induced
¦changes in the size and composition of algal popula-
tions. In river systems, detrital material of
terrestrial and aquatic origin is an important food
source to many organisms. Thus the relationship pre-
sented, which does not consider this food source,
produces conservative (i.e., high) estimates of impact.
Assessment of Thermal Effects
To illustrate the use of the concepts on thermal
effects discussed in the preceding sections, a hypo-
thetical comparison was constructed. Important
organism groups were designated as follows:
Diatoms
Green Algae
Blue Algae
Tolerant Arthropods (midges)
Intolerant Arthropods (may-
flies, caddisflies)
Other Invertebrates (oligo-
chaetes, nematodes,
molluscs)
Zooplankton (rotifers, proto-
zoans, planktonic
arthropods)
Forage Fish (minnows, sunfish,
shiners)
Rough Fish (carp, gizzard
shad, bullheads)
Sport Fish (warm water
species: bass, channel
catfish)
Profile outputs for two hypothetical river-plant
systems are depicted in Figure 1. A matrix of feeding
preference factors was formulated to define the nature
of the links in a food web model. Conversion effi-
ciencies of 20* were used for all feeding activities.
The example case considered is for a single time
period, a high temperature period for the hypothetical
river system. An ideal application would consider at
least the four seasons so that key seasonal biological
activities such as spawning and migratory movements
could be addressed.
For the two plant-river systems, longitudinal
temperature profiles were produced using methods given
23
by Edinger and Geyer. This information was used with
relationships of biological response discussed earlier
to project longitudinal profiles of temperature-induced
biological effects. Such a graphical presentation is
certainly one method for comparison of alternatives.
However, such a comparison methodology would probably
be of limited utility to the lay public and govern-
mental decision makers. For the technical expert, such
impact profiles provide valuable information on the
effects of plant-induced temperature stresses. How-
ever, a graphical presentation is cumbersome for the
comparison of numerous alternative river-plant con-
figurations.
To summarize the information found in the set of
profiles shown in Figure 1, a critical change level
was first defined. This level is that above which a
change is sufficient to be considered important; i.e.,
smaller changes are labeled as insignifcant. For the
hypothetical example presented here, a critical change
level of 10 percent was used. This operational defini-
tion of a significant impact is largely a matter of
professional judgment.
For each biological group or category, an exceed-
ance mileage (the distance [in river miles] over which
the estimated change exceeds the defined critical
change level) was measured. This defines the extent
of zones of significant impact. To avoid the use of
cumbersome large numbers, these exceedance mileages
are divided by the total river miles in the system
being evaluated. This defines the relative or pro-
portional extent of the zones of significant change.
Because of the criterion used to define the biological
groupings, signs can be given to these relative im-
pacts: positive (+) for a non-detrimental or insigni-
ficant change and negative (-) for a deleterious
change. The results of this analysis are presented in
Table 1 for the hypothetical case comparison in Figure
1. Summing the non-detrimental (+) and detrimental (-)
relative changes results in a measure of overall impact
for the two systems. In this case, System A is indi-
cated to be the poorer of the two systems since its
overall impact is more negative (i.e., more detri-
mental ).
This methodology defines an index of overall
environmental impact for the river reach considered.
It is based upon the definition used here to define the
important biological groups used in the food web model.
Although this point will certainly be subject to dis-
cussion, because all of the defined groups are impor-
tant, then the summation of relative changes in each
group may be used to express the amount of total impact
without the use of inherently subjective weighting
factors for combining the effects on individual groups
into an overall impact indicator.
The thermal stress evaluation scheme described
above is only one segment of the process of assessment
for the siting and operation of power plants. Other
components include evaluation of the ecological reason-
ableness of the best of the alternatives considered
and consideration of site and plant specific factors
which define the nature of the plant system-biotic
system interface. For example, consider the comparison
of Systems A and B just presented. Clearly System B is
the better of these two choices. However, System B is
projected to result in considerable reduction of the
sport fishery in this river reach and promote a dom-
inance of tolerant species. As such, even System B
does not appear very environmentally desirable. Pro-
fessional Judgment, experience, and other non-quanti-
fiable factors play an important role in such determin-
ations. Models such as that proposed offer a
6 17-3
Primary Producers:
Primary Consumers •-
Fish:
-------
TABLE 1
Extent of Relative Impact for
Hypothetical Example Systems A and B
Illustrated in Figure 1
Biological Group
Extent of Relative Impact
System A
No
Change
Deleterious
Change
System B
No
Change
Deleterious
Change
Diatoms
Green Algae
Blue-green Algae
Tolerant Arthropods
Intolerant Arthropods
Other Invertebrates
Zooplankton
+0.00 -1.00
+0.12 -0.88
+0.00 -1.00
+0.70 -0.30
+0.00 -1.00
+0.40 -0.60
+0.14 -0.86
+0.S0 -0.50
+1.00 -0.00
+0.60 -0.40
+0.04 -0.96
+0.40 -0.60
+0.62 -0.38
+0.50 -0.50
Forage Fish
Rough Fish
Sport Fish
Sub-Totals
TOTAL IMPACT LEVEL
+0.10
+0.00
+0.00
+1.46
-0.90
-1.00
-1.00
-8.54
+0.54
+0.00
+0.00
+4.20
-0.46
-1.00
-1.00
-5.80
-7.08
-1.60
^Estimated change less than defined critical change level; 10 percent used here as critical change level.
capability for making quantitative estimates of impact
and thus provide valuable information for these more
subjective evaluations. Projection of potential en-
vironmental impacts at the preliminary stages of power
system development allows the earliest possible identi-
fication of ecological problems. With this "early
warning" capability, costly delays in design and
licensing may be avoided. Mathematical methodologies
offer relative ease for comparing the effects of
various siting, design and operational schemes. As
operational definitions of such characteristically
subjective terms as significance of impact, stability,
and desirability are made, they can be readily adapted
to these quantitative methodologies.
Literature cited
1. Schubel, J.R. 1973. Effects of exposure to time-
excess temperature histories typically experienced
at power plants on hatching success of fish eggs.
Chesapeake Bay Instit., The John Hopkins Univ.
Spec. Rep. No. 32-PPRP-V, Ref. No. 73,11.
2. Hoss, D, E., W. F. Hettler, Jr., and L. C. Caston.
1973. Effects of thermal shock on larval estua-
rine fish-ecological implications with respect to
entrainment in power plant cooling systems. Paper
presented at the Proceedings of the Symposium on
the Early Life History of Fish, Oban, Scotland.
3. Schubel, J.R. Personal Communication.
4. Marcy, B. C., Jr. 1971. Survival of young fish
in the discharge canal of a nuclear power plant.
J. Fish. Res. Bd. Canada 28:1057-1060.
5. . 1973. Vulnerability and sur-
vival of young Connecticut River fish entrained at
a nuclear power plant, J. Fish. Res. Bd. Canada
30(8):1195-1203.
6. Marcy, B.C., Jr. 1975. Entrainment of organisms
at power plants, with emphasis on fishes—an
overview. In: S.B. Saila (ed.) Fisheries and
Energy Production: A Symposium. Lexington Books,
D.C. Heath and Co., Lexington, Mass. pp. 89-106.
7. Lawler, J.P. 1972. The effect'of entrainment at
Indian Point on the population of Hudson River
striped bass. Testimony before AEC (Docket No.
50-247) for Consolidated Edison Company, Indian
Point Station, Unit No. 2.
8. Lake Michigan Cooling Water Intake Technical
Committee. 1973. Lake Michigan intakes: report
on the best available technology. 45 pp.
9. U.S. Environmental Protection Agency. 1973a.
Reviewing environmental impact statements - power
plant cooling systems, engineering aspects. Envir.
Prot. Tech. Sec. EPA-660/2-73-016. 93 pp.
10. U.S. E.P.A. 1973b. Development document for
proposed best technology available for minimizing
adverse environmental impact of cooling water
structures. EPA 440/1-74/015. ,
11. American Nuclear Society. 1974, Entrapment/
Impingement: Guide to Steam Electric Power Plant
Cooling Systems Siting, design and operation for
controlling damage to aquatic organisms at water
intake structures. Draft standard N223, December
13, 1974. 24 pp.
12. Canale, R.P. and A.H. Vogel. 1974. Effects of
temperature on phytoplankton growth. J. Env. Eng.
Div., Proc. ASCE. Vol. 100, No. EE1, pp. 231-241.
13. Parker, F.L. and P.A. Krenkel (ed.). 1969a.
Engineering Aspects of,Thermal Pollution,
vanderbilt University Press<
17-3
-------
14. Bush, R.M., E.B. Welch and B.W. Mar. 1974.
Potential effects of thermal discharges on aqua-
tic systems. Env. Sci. and Tech., Vol. 8, No. 6,
pp. 561-568,
15. Wurtz, C.B. and C.E. Renn. 1965. Water tempera-
tures and aquatic life. Cooling Water Studies
for Edison Electric Institute, John Hopkins
University, Research Project RP-49.
16. Jensen, L.D., R.M. Davies, A.S. Brooks, and-C.D.
Meyers. 1969. The effects of elevated tempera-
ture upon aquatic invertebrates. Cooling Water
Studies for Edison Electric Institute, John
Hopkins University, Research Project RP-49.
17. Coutant,. C.C. and S.S. Talmage. 1975. Thermal
effects. JWPCF (Annual Literature Review), Vol.
47, No. 6, pp. 1656-1711.
18. Drost-Hansen, W. and A. Thorhaug. 1974. Bio-
logically allowable thermal pollution limits.
Ecological Research Series, USEPA, EPA-660/3-74-
003.
19. National Academies of Sciences and Engineering.
1973. Water Quality Criteria - 1972. Ecological
Research Series, USEPA, EPA-R3-73-033.
20. Parker, F.L. and P. A. Krenkel (ed.). 1969b.
Biological Aspects of Thermal Pollution,
Vanderbilt University Press.
21. Carlander, K.D. 1965. Handbook of Freshwater
Fishery Biology. Volume One. Iowa State Press,
Ames, Iowa.
22. Altman, P.L. and D,S. Dittner (ed.). 1966.
Environmental Biology, Committee on Biological
Handbooks, Federation of American Societies for
Experimental Biology, Bethesda, Md.
23. Edinger, J.E. and J.C. Geyer. 1965. Heat
exchange in the enviornment, John Hopkins
University, Research Project RP-49.
a
17-3
-------
MONITORING OF ENVIRONMENTAL EFFECTS OF COAL
STRIP MINING FROM SATELLITE IMAGERY
Ronald L. Brooks
EG&G/ Washington Analytical
Services Center.
Las Cruces, New Mexico
SUMMARY
The objective of this investigation is to evaluate
satellite imagery as a means of monitoring coal strip
mines and their environmental effects. The satellite
Imagery employed in this study Is SKYLAB EREP S-190A
and S-190B from SL-2, SL-3 and SL-4 missions; a large
variety of camera/film/filter combinations has been
reviewed. The investigation includes determining the
applicability of satellite imagery for detection of
disturbed acreage in areas of coal surface mining,and
the much more detailed monitoring of specific surface
raining operations including: active mines, inactive
mines, highwalls, ramp roads,pits, water impoundments
and their associated acidity, graded areas and types
of grading, and reclaimed areas. Techniques have been
developed,via this investigation, to enable State and
Federal mining personnel to utilize this imagery In a
practical and economic manner, requiring no previous
photo interpretation background and requiring no pur-
chases of expensive viewing or data analysis equip-
ment. To corroborate the photo interpretation results
on-site observations were made in the very active mi-
ning area near MadisonviIle, western Kentucky. Strip
mining and reclamation officials of the state of Ken-
tucky assisted greatly during the on-site observation
and Inspection; they also reviewed the results of the
investigation and are very enthusiastic about the po-
tential of satellite monitoring to assist them in the
various duties performed by their agencies. This In-
vestigation was funded by NASA/Lyndon 8. Johnson Spa-
ce Center under contract NAS9-13310.
BACKGROUND
Coal mining has been accomplished in two manners:
underground mines and surface mines. When coal beds
are a few feet below the surface and In a relatively
horizontal position,they can be mined by removing the
material on top and shovelling out the coal.Presently
over 50? of the total coal production in the United
States is by the surface mining method because it can
be done cheaper by using modern machinery such as
gigantic coal shovels that can remove many tons of
material at a time.
SURFACE COAL MINING
Surface coal mining in the United States Is rela-
tively modern having started during the last decades
of the last century. It should be noted that here and
for the remainder of this paper we shall adopt the
terms strip mining / stripping because their negative
connotation has made the terms popular synonyms for
surface coal mining. Basically there are two types of
strip mines: area and contour.
Area Strip Mining This type of mining is usually
practiced on relatively level terrain with flat or
only gently sloping coal beds.The technique calls for
a first cut or trench which removes overlying soil,
commonly referred to as "overburden" or "spoil", and
exposes the coal bed. After the coal is removed, ad-
ditional cuts are made paralleling the first. As each
cut Is made, the overburden is deposited in the cut
previously excavated producing narrow and very steep
ridge-like features called "spoil banks" and a fairly
perpendicular scarp called the "hlghwall". After the
final cut is made and mining ceases, an open trench
or pit remains, bounded on one side by the last spoil
ridge and on the other by the undisturbed hlghwall.
Carlos G. Parra
Geoscience Research Group
New Mexico State University
Las Cruces, New Mexico
Contour Strip Mining Where hilly or mountainous
terrain makes area strip mining uneconomical, contour
mining Is practiced.Mining advances along the natural
contour of a hill being stripped.The first cut Is made
along the hillside above the coal bed.The hillside is
cut back as far as the value of the underlying coal
will allow,with the resulting overburden dumpedon the
downhtli slope. Mining proceeds laterally along the
slope creating a shelf or terrace in the profile of
the hill. Contour stripping leaves a near vertical
highwall above the mining ledge and a steeply Incli-
ned spoil ridge.
With advances in modern technology, the area strip
mining method is becoming more commonly practiced due
to the relatively lower operating costs of stripping
machinery.HI I Is and topographic Irregularities do not
present insurmountable economic barriers as they did
a few years ago. Today strip mining is considered to
be more advantageous than underground mines In reco-
very rates, grade control , economy, f lexibi I ity of
operation, safety and working environment.
Reclamation However,coaI strip mines have recently
received much adverse publicity. Problems of earth
waste, erosion, scenic degradation, acid drainage and
general ecological disturbances have plagued strip mi-
ning operations. For several years strip mines were
left untouched after mining operations ceased, leav-
ing un-reclaimed, torn landscapes which have been
given.the name "orphan areas".Due to the vast acreage
of strip mined areas and Increased public concern,
coal strip-mine reclamation legislation has been en-
acted in many states resulting in the reclamation of
considerable acreage.Coal strip-mine reclamation con-
sists mainly in returning the land to the best ecolo-
gic, topographic and aesthetic conditions.ActuaI rec-
lamation practices start before the mining operation
ceases. The actual procedure is to bulldoze the spoil
banks into the floor of the pit. Regulatory contracts
usually require this backfilling to reach the top of
the highwall and to be graded to the original slope;
elimination of spoil ridges;Impoundment of water;neu-
tralization of acid waters;burying all debris;fiI I Ing
the pit floor to a depth of 4-5 feet;and revegetatlon
resulting in acceptable growth. At present, 20 states
have set up laws and regulatory agencies to monitor
strip mining activities and to enforce regulations.
Each agency enforces these regulations by periodic
inspections and fines if necessary.However, the large
area I extent and a shortage of mining and reclamation
inspectors cause anexcessive work load forthe agency.
THE STUDY AREA
For the analysis of strip mining and reclamation
monitoring by satellite imagery It is necessary to
select a representative region.First, the region must
have a long history of Intensive coal strip mining so
that a proper study of pre-law and post-law areas can
be made.The area should presently be mined and ideal-
ly several stages of reclamation be present.Secondly,
the study area should also have ample coverage by the
satellite from which the Imagery for the study is to
be obtained. In the case of this particular study the
satellite used was Skylab,so ample coverage of a re-
gion by Sky/ab Imagery was sought. The coverage also
should include a wide variety of camera/fiIm/fI Iter
combinations for proper assessment.
1
17-4
-------
WESTERN KENTUCKY
The Western Kentucky coal fields meet these re-
quirements. These are horizontal to gently sloping
strata of bituminous coal of Pennsyivanian age. Coal
production from this field started more than lOOyears
ago and at present more than 33mlillon short tons are
extracted year Iy.Strippable bituminous coal resources
in the Western Kentucky coal fields are estimated at
more than 4b i11 I on tons, particularly centered around
Muhlenberg and Hopkins Counties.
KENTUCKY RECLAMATION LAWS
In 1964 the first effective reclamation law was
passed In the state of Kentucky. Prior to this law
more than 700,000 acres had been mined in the Western
Kentucky coal fields. Area strip mining left spoil
ridges;whatever little vegetation is present was nat-
ural growth under harsh conditions since no efforts
were made to cover the acid producing materlals.Pres-
ently there are more than 40,000 acres of these so
called orphan areas In Western Kentucky. Since the
newer, more effective law of 1966, orphan areas ore
not being produced. It should be noted here that no
one is responsible to reclaim these areas. Another
purpose of the law was to create the Division of Rec-
lamation In the Department of Natural Resources and
Environmental Protection, Bureau of Land Resources.
INSPECTIONS
One of the principal offices of this newly created
agency is located In MadlsonviIle, Kentucky, and some
of the specifications that thfs office must check are
Iisted below:
-Check reclamation plan, determine
damage to public roads and streams
-Check breakthrough of acid waters
-Check the extent of ponded runoff
-Check dumping and piling of spoil
-Check proper complete backfilling
-Check vegetativecover requirement
-Check type of planting or seeding
-Check date of planting or seeding
-Check area of I and to be reclaimed
-Check effectlvenessof reclamation
The law also gives the director of the Division of
Reclamation the power to suspend a permit and thereby
halt all operations for non-compliance with the regu-
lations. Mine Inspections usually take place without
notice by one of the ten to twelve mine inspectors
from the field office. It should be noted that one
mine inspector may be responsible for up to ten mines
ranging In size from a few acres to thousands of acres
plus being responsible for the continued surveillance
of the landscape against Illegal operations.Any fast,
accurate monitoring technique, such as that presented
here, will be welcomed.
VIEWING TECHNIQUES
A number of techniques were studied for analyzing
the Sky lab transparencies such as standard photo en-
larging equipment, stereoscopes, magnifying reticles
used with a light table, and Map-O-Graphs. The best
method for viewing the Imagery was found to be a sim-
ple overhead transparency projector, since transpar-
encies of this type(9"x9"copy) having scales approxi-
mately 1/725,000 (S-190A) and 1/475,000 (S-190B5 are
too small to be viewed without enlargement. The less
expensive models using plastic fresnel colllmatlng
lens were found to be too poor for ana lysis.The over-
head projectors of a type similar to the Transpaque
Auto level Model 20400 used In this study, use a par-
abolic reflector In their optical system and produce
Images which are perfectly suitable for this type of
study. This projector when used with a smooth white
screen, such as poster board, can produce high reso-
lution Images In any type of darkened room. The Image
produced by the overhead projector has the advantage
of being viewable by any number of people, can be
studied closely and features can be traced when pro-
jected onto a suitable paper. The scale Is infinitely
variable just by changing the distance from the screen
to the projector. The S-190B transparencIes were en-
larged to a scale of 1/18,000 for detailed area study
and could be enlarged further with little loss of re-
solution. Most of all, the technique is simple,effi-
cient and produces excellent images which can be used
by laymen with no photo-interpretation experience.
IDENTIFICATION OF DISTURBED AREAS
The first analytical section of this study is to
determine the applicability of satellite imagery for
the overall recognition of coal strip mined areas.The
general aim is to be able to distinguish this type of
land-use from all others. Due to the physiographic,
climatic and economic elements of western Kentucky,
the land-use types generally fall In the following
categories:
-Farms
-Urban areas
-Strip mines
-National Parks
-Natural Forests
-Highways and Railroads
-Lakes, streams and reservoirs
The procedure followed for this analysis Is similar
to that explained above. Each transparency was viewed
independently through an overhead projector at scales
around 1/100,000 so that the entire photo area could
be viewed at one time. At this scale It Is difficult
to recognize some of the features peculiar to strip-
mines such as pits, ramp roads, etc... Therefore an
initial recognition of the sites disturbed by strip-
mining Is necessary. Individual transparencies were
viewed and visual Inspections of land-use patterns
were made and recorded.The two Most applicable trans-
parencies are:
SL-3, S-3 90A. CAMERA STATION 2
Two transparencies from Mission SL-3, Camera S-190A
station 2 In black and white Infrared were obtained;
one positive and one negative. The lower 1/4 of the
photo is almost totally obscured by cloudsjfortunate-
ly» +he strip mines are located In the upper right
sector where cloud cover Is minimal.The most outstand-
ing feature In these transparencies is the elongated,
U-shaped feature formed by the Kentucky and Bark ley
Lakes.Medium size lakes such as Lake Beshear and Lake
Malone are also Identified. At this point It was de-
cided to use the negative transparency because It af-
forded better tonal contrasts.In this negative trans-
parency even smal) lakes such as Lake Peewee appeared
as light features contrasting with darker tones sur-
rounding them.The areas disturbed by strip-mining are
distinguished by their light tone and t>y a series of
lighter tone lines representing the hlghwalls exposed
to direct sun IIght.
SL-3. S-)908
The transparencies from Mission SL-3,Camera S-19QB,
In color Infrared proved to be also of good quality
for overall s+rlp-mlne recognition.SI nee these photos
were taken during the same pass as the one described
above, the area Is under the same cloud cover condi-
tions. Detail is greater, not only due to a larger
scale but also due to the film's higher spectral res-
olution. Medium size water bodies such as Rough River
and Norlln Reservoirs are easily Identified. Mammoth
Cave National Park Is well defined since It represents
a change In red color not found anywhere else In the
transparency. Strip-mines are prominent appearing as
the only light blue tones with lighter white lines In
the entire transparency.
2
17-4
-------
DETAILED MONITORING
Detailed monitoring of coal strip-mines is per-
formed using the same equipment and procedures as for
disturbed areas studied above.The main viewing differ-
ence is that the overhead projector was placed far-
ther from the screen to obtain larger scales 1/36,000
for S-190A and 1/18,000 for S-190B. Since mapping the
entire 9" x 9" area would be time consuming it was
decided to choose representative areas. It should be
noted here that the transparency was masked by a black
heavy-weight paper with a 2" x 2" cut out, so as to
allow only the selected area to be projected on the
screen. This procedure enhanced viewing because it
eliminated the projector's light except on the study
area, resulting in an image which appears sharper to
the eye.The actual analysis proceeded as follows:once
the image was projected onto the screen, a general
reconnaissance was performed aided by topographic and
geologic maps. Tracing of features was followed by a
classification of tones and/or colorsand intensities.
The app I icabiIityof three selected transparencies for
detailed monitoring follows:
SL-2, S-190A. CAMERA STATION 6 POS.
This transparency is underexposed by 1/2 fstop and
this seems to be partly responsible for its poor qual-
ity. A variety of tonal patterns were discernible and
traced, yet this is somewhat misleading because simi-
lar features often had different tonal patterns.Close
inspection reveal that light tones (Kodak Gray Scale
values of 0 to .2)or areas of high ref 1 ectanceusua My
represent southfacing highwalls, ramp roads and ac-
tive mines;medium tones (.3 to 1.5) usually represent
orphan areas, reclaimed areas and farming land; dark
tones (1.6 to 1.8) usually represent the pit of mines
or the corona of trees in contour-mined areas; very
dark tones ( >1.9) represent natural vegetation. It
should be stressed that these are gross generaliza-
tions and the transparency Is not really adequate for
detailed monitoring.
SL-2. S-190A CAMERA STATION 3
Similar procedures were used to analyze the color
infrared transparency obtained from camera station 3.
Again, there is a great variation in colors rather
than in gray tones. Naturally vegetated areas appear
as dark red, vegetational differences are evident; for
exampIe,timbers growing in the floodplain ofpond Riv-
er were of a darker red than those inthe interfluvial
areas. Generalizations were necessary because the va-
riety of colors and intensities tendto obstruct rath-
er than enhance recognltion.The following generaliza-
tions tend to conform rather well to their features:
Very light to light reds— undifferentiated
farming or natu-
ral grasses
Very dark to medium reds— natural vegeta-
tion (timbers)
Very light blue to white— exposed hlghwall
shallow mi nes
Light blue to dark blue— strip mine rec-
lamation, active
m i nes
Very dark blue tones— deep pit floors
In general this transparency Is much better than
the one discussed above yet it falls to provide exact
information and therefore we classify it as of medium
qua Iity.
SL-4, S-190B
This is by far the most outstanding Sky lab trans-
parency for monitoring coal strip mines.Its sharpness
and high resolution are truly amazlng.lt Is felt that
it is best to make a few general observations and then
to devote a few paragraphs to examples of detailed
monitoring using this imagery.
Genera Ii 11es Most open water bodies such as Peewee
Lake appear as medium blue with intensities varying
according to water depth; siltation and sedimentation
are identified by their brownish color. Highways, re-
garless of the width and composition appear as very
sharp I ines.Because roads are so clearly visible this
imagery is moit helpful for mine inspectors.Very fine
detail is recognized in the town of Madisonvi I Ie, Ky.
where streets, golf courses, trailer parks etc... can
be recognized. Strip mines are shown bIuish,highwaIIs
as white sinuous I ines,pits as darker blues or black.
Orphan areas are distinguished by the large number of
highwalls. Reclaimed areas appear in medium blues to
brownish colors depending 'on the type of reclamation
practiced (grasses = blue; trees = brown ) or their
stage of reclamation (recently planted trees = blue;
successful tree growth = brown ). Farmed areas appear
with unique rectangular pattern varying in color from
light reddish-brown to bIuish,depending on their cul-
tivation stage.
Inactive Pre-1 aw contour mines Both types of strip-
mining methods are identified. Inactive contour strip
mining in an area just west of US Highway 41 and iust
north of the western Kentucky Parkway is identified.
The roughly circular features are local topographic
highs produced by domal structures' of sediments.These
are old pre-law mines now inactive. There is a varia-
tion of colors and intensities. The corona of trees
left undisturbed above the coal seam appears evident.
Highwalls appear light gray,particuIarIy those facing
south which with higher solar exposure and thus less
moisture available have not been naturally reclaimed.
Colonial Coal Company Area Mine A large feature Im-
mediately recognized as an area strip-mine Is that
operated by Colonial Mine Company just west of Madl-
sonville. The mine consists of a large north-south
trending pit which Is being worked in an easterly di-
rection. A series of ramp roads are leading into the
pit and trending east-west which parallel a series of
spoiI ridges. The area west of the present pit has
been reclaimed as the operations progress. There is a
gradation of blues indicative of successive stages of
reclamation. A reclaimed area which has suffered an
army-worm infestation which has drastically altered
the reclamation process is readily visible.
Off-strike grading Another Interesting detail obser-
ved from this imagery is the off-strlke grading shown
at Cimarron Coal Company mine.Off-strIke grading con-
sists of a series of parallel ridges of spoil graded
to trend perpendicularly with the strike of the coal
seam for soil conservation purposes. In this totally,
well established, reclaimed area the off-strlke grade
appears as subtle linear changes of medium blue.
Detection of acid waters Last It should be noted that
this imagery is also capable of detecting concentra-
tions of Iron oxides, also known as "Yellow boy" at
different water Impoundments. There Is an un-named
Lake (herein referred to as Red Lake) just west of
Madlsonvllle and east of the Colonial Mine Company
which appears quite reddish in color In this S-190B
imagery. Further field investigations revealed that
this water impoundment has a real reddish coloration
due to a high concentration of "yellow boy".Mine Ins-
pectors have tested Red Lake waters and found It to
have a pH of 4.5, an indication of acidity and proba-
ble environmental detriment.
ON SITE VERIFICATION
For the purpose of verifying the results of our
ph6+o interpretation, two trips were made to the Mad-
lsonvllle, Kentucky, area. The Initial trip was for
gathering general Information. Liaison was established
with the officials of the Kentucky Department of Nat-
ural Resources and Environmental Protection, Division
of Reclamation at Madlsonvllle. Mine census Informa-
tion and topographic maps were provided. Ideas were
3
17-4
-------
exchanged on the potential benefit of satellite im-
agery providing large-area monitoring. During the in-
spection tour, we were shown pre-law orphan areas,
with remaining highwalls and poor grading;we also saw
post-1 aw stripping methods and much improved reclama-
tion techniques. The Sky lab imagery was taken to Mad-
isonville for our second visit. The imagery, via an
overhead projector,was reviewed by the mining inspec-
tors. These gentlemen, although not accustomed to the
scope of coverage by satellite imagery, verified all
our prel iminary findings as I isted above within ha If
an hour. The Assistant Director of Kentucky's Depart-
ment of Natural Resources and Environmental Protec-
tion concluded that "The Sky lab photographs...wouId
be a most useful tool for environmental law enforce-
ment and (coal) resource surveys..."
CONCLUSIONS AND RECOMMENDATIONS
The resulsts obtained in +his study prompt us to
make the following conclusions and recommendations.
It is our sincere hope that satellite monitoring may
be regarded as a useful tool not only to federal and
state agencies but also to the many industrial, envi-
ronmental, and other groups associated with coal strip
mining. The recognition of areas which have been dis-
turbed by strip mining is a task which can be well
aided with the use of satellite imagery. The large
areas covered by orbital photography allows the user
to estimate the acreage and assess the environmental
effects of strip mining activity from just a few
frames. The major feature of such imagery is the ease
of differentiation between strip mines and all other
types of land use.Infrared-photography both in colors
and black and white was found to be the best suited
for this purpose, it is definitely an asset for Strip
Mining and Reclamation agencies to use Sky lab type
imagery (particularly S-I90B transparencies) to moni-
tor the several environmental factors associated with
strip mining as outlined above. The high resolution
of these imageries is well suited for detailed moni-
toring of strip mines. It is recommended that imagery
taken with a 6-month interval would be best because
it would show significant changes in operation, recl-
amation and environmental effects without too long an
interval for law enforcement and proper management.
BIBLIOGRAPHY
Cubbison,E and L.Dunlap Stripping the Land. Only the
Beqinninq. COALition Against Strip Mining,
Friends of the Earth, Feb 1972.
Kentucky Department of Natural Resources and Envi-
ronmental Protection Title XXV I I I:Mi nes and
Minerals Chapter 350 Strip Mining
Kentucky Department of Commerce and Kentucky Geolog-
ical Survey, Mineral Resources and Mineral
Industries of Kentucky (map) 1962.
Kentucky Division of Reclamation Surface Mining and
Reclamation in Kentucky 1972.
NASA/JSC Sensor Performance Report for SL-2 and SL-3
Principal Investigators Management Office.
NASA/JSC Sensor Performance Report v.l S-190A Systems
Analysis and Integration Office. 1974
US Bureau of Mines Information Circular No.6772, 1966
No.8406, 1969
No.8531, 1972
No.B642, 1974
4
17-4
-------
AERIAL REMOTE SENSING APPLICATIONS IN SUPPORT OF
OIL SPILL CLEANUP, CONTROL AND PREVENTION
Donald Jones
U.S. Environmental Protection Agency
Washington, D.C.
and R. Landers and A. Pressman
U.S. Environmental Protection Agency
Environmental Monitoring and Support Laboratory
Las Vegas, Nevada
The Environmental Protection Agency and the U.S.
Coast Guard, in response to the oil spill law require-
ments embodied in the Federal Water Pollution Control
Act Amendments of 1970 and 1972, have developed remote
sensing surveillance systems for both detection of oil
discharges and mapping the areal extent of major oil
spills in support of cleanup and control operations and
environmental damage assessment. This paper describes
the aerial systems and several of the operational sup-
port applications in the National Oil Spill Prevention
and Control Program being undertaken by these Federal
agencies.
BACKGROUND
Oil spill surveillance, detection, and assessment
by remote sensing came into being as a result of the
Santa Barbara Channel major oil pollution incident in
January 1969. The critical need for aerial observation
by visual or whatever means to identify the location,
areal extent, and movement of the oil was clearly dem-
onstrated in this dramatic ana unusual spill. Conser-
vative estimates placed the release of oil from an
offshore platform at three million gallons over a 100
day period, with the oil slick at one time covering
approximately 500 square miles. Day to day mapping of
the oil spread was carried out by visual observations
from aircraft in support of the control and cleanup
activities. Aerial cameras and multispectral scanners
were also utilized, but primarily in an experimental
role to determine their capability for discriminating
oil on water. Delays experienced in data processing
and interpretation precluded their use in the planning
of day to day control and cleanup activities. Import-
antly, however, it was learned that standard aerial
color photography and the ultraviolet, thermal infrared,
and microwave regions of the spectrum all provided good
contrast between oil and water, with resultant potent-
ial for aerial mapping of large oil spills and as well
for detection surveillance of oil discharges from ships
and offshore platforms. There remained the data hand-
ling problem of providing adequate timeliness to be of
use in support of the cleanup and control operations,
and the integration of these sensors into a surveillance
system.
The Santa Barbara incident also provided the impe-
tus for the first U.S. comprehensive and enforceable
law dealing with oil spills, which was enacted in April
1970. A key feature of the law was the requirement to
prepare a National Contingency Plan for removal of oil
by providing for efficient, coordinated and effective
Federal action to minimize damage from oil discharges.
A revolving fund of $35,000,000 was authorized to carry
out the Federal responsibility for oil spill removal in
those circumstances of an unknown violator, inadequate
cleanup response by a known violator, or in natural
disaster situations. The U.S. Coast Guard and the Envir-
onmental Protection Agency were designated as the lead
Federal Agencies to implement the National Contingency
Plan, to include a system of surveillance and detection
to insure earliest possible notice of discharges of oil.
From this statute authority a research and development
program evolved, sponsored by the U.S. Coast Guard and
EPA, leading to an operational surveillance system for
U.S. territorial and contiguous waters to detect oil
discharges and map the areal extent and thickness pat-
terns of major oil spills when they do occur.
CURRENT STATUS OF AERIAL OIL SPILL
SURVEILLANCE
U.S. Coast Guard
The Coast Guard is responsible for oil pollution
surveillance, enforcement of antipollution regulations,
and initiation of various antipollution countermeasures
to minimize oil pollution hazards in coastal and navi-
gable inland waters of the U.S. In meeting this respon-
sibility, the Coast Guard has undertaken the develop-
ment of an airborne sensor system which will detect,
classify, quantify, and map oil spills on the ocean
surface. The Airborne Oil Surveillance System (A0SS)
is intended to accomplish these tasks.
Functions of the AOSS are to (1) detect oil slicks,
(2) indicate the magnitude of spills - areal extent and
thickness, (3) identify and document the source(s) of
discharges, (4) assess cleanup operations, and (5) ga-
ther data regarding the frequency and magnitude of sig-
nificant spills. The AOSS was designed to meet the
needs of both the law enforcement mission (large search
area with infrequent targets) and the countermeasures
mission (oil-spill assessment under a wide range of
weather, lighting, and ocean surface conditions). Addi-
tional requirements placed on the prototype system in-
clude (1) effective surveillance of a 50-n mile-wide
coastal zone, (2) long-range (25 nmi) detection of
ships and long-range detection of oil slicks, and (3)
adverse weather operations. In addition to the AOSS,
the Coast Guard is sponsoring research for an active
fluorescence sensor for airborne classification of oil
spills, multi-frequency passive microwave techniques
for oil thickness measurement/quantification of oil
spills (Naval Research Laboratory), and a television
system which detects oil spills by displaying the dif-
ference of signals of opposite polarization (a NASA pro-
ject) .
The prototype system was developed by Aerojet Elec-
tro Systems Company and consists of (1) an X-band
sidelooking radar for long-range ship and oil detection
and adverse weather oil-spill mapping; (2) a 37-GHz
passive microwave imaging system for adverse weather
spill mapping and spill thickness approximation; (3)
a multlspectal low-light level TV system for high reso-
lution spill documentation and violator identification;
(A) a multichannel line scanner for spill confirmation
and discrimination; (5) a position reference system for
legal and operational effectiveness; and (6) a real-
time processor-display console for maximum operational
effectiveness. The system provides for false target
discrimination, automated detection alarming, and a
color display to achieve maximum coupling of the sen-
sor information to the equipment operator.
Design of the AOSS was initiated in 1972 and Inte-
gration was completed in early 1974. The AOSS was
installed aboard a Coast Guard HU-16E Albatross in
June 1974 and was fLight-tested off the southern Cali-
fornia coast during a series of shakedown and back-
ground data flights from June through August 1974.
Background data flights were conducted mostly over the
Santa Barbara Channel where natural seeps and routine
shipping activities are commonplace. The AOSS was
17-5
-------
subsequently based at the Coast Guard Air Station, San
Francisco, and it participated in a comprehensive flight
test program off the northern California coast during
September and October 1974. The test program included
AOSS evaluation of (1) a series of controlled static
and dynamic spills on an ocean site approximately 100
nmi west of Santa Cruz, California; (2) routine tanker
discharge operations (3) surveillance of routine shipp-
ing and harbor activities; (4) potential false targets;
and (5) selected targets of opportunity. Details of
test results and of the AOSS program from which this
synopsis was taken are noted in the Proceedings of the
1975 Conference on Prevention and Control of Oil Pollu-
tion March 25-27, 1975, San Francisco in and article,
"Flight Evaluation of U.S. Coast Guard Airborne Oil
Surveillance System" by Lt. A, Maurer, USCG and A.T.
Edgerton, Aerojet Electro Systems Company, the sur-
veillance system was generally reliable and the flight
test results demonstrated that a practical airborne oil
surveillance system is feasible.. The tests also showed
the system to be useful for other Coast Guard missions.
Future implementation envisions sensor package
production to be mounted in six to eight jet aircraft
with initial surveillance operations scheduled for late
1977. A recently completed contract with Purdue Univer-
sity, Wright & Olson, principal investigators, proposed
pollution forecast models to develop a large scale,
random surveillance schedule for pollution, detection in
the coastal waters of the U.S. including Alaska, and
the Great Lakes. Personnel from the Environmental Moni-
toring and Support Laboratory (EMSL) Las Vegas are
currently assisting the Coast Guard in developing a
plan for data processing and evaluation facllitites to
effectively handle the data inflow.
The'U.S. Coast Guard currently has an interim aeri-
al surveillance system using six Grumman ECU" - 16E
fixed-wing aircraft equipped with infrared and ultra-
violet sensors. This sytem is utilized primarily in
detecting oil discharges from ships sailing in U.S.
coastal waters and the Great Lakes. In addition, Coast
Guard shore units, harbor patrol craft, and helicopters
all have an oil pollution detection mission.
EPA
When EPA was formed in December 1970 from elements
of a mi»ber of other Federal agencies, a fortuitous in-
heritance from HEW was the Western Environmental Re-
search Laboratory at Las Vegas. At this Laboratory
were several prop aircraft vintage WW II that were uti-
lized for air sampling probes in support of underground
nuclear weapon tests and monitoring of nuclear power
sites. From this "desert escadrille" a remote sensing
unit was developed during the next several years in
response to varied and considerable Headquarters £nd
Regional requests for water, land and air pollution
surveys, one of the earliest of which was for aerial
mapping of large oil spills. Importantly, all the
elements and requirements of a remote sensing system
were considered such that a modern photographic labora-
tory and relevant,data processing and reduction facili-
ties were established. Also, skilled personnel, both
Government and contractor, were brought aboard to en-
sure meaningful analysis, evaluation and presentation
of the data acquired. The Las Vegas remote sensing
unit, now a Division of EMSL, along with its field
station, the Environmental Photographic Interpretation
Center (EPIC) at Warenton, Virginia, and contractor
data acquisition support has proven to be very effect-
ive in remote sensing applications with comprehensive
operations commencing in 1973. A number of successful
enforcement actions in air and water pollution cases have
taken place, based largely on remotely sensed data, and
aerial photographic mapping of large oil spills has be-
come an integral and necessary element in support of
effective cleanup operations and oil pollution environ-
mental damage assessment. Figure 1 identifies the type
aircraft and sensors used in a variety of remote sens-
ing missions.
SNSl • l» Itwm SM6i kliutll DMA WttKlIM lAtlllMS
AlnCnAfT 121
MOHAWK
C-4S
S£flfS0fiS: AERIAL CAMERAS
«c i 1 INCH MAPPING
"C « (2 EACH) fl INCW MAPP'AfG
KC-ia 8 WtCK MAPPING
*•31 »•!«
-------
common for large spill incidents to have Coast Guard
helicopter overflights for visual observations, which
normally can be conducted within only a few hours sub-
sequent to knowledge of a big spill. Finally, the
mounting and assembly of Federal and contractor clean-
up resources can require twelve hours or more to com-
mence operations. In any event, experience with some
six major spills over the past two years has demon-
strated the need for and value of initial data acquisi-
tion within 12 and not more than 24 hours after the
spill event, and as frequently thereafter as the
nature of the spill and circumstances dictate.
After each flight the Project Officer is briefed
by the flight crew as to the visual spread and loca-
tion of the oil which information is in turn relayed
to the OSC. The film data, usually from a Wild or
Zeiss aerial camera color negative or color reversal,
in 9" format, is immediately dispatched by first com-
mercial aircraft to Las Vegas for pickup and immediate
processing and photo interpretation at EMSL, where the
personnel undertaking these tasks have been placed on
a 24-hour work schedule. Initial analysed results are
telephoned to the Project Officer (then relayed to the
OSC) after which the analysed flight line transparen-
cies and a color coded (light and heavy oil accumula-
tion) large scale map of the spill scene are dispatched
to the Project Officer by first available commercial
air. Upon receipt the Project Officer then conducts a
full briefing for the OSC and the Federal members of
the cleanup control and response team at earliest
opportunity. Time elasped from overflight to full
briefing has averaged 24 to 30 hours.
Hard copy documentation such as this is essential
for post-spill environmental damage assessment, and
has proven timely enough for effective support in the
day to day planning and monitoring of cleanup and con-
trol operations. Frequent helicopter visual observa-
tions of the spill area and cleanup operations continue
to be necessary and indeed can be adequate alone for
spills not exceeding two or three hundred thousand
gallons. However, geographical and environmental cir-
cumstances associated with the spill are also criteria
to be considered in the decision for need of permanent
documentation. The advantages of aerial color film
over black and white (B/W) panchromatic in detecting
oil on shorelines and marshlands as well as for accumu-
lation determination far outweigh the locally available
rapid processing advantage of B/W. Relatively few
photographic laboratories in the U.S. have the capabi-
lity and quick turnaround time for processing aerial
color film. Thus, the Las Vegas photo facility is
utilized in the valuable dedicated role along with the
standby photo interpretation function. Frequency of
commercial jet air schedules is adequate for 24-hour
turnaround from data acquisition to complete briefing.
For some missions aerial B/W has been flown, processed
by local photographic facilities, and interpreted by
the On Scene Project Officer for briefing within a 12-
hour time frame. Similar quick turnaround as well as
nighttime aerial mapping can be achieved through the
use of thermal IR scanners. For large coastal spills
threatening many miles of shoreline, wide coverage IR
scanning of oil slicks covering hundreds of square
miles is ideal in support of countermeasure plans in
the distribution and location of cleanup and control
resources. Such missions were conducted for the mass-
ive four million gallon oil spill in the Gulf of Mexico,
resulting from an off-shore platform blowout with the
oil release lasting for three months in early 1971. Fi-
gure 2 illustrates this type of mapping mission.
Figures 3 and 4 are examples of photography taken
in the two major oil spills during February and March,
1975 in the Delaware and Mississippi Rivers. In each
incident, an oil barge tow breakaway and collision with
a bridge and an oil tanker/ship collision, more than
100 miles of shoreline was contaminated to varying de-
gree. Three overflights were conducted for each spill,
with the final data documentation (not shown here)
clearly showing the effective cleanup results of most
of the oil removed or dissipated and dispersed by natu-
ral forces. Remaining vestiges of the oil were inves-
tigated by cleanup crews and the environmentally sensi-
tive areas affected were subjected to environmental da-
mage assays by EPA scientists. ,
INFRARED IMAGERY - SHELL OIL PLATFORM-B I IRE, GULF 01 MEXICO
PREPARED FOR ENVIRONMENTAL PROTECTION AGENCY
j * "h' vEGfTMUlH *S
PROJECT 7514
1302 CST 3/6/75
Oil SUCKS
DAMAGED IAAGE
MISSISSIPPI RIVER
OH SUCK
&jp i
''ill
> ,% * DELAWARE RIVEK OIL SPILL
' RAMMED OIL TANKER ON FIRE
V fjjH? ,Marcus Hook, Pennsylvania
r | February 2, 1975
:
3
giife ' ¦ **•<>¦
Figure 4. B/W reduction of 9" color view of burning
tanker and resultant oil pollution of one million gal-
lons .
17-5
-------
OIL POLLUTION PREVENTION COMPLIANCE
"MONITORING by AERIAL PHOTOGRAPHY
EPA issued its oil pollution prevention regulation
in the Federal Register In December 1973, effective on
January 10, 1974. This regulation established proce-
dures, methods and equipment and other requirements
for equipment to prevent the discharge of oil from
non-transportation-related onshore and offshore facili-
ties into or upon the navigable waters of the United
States or adjoining shorelines. The regulation was
directed to owners or operators of non-transportation-
related onshore and offshore facilities engaged in
drilling, producing, gathering, storing, processing,
refining, transferring, distributing or consuming oil
and oil products, and which, due to their location
could reasonably be expected to discharge oil in harm-
ful quantities into or upon the navigable waters of
the United States or adjoining shorelines. Owners or
operators of these facilities were required to prepare
a Spill Prevention Control and Countermeasures Plan,
utilizing guidelines set forth in the regulation.
These guidelines identified means and measures to pre-
vent oil spills in the first place and, when occuring
to prevent the oil from reaching a waterway.
Several hundred thousand facilities in the U.S.
have to comply with this regulation, ranging from huge
refineries to small stripper wells having a storage
capacity in excess of 1320 gallons. With limited
Federal and State inspection personnel, it is patently
impossible to routinely monitor all those facilities
for regulation compliance. In practice, spot checking
is carried out with particular attention paid to prior
spill violators and as well to those facilities with
high spill potential adjacent to major waterways.
Also, for oil spills occuring after January 1974, the
facility's Plan is subjected "to a review process and
the facility inspected to ensure the carrying out of
the required Plan modifications.
Aerial photographic missions in support of compli-
ance inspections have proven useful in a variety of
ways. Both high altitude and low altitude overflights
have been conducted for several EPA regions. Data ob-
tained has served inventory purposes for stripper well
production and storage inspection and drainage pattern
identification from storage facilities. The latter
use, for example, has served to test the applicability
of the regulation to a facility which, due to its lo-
cation, could reasonably be expected to discharge oil
into or upon the navigable waters of the U.S. The
photographic data enlargements also reveal the general
operating conditions of the facility and whether the
regulation prevention guidelines are being carried out.
Figure 5 is such an example of oil storage tank leaks
that need to be corrected.
I
Finally, for refineries, tank farms and industrial
complexes transferring, storing and using oil in the
production process, aerial photographs serve as an
excellent guide for ground inspection of indicated or
suspected violations and spill potential areas within
the complex that may require Plan modification to re-
duce spill risk. In sum, aerial photography has proven
to be an effective tool in the compliance monitoring of
the oil spill prevention regulation in optimizing the
effectiveness of the limited resources of the facility
compliance inspection program. For example, Figures 6
and 7 show an aerial and ground view of a stripper well
storage and oil separator facility, typical of thousands
In several States. This aerial survey in Kansas of
several hundred such facilities enabled ground inspec-
tion of just those where aerial data enlargements had
indicated oily discharges or other regulation viola-
tions.
Figure 5. High altitude view of Hudson River and
Albany, New York with llx and 25x data enlargements
showing oil tank leaks.
EPA's EMSL with its field station EPIC at Warren-
ton, Virginia and supported by contractor, NASA, and
USGS data acquisition is carrying out the aerial com-
pliance monitoring program, both in support of regula-
tion enforcement and in an experimental role to
determine additional applications for the oil pollution
prevention program. For example, flood prone areas
where oil production, storage and processing facilities
are located need to be assessed for spill potential.
Importantly, expertise gained here can be applied to
prevention programs for hazardous substances spills,
pnce these substances have been identified by Federal
regulation.
Remote sensing operational applications to solving
pollution problems and for enforcement purposes are at
a beginning stage, for which the sensor state-of-the
art and remote sensing techniques have by no means been
fully exploited. Rapid recognition of the remote sens-
ing potential is now taking place within EPA and the
industrial and environmental community. Such applica-
tions as outfall inventory, thermal plume and industri-
al effluent movement, ocean outfall and ocean dumping
dispersion, and lake eutrophication surveys are but a
few of major water pollution problems for which remote
sensing is now providing useful and valuable assessment
data, thereby contributing to the achievement of the
clean water goals set forth in the Federal Water Pollu-
tion Control Act of 1972.
Waste pond
Drain Ditches to Ri.ver
Ground Photo
Figure b. B/W reduction of 9" aerial color view of
stripper well storage facilities and separator waste
ponds in Kansas.
EPA Inspector
Oil Discharge j*. M ¦¦ ¦
Figure 7. Ground view of storage tanks and oily waste
discharge. Magnification of aerial data and stereosco-
pic viewing detected this violation of the oil pollu-
tion prevention regulation.
Oil/Water Separator
100 bbl (4200$) Tanks
i- V
n
4
17-5
-------
POWER VS. POLLUTION : A NUMERICAL APPROACH
by
Herold I. Zeliger and Marsha Funk
Harold I. Zeliger, Ph.D.
Chemical & Environmental Consultants
18 Spring Hill Terrace
Spring Valley, New York 10977
The second law of thermodynamics implies that one
form of energy may not be converted into a second form
with" 100% efficiency. Stated another way, all energy
transformations have inefficiency factors associated
with them,The8e inefficiencies give rise to environ-
mental pollution whenever man taps nature's energy
reserves for his own use.
In determining the extent of environmental poll-
ution from the inefficiencies of different energy
transformations, one is presented with a serious
dilemma. How, for example, are the effects of nuclear
fallout to be compared with nitrogen oxide exhausts
from an automobile engine for environmental Impact?
Such comparisons may be made by defining poll-
ution in units of energy released. We define the poll-
ution coefficient, Qp, as that fraction of the total
energy transformed,AEt, which does not accomplish the
desired task. Numerically,
Qp -AEp/AEt (1)
where equals the pollution energy.
For a given energy transformation, the total
energy available must first be considered. For combust-
ion of carbonaceous fuel (as shown In equation 2) the
heat of reaction does not account for the total energy
Impact on the environment. Carbon dioxide and water
vapor cause a greenhouse effect In the atmosphere and
sulfur dioxide and oxides of nitrogen are further
oxldlced and hydrated in the air to corrosive acids,
as shown by eq. 3 and 4.
CnHinSpNq + 0;' ^ nC02 + m H2O + pSOj + qN02 (2)
so2 + o2
no2 + o2
*£) H2S04(aq)
-2*2^ HNO-(aq)
(3)
(4)
To take this factor into account, we are propos-
ing that the energy expressed by the standard heats of
formation, ARf, of the ultimate products be added to
the energy of the reaction, AEm, In computing the
total energy released when the products are pollutants
(as shown in eq. 5).
^"Et * ^Erxn + ^Hf(products)
For the combustion reaction shown in eq. 2,
•^(products) " nAHf(C02) + 5
-------
From eq. 7,
AE„
AE
AEp - -20,650 - 0.22(-10,510)
Ae - -18,338 cal
P
Substitution in eq. 1, gives,
Qp - AKp/AEt
Q - -18.338/-20.650
P
Q « 0.89
P
r.nat. rfiMRitRTtnw tn a steam turbine electrty^t.
GENERATOR-
Assuming an average composition for coal of; 88%
carbon, 1% hydrogen, 1% sulfur and 10% inert ash,* an
efficiency of 40% conversion of heat of combustion to
electricity' and an average heat of combustion of
-6733 calories per gram , a calculation similar to
that shown for petroleum above yields the following
results:
^E.
AErxn + ^Hf (products)
AE - -6733 + (-8462)
A! ¦ 4E„ -
-------
serves only to pollute the environment. The globe can
accomodate only a limited number of calories of pollut-
ion per year. Based on fuel consumption predictions*®
we estimate that man is releasing about 4.5 x 10*9
calories per year at the present time. Though this is
only a small fraction of the solar energy reaching and
leaving the earth's surface (estimated at 5.4 x 10
calories per year)^ it is too much for the globe to
handle, as evidenced by polluted streams and fouled
air, etc.
We must first determine how much energy release
the globe can tolerate. We believe this can be done
by choosing the last year in which there was little or
no pollution of the environment, estimating the energy
release of that year and setting that as a maximum
allowable annual release. More efficient ways of
generating power, e.g., direct conversion of solar
energy into electrical energy, or the use of solar
energy to produce hydrogen, would permit us to use more
energy without incrasing pollution.
Let us assume that the year 1940 was the last
non-polluting year. In 1940, the total energy release
is estimated to have been about 1.5 x 10 calories}"
or one third of the current rate. Our challenge is to
at least triple the efficiencies of our methods of
power production just to keep from further polluting
our environment. As the world's energy needs grow, we
shall have to increase energy conversion efficiencies.
Even if this were possible, the growing demand is such
that the requirement of 100% efficiency would be reach-
ed and surpassed in a very few years.
It seems obvious that the only long term solution
to the world's energy/pollution problem is the use of
solar energy. We must find ways to capture some fract-
ion of the incoming solar radiation. Winds, tides,
ocean currents, solar cells, solar hydrogen generators,
etc. These are but a few of the options available to
man.
6. Ref. 1, Vol. II, p. 131.
7. M. Benedict, Bui, of Atomic Scl. . Vol. XXVII,
No. 7, 8 (1971).
R.T. Weidner and R.L. Sills, Modern Phvalea. 2nd.
ed., Allyn and Bacon, Boston, 1968, p. 460
J.F. Hogerton, Atomic Fuel. U.S. At. Energy Comm.,
1968, pp. 1-2.
8. S. Glasstone, Controlled Nuclear Fusion. U.S. At.
Energy Comm., 1968, pp. 6-13.
9. ibid, pp. 9-14.
W.C. Gough and B.J. Eastlund, Scientific Amaerlcan.
Feb., 1971, p. 223.
10. Committee on Resources and Man, National Academy of
Sciences, National Research Council, Resourcea and
W«nr w.H, Freeman, San Francisco, 1969, p. 163.
2
11. Based on a solar constant of 1.97 cal./min./cm ,
the earth'8 albedo of 0.39 and earth's surface
area of 5.2 x 10® km^.
SUMMARY
Pollution coefficients, the fractions of transfor-
med energy that pollute the environment, have been
calculated for power generating processes by determin-
ing the total energy released and the polluting energy.
The total energy la taken as the caloric energy of
reaction plus the standard heats of formation of the
ultimate products and/or the energy released by sub-
sequent reaction of the Initial reaction products,
when the products are pollutants. The pollution energy
is the total energy less the useful energy, or that
energy for which the transformation was intended.
The Q values (pollution coefficients) determined
for four p power generating processes show that between
71% and 91% of the total energy transformed Is poll-
ution energy.
Although nuclear fission yields the smallest poll-
uting energy, it, like the other processes, Is unable
to supply man's power needs without dangerously poll-
uting his environment to the point where the environ-
ment will be unable to sustain life Only solar energy
seems to provide any real hope for the future.
REFERENCES AMD NOTES
1. International Critical Tables. Vol. II, McGraw-Hill,
New York, 1927, p. 137.
2. ibid, Vol. V, p. 162.
3 C.M. Summers, Scientific American. Sept. 1971, pl37.
4. H. Davis, Ed., Coal Age, personal communication.
5. Ref. 3, p. 151.
3
17-6
-------
AIR POLLUTION IN THE VENETIAN AREA
S. CERQUIGLINI MONTERIOLO
Istituto Superiore di Sanitk
v.R.Elena 299 - 00161 ROMA
E. BIANCO
Tecneco (ENI)
p.t.47 - 61032 Pano (Pesaro)
Summary. Data for the first two years
of a research program on the air pollution
situation in the Venetian area are elabora-
ted and compared with the air quality stan-
dards of the EPA and the air quality cri-
teria of the WHO. The area covered by the
research program (300 km^) includes the
city of Venice, the inhabited centres of
its immediate hinterland includino the in-
dustrial zone, and some of the Venetian is-
lands .
The Istituto Superiore di Sanity (ISS)
in collaboration with the Ente Nazionale I-
drocarburi (ENI) have in progress a three
year research program for the study of the
air pollution in the Venetian area and its
immediate hinterland. To carry out this re-
search an automatic monitoring network for
sulphur dioxide has been installed; the net-
work is completed in some of the sampling
stations by the continuous determination of
suspended particulates and the monthly col-
lection of rain and dustfall. Since the be-
ginning of 1975, the nitrogen oxides are au-
tomatically monitored through chemiluminis-
cence.,
Monitoring network. The basic monito-
ring network is made up of 34 air condition-
ed housing cabins, 24 of which equipped
with automatic coulometric analysers for sul-
phur dioxide. The monitored area (300 km2)
include the city of Venice, the islands of
Murano, Lido arid S.Elena, urban centres as
Mestre and Marghera, small inhabited zones
in rural areas,arid the industrial zone.
The monitoring was initiated in January
1973 with 10 stations; by the beginning of
1974 the 24 stations were,all at work. The
siting of the stations has been programmed
a) with the help of a Con.ca.we mathematical
model using emission data for the industrial
zone, which roughly lies in the geographic
centre of the zone being studied, and b) ta-
king into account the need of a suitable mo-
nitoring in the more densely populate^ areas
(Venice, Mestre, Marghera). or in those where
there is a particular interest such as ar-
tistic (Venice) or touristic (lido).
The analogical output from the analysers
is transformed and teletransmitted via tele-
phone line to an Operative Center in Venice,
where it undergoes further transformation and is
continuously fed into 24 recorders, as well
as to a computer which samples with a fre-
quency of 1 signal per sec. The computer
pilots a key-board which prints the hourly
mean concentrations and, each 24 hours, the
data for maximum and minimum in the day
(1 hour mean concentration) as well as the
data for the daily mean . Continuous indica-
tion of the values for the 30 min mean higher
than 0.3 ppm is also printed.
The following meteorological data, con-
tinuously monitored, complete the analytical
ones: direction and speed of wind, tempera-
ture, atmospheric pressure and humidity, amount-
of rainfall and presence of fog and cloudi-
ness.
Results of two first years of monitoring
The daily data obtained are printed by the
local press and the trends of the mean con-
centrations for 1 hour are exposed to the
public in.special show cases situated in the
historical centre of Venice. All the collec-
ted information is annually published in
special volumes which also contain the ela-
boration and evaluation of the results ob-
tained. Publication of a final complete
study is foreseen at the end of the programme.
The extension of this, report allows only to
present some observations on the results of
the first two years of sulphur dioxide moni-
toring. As far as the other parameters stu-
died are concernedj and in order to give a
general picture of the situation, it is use-
ful to summarize briefly the following data.
Suspended particulates continuously monitored
are, in Venice, in the range from 50 to 100
fxg/m.3 (daily mean value) in summertime, rea-
ching occasionally 250-300 jug/m^ in winter.
For the hinterland areas of Marghera and
for the industrial zone, the daily mean va-
lues for suspended particulates lie in the
range from 50 to 250p.g/m in summer, reaching
maximum values of around 350 and even 400
yug/m^ in winter. The dustfall ranges from
minimum values in Venice (from 0.033 to
0.084 g/m.d) to maximum values in Marghera
(from 0.140 g/m?d to 0.917 g/m?d); to noti-
ce that the last inhabited centre is situa-
ted in immediate contact with the industrial
18-1
-------
VENEZIA
FIGURE 1.- Location of sampling points in the monitored
area. Black circles indicates stations started in 1973•
zone. The pH of the collected rain is in the
range from 4.4 to 7.4; rain analysis reveals
a greater concentration of pollutants (sul-
phates, chlorides, fluorides) in the areas
adjacent to the industrial zone. As has been
already mentioned the monitoring of NO and
NO^ was started in 1975.
The data from sulphur dioxide monitoring
have been elaborated by calculating the cu-
mulative frequency of daily means which are
not higher than some reference values. For
the choice of these,,the WHO air quality
criteria3 and the EPA^air quality standards
for sulphur dioxide were taken into account.
Data has been also directly compared with
the above standards and with the first level
of the quoted criteria on the basis of the
values for the annual arithmetic means and
from the indications to the number of times
in the year in which daily mean values has
exceeded 0.140 ppm. The elements for this
comparison are reported in table I for 1974
for those sampling points for which a whole
year's measurements were available. From
Table I data comes the fact that Venice, the
islands and the inhabited minor centers of
the hinterland can be considered relatively
clean areas as far as pollution by sulphur
dioxide is concerned. However, it is not so
for the inhabited areas near the industrial
zone (Mestre and Marghera) nor for the in-
dustrial zone itself.
The cumulative frequency distribution
of daily mean data within concentration of
sulphur dioxide classes has been carried out
for every year and for every sampling sta-
tion according to the following reference
values: 100, 250, 365 and 500^ug/nr (0.038,
0.096, 0.140 and 0.192 ppm). The first re-
ference value as a cognitive limit, the se-
cond and fourth taken from air quality cri-
teria and the third from the EPA primary
air quality standard. The remarkable amount
of elaborated data (3515 daily averages for
2
18-1
-------
TABLE I
AREA
Sampling
station
NUMBER
SULPHUR
DIOXIDE
ANNUAL
MWHKET1C
MEAN
PPH
SULPHUR
0I0XIDE
DAILY AVERAGE.
HUK8ER OF DATA
HIGHER THAN
0.14 m IN 0.
ne ran (197+)
VENICE AND
22
0.029
1
ISLANDS
24
0.028
0
26
0.024
1
21
0.032
2
28
0.028
0
31
0.024
0
INLAND MINOR
13
0.029
0
CENTERS
20
0.031
0
MIXED AREA
7
0.030
0
WITH URBAN
15
0.033
1
PREVALENCE
16
0.047
24
17
0.034
2
MIXED AR3A
3
0.033
0
WITH INDU-
9
0.053
1
STRIAL PRE
10
0.049
1
VALENCE
11
0.041
0
INDUSTRIAL
2
6
29
30
,0.042
0.049
0.061
0.056
1
1
24
5
NOT CLASSI
PIED ~
18
21
0.053
0.034
2
10
Reference values
0.03 ppm - annual arithmetic mean (EPA na-
tional primary ambient air qua-
lity standard)
0.14 ppm - daily average, not to be excee-
ded more than once per year (EPA
national primary ambient air qua
lity standard) ~
0.038 ppm - annual arithmetic mean concen-
tration (OMS air criteria)
1973 and 7595 for 1974) are reported in the
annual volumes already mentioned (in Ita-
lian) which can be sent on request to who-
ever interested. In this report, only the
results are presented of the cumulative fre-
quency distribution of data for the stations
which are worked for two consecutive years,
classified in groups of stations with analo-
gous characteristics. To carry out this
grouping, the nature of the area covered
3
by each station has been taken into account
as well as aa the value of the ratio of
mean 3 summer months concentration and mean
3 winter month concentration for sulphur
dioxide. The calculated ratios are reported
in Table II both for each station and for
the grouped stations (area). The ratio's va-
lues can furnish a certain indication as to
the prevalence in each area of pollution
coming from permanent sources of emissions
(presumably industrial) and from seasonal
sources (presumably domestic heating).
The cumulative frequency distribution
of sulphvr dioxide daily mean data for 1973
and 1974 is graphically represented in fig.
2. Prom the reported distribution it is easy
to see that the quality of the air, accepta-
ble in the Venice zone, undergoes deteriora-
tion in the mainland urban areas in the
neighbourhood of the industrial zone and in
the industrial zone itself. An increase can
be observed in the percentage of days with
sulphur dioxide concentrations lower than
TABL3 II
1973
1974
AREA
8
K
si §
55
CO se
R
EACH
STATION
R
AREA
s
g
i e
a. co
*: x
< s
w se
R
EACH
STATION
R
AREA
VENICE AND
ISLANDS
22
24
4.00
5.36
4.75
22
24
26
27
28
31
3.73
2.83
4.57
10.47
9.25
3.11
5.66
INLAND i.. "OP.
CENTERS
13
20
4.91
4.17
4.54
MIXED AREA
WITH URBAN
PREVALENCE
16
17
8.18
3.60
5.53
7
15
16
17
3.06
2.94
6.6 7
5.33
4.50
MIXED AREA
WITH INDU-
STRIAL PRE
VALENCE "*
9
10
1.48
2.14
1.80
3
9
10
11
1.76
1.49
1.82
2.06
1.78
INDUSTRIAL
2
6
29
30
2.05
2.37
2.10
1.27
1.88
2
6
29
30
1.47
2.00
3.21
1.58
2.06
NOT CLASSI
PIED
18
21
1.71
8.00
B _ winter average (3 months)
summer average (3 months)
18-1
-------
VSNICS (STATIONS 11-14
r
%
I N III
MSSTM ( STATIONS IS — 1T>
r
i n m iv
MANONSNA (STATIONS*-«*> INOUSTHIAL ZONI (ST M-m-M9 90i
r*
% %
I
I
FIGURE 2.- Cumulative frequency distribu-
tion of daily mean concentrations for sulphur
dioxide: [J 1973, 10 1974.
Distribution classes: data lower than: I, 100;
II, 250; III, 365 and IV, 500 ^ig/m3.
the second level in 1974 compared with 1973•
However, such an improvement is difficult to
interpret, due to the diversity of meteoro-
logical conditions in the two years. A dis-
cussion of the meteorological data is not
possible in the space given to this report,
and again we must refer to the annual vo-
lumes quoted for a complete elaboration.
We can only indicate here that the num-
ber of hours per year in which each area
has been exposea to wind from the industrial
zone is higher for 1973 than for 1974. Fur-
thermore , the total number of hours with
meteorological calm is higher in 1974 than
in 1973, while the percentage of rain, fog
and mist does not show significative dif-
ferences in the two years considered (rain
5.4$ in 1973 and 6.0$ in 1974; fog and mist:
16.9$ in 1973 and 18.7$ in 1974).
References
1. Stichting Con.ca.we, "The calculation
of atmospheric dispersion from stacks",
The Hague, august 1966.
2. ISS - ENI, "Indagini sullo stato del-
1'inquinamento atmosferico nell'area di
Venezia", Vol. I, Roma, 1974.
3. World Health Organization, "Air quality
criteria and guides for urban air pollutant"
Geneva, 1972.
4. Environmental Protection Agency, "Natio-
nal Primary and Secondary Ambient Air Qua-
lity Standards", Federal Register, 36:8187
(1971).
4
18-1
-------
EVALUATION OF AIR QUALITY IN THE VICINITY OF AN INTERSECTION
IN WASHINGTON, D.C,
Donna M. O'Toole and Ronald C. Hilfiker
Environmental Research § Technology, Inc.
Concord, Massachusetts
Introduction
In recent years the variation in carbon monoxide
concentrations in urban areas has been studied exten-
sively. These studies have indicated that carbon
monoxide levels vary significantly in the vicinity of
urban intersections. In an effort to quantitatively
evaluate the spatial and temporal variations of
carbon monoxide in the vicinity of the intersection
of Wisconsin and Western Avenue N.W., an air quality
measurement and analysis program was performed. The
objectives of this program were to: 1) evaluate car-
bon monoxide concentrations in terms of compliance
with ambient air quality standards in the vicinity of
a roadway intersection, 2) perform continuous monitor-
ing of carbon monoxide levels in the vicinity of this
intersection over a 4-week period, and 3) evaluate
and understand the results of this program. The
measurement program was conducted between April 29
and May 24, 1974. Continuous measurements were made
of CO concentrations, utilizing both stationary and
portable analyzers. This paper details the monitor-
ing program employed and an evaluation of the data
obtained.
Summary
The measurement and analysis program conducted
to evaluate the spatial and temporal variation of
carbon monoxide levels in the vicinity of the inter-
section of Wisconsin and Western Avenues fulfilled
its objectives. The monitoring program provided data
which permitted the evaluation of air quality levels
both at the 5- and 30-foot heights.
An examination of the monitoring data indicated
that while the 1-hour NAAQS* was not exceeded, 8-hour
average concentrations greater than the 9 ppm NAAQS
were experienced at and in the vicinity of the inter-
section. The maximum 1-hour average monitored during
the study was 24.2 ppm. The highest 8-hour averages
monitored at and away from the intersection were
16.62 ppm and 15.1 ppm, respectively. An analysis of
8-hour concentrations greater than 9 ppm indicated
that concentrations at the intersection were of a
greater magnitude, occurred more frequently and were
of a longer duration in comparison to concentrations
monitored away from the intersection.
The diurnal Variation of CO concentrations cor-
related with traffic volumes. This correlation was
observed in the concentrations monitored at the
intersection at both the 5- and 30-foot heights.
Additionally, this correlation was indicated by
concentrations measured at the 30-foot level external
to and in the interior of a building 500 feet from
the intersection.
The spatial variation of CO concentrations in the
vicinity of the intersection exhibited no single pat-
tern. However, CO concentrations usually peaked at
the intersection and significantly decreased within
300 feet of the downwind corner of the intersection.
The concentrations in the vicinity of the intersection
were significantly influenced by local meteorological
conditions and the volume of traffic.
Site Description
The intersection at Wisconsin and Western Avenues,
N.W. is a signalized at-grade, urban, heavily traveled
crossing. Wisconsin Avenue is a six lane, two direc-
tional roadway which carries 30,000 vehicles per day,
and Western Avenue is a four to five lane, two direc-
tional roadway which carries 24,000 vehicles per day.
The intersection is situated within a central
business district surrounded by single family resi-
dential neighborhoods. The building configurations
at the intersection consisted of a two-story depart-
ment store (west corner), a one-story restaurant with
a peaked roof (north corner), a gas station (east
corner) and a construction site (south corner). For
up to 500 feet around the intersection, buildings did
not exceed a three-story height so that an urban
street canyon effect is minimal.
Measurement Program
As previously mentioned, a significant portion
of this study included a field measurement program
for the purpose of providing continuous data for the
analysis of carbon monoxide levels at Wisconsin and
Western Avenues, N.W. Carbon monoxide concentrations
at the intersection are attributable to vehicular
traffic flow through the intersection and background
CO concentrations. Background CO levels may be
defined as those levels of pollutant concentration
not directly attributable to an identifiable source
or combination of sources within the study area.
Both background and intersection related concentra-
tions vary daily as well as diurnally. The monitor-
ing site locations and configurations were chosen
upon consideration of the above factors as well as
expected traffic and meteorological conditions and
the topography in the vicinity of the intersection.
Instrumentation and Procedures Utilized During
the Monitoring Program
A portable instrument shelter on a trailer was
set up on the east corner of the Wisconsin and Western
Avenue intersection. The instrument used to monitor
carbon monoxide concentrations at the intersection
was a URAS-2 non-dispersive infrared gas analyzer
(NDIR) manufactured by Intertech Corporation. The
URAS-2 analyzer was modified with a timer and multiple
sampling pump to enable two sampling probes to be
attached to it. One sampling probe was mounted on a
20-foot portable tower set on top of the 10-foot high
shelter to allow measurements to be taken at the 30-
foot level. Another extended frpm the intersection
side (SW) of the trailer at approximately the 5-foot
level. The monitoring cycle consisted of sequential
sampling at each height foT approximately fifteen
minutes. The hourly averaged concentration at a
given height was determined by averaging the measure-
ments made at that height during the previous hour.
Wind measurements were made at the 30-foot level
with a CLIMET Instrument Wind System. The recorder
was inside the shelter and continuous simultaneous
readings were registered for wind speed and direction.
~National Ambient Air Quality Standards
1
18-2
-------
A second URAS-2 was installed on the third floor
of Washington Clinic. The Clinic, a three-story
structure, is located Northeast of the intersection,
at the corner of Military Road and Western Avenue.
The URAS-2 monitored ambient air inside the Clinic
for the first 18 days with random sampling made out-
side a window, which was adjacent to Western Avenue
and approximately 30 feet above street level. For the
remaining portion of the program, the intake of the
URAS-2 was left outside the building to enable compari-
son between the 30-foot recordings at the intersection
and the 30-foot readings at the Clinic approximately
500-feet Northeast of the intersection. This instru-
ment was equipped with a timer so that it automatically
checked the baseline every six hours with CO-free
nitrogen.
To supplement the CO monitoring at the East
corner of the intersection, five Ecolyzers with
Simpson Electric Strip Chart Recorders (Model 2750)
were used. The first measure taken to insure maximum
data capture and reliability was to equip each Ecolyzer
with an auxiliary D-cell battery pack. Next a humidity
control unit was mounted on each instrument. These
modifications, along with the recharging of batteries
and on-line electricity hook-up, when instruments
were not monitoring, insured accurate data capture
with a minimum of down-time. Each day before the
monitoring network was set up, each Ecolyzers' base-
line and span readings were checked. During the
monitoring period, periodic visual checks were made
to verify data capture. While monitoring, the Ecolyzers
and recorders were kept in a shelter. From the top
of this shelter extended a sampling probe located at
approximately the 5-foot level. This is the approxi-
mate level at which pedestrians would inhale the
ambient air.
In order to validate the Ecolyzer readings, each
instrument was run in parallel with the URAS-2 at
the intersection for at least 9 hours. The simultan-
eous readings for these time periods were compared.
An analysis of the linear relation between each of
the Ecolyzers and the NDIR resulted in correlations
from +.90 to +.98. A linear regression analysis was
computed for each instrument so that the Ecolyzer
readings could be calibrated. This numerical adjust-
ment was minimal. When the zero drift was checked on
the instruments, it was found to be almost negligible
on the days of valid data capture.
All of the above-mentioned procedures and pre-
cautions were implemented to assure the best possible
data capture and the least possible amount of down-
time.
Monitoring Configurations Design
The design of an adequate monitoring network is
generally constrained by the cost. Due to the cost
of the reference method systems, only two of these
instruments were used for this study. The reference
method instruments were supplemented with portable CO
analyzers. Five portable CO analyzers were available
for this study. Since one of the portable instruments
was always running in parallel with the NDIR in the
trailer for calibration, there remained four Ecolyzers
for ambient air monitoring. To optimize the use of
these instruments, a system of monitoring configura-
tions, which was wind and traffic dependent, was de-
signed.
The first parameters considered when designing a
measurement program to monitor maximum peak level CO
concentrations at the intersection were wind direction
and speed. The general horizontal wind flow and the
intensity of local turbulence are responsible for
dispersing CO from a source region such as a road.
With these considerations in mind, monitors were
stationed so that they were downwind of the inter-
section. In order to differentiate between concen-
trations attributable to intersection emissions and
those resulting from background sources, monitors
were deployed both upwind and downwind from the
intersection. Table 1 lists the monitoring config-
urations determined for this study.
The next parameter that was considered in de-
signing the monitoring network was traffic. The
volume and spatial distribution of traffic yields
significant effects upon the near field CO concen-
tration values. Higher values for CO are usually
observed when traffic levels are at their highest,
that is, at morning, lunch and evening commuter rush
hours.
The traffic at the intersection of Wisconsin and
Western exhibits a diurnal pattern. This pattern
exhibits significant levels of vehicles between 7 AM
and 11 PM with peaks during 7 AM to 9 AM, 11 AM to
1 PM and 4 PM to 6 PM. From this information, three
nine-hour intervals were designated for monitoring.
These intervals were 6 AM to 3 PM, 11 AM to 8 PM and
3 PM to 11 PM. The first time period provided data
to evaluate the build-up in CO concentrations from
minimal traffic flow through the lunch time traffic
peak. The second period provided data on the lunch-
time CO levels in comparison to the evening rush hour
CO levels. The third period provided data on the
expected maximum CO levels, from peak traffic volume,
and the necessary time for CO levels to drop off from
this expected maximum.
Having defined the sampling periods, it was
necessary to determine the priorities for these
periods. Since Washington, D.C. is southeast of the
intersection, the majority of the morning commuter
traffic heads southeast while the evening commuter
traffic heads north. To monitor peak CO levels at
the intersection, it was most appropriate to monitor
when the wind was parallel to the street. With the
wind from a direction of 255° to 75° (approximately
WSW to ENE), it was more appropriate to monitor with
traffic heading south. This was more predominant
during the 6 AM to 3 PM sampling period. Therefore,
for WSW to ENE winds, a 6 AM to 3 PM sampling period
was considered the first priority, 11 AM to 8 PM the
second priority and 3 PM to 11 PM the third priority.
With the wind from a direction of 75* to 255° (approx-
imately ENE to WSW), it was appropriate to monitor
traffic heading north. This is more predominant
during the 11 AM to 8 PM sampling period. Therefore,
for ENE to WSW winds, an 11 AM to 8 PM sampling
period was considered the first priority, 6 AM to
3 PM the second priority and 3 PM to 11 PM the third
priority.
The frequency of monitoring in any one sampling
period, for each wind sector, was defined as 3 sep-
arate days. If any configuration was monitored three
times in its first priority, the monitoring proceeded
to its second priority. After three 9-hour samplings
in its second priority, the monitoring proceeded to
its third priority sampling period. This allowed for
the distribution of the monitoring throughout the
6 AM to 11 PM period. Table 1 lists the sampling
period priorities defined for each wind sector. (It
must be noted that due to wind direction variability,
monitoring never proceeded to the third sampling
priority.)
2
18-2
-------
TABLE 1
WISCONSIN AND WESTERN AVENUE N. W.
ECOLYZER LOCATIONS AS DETERMINED BY WIND
DIRECTION, WIND SPEED AND TRAFFIC DISTRIBUTION
Wind Monitoring Sampling Time Priority
Direction Configurations 6AM-3PM 11AM-8PM 3PM-11PM
(deg)
345°- 15°
A, G,
C, H
15°- 45°
A, H,
N, Q
45°- 75°
B, I,
M, W
75°-105°
B, D,
I, J
105°-135°
B, G,
J, R
135°-165°
C, L,
U, X
165°-195°
A, C,
L, M
195°-225°
C, I,
M, Y
225°-255°
A, H,
N, S
255°-285°
D, N,
T, U
285°-315°
A, L,
U, V
315°-345°
A, G,
J, P
CALM
E, F,
K, N
Alternate
Schemes
Category A
N, B,
A, I
C, A,
U, G
A, C,
Y
H, A,
D, N
Category B
L, U,
V, Al
N, Z\
. z2'
3
3
3
3
3
3
3
3
3
3
3
3
3
I, Nl U, BKjT bk2
B, C, D, PL, RE, CLi, CL3, BK3
A, B, C, N, M, Mi, M2, AL2
V
SHOPPING
COMPLtX
~7n:
LoAnJ
mrnmI inr~in
¦ Stationary NDIR
A Eco/yzer Monitoring Sites
• Alternate Eco/yzer
Monitoring Sites
Figure 1 Monitoring Site Locations In the Vicinity
of the Intersection of Wisconsin and Western
Avenues, N. W.
In addition to the above, two categories of
alternate schemes were determined. The first cate-
gory was determined when the wind fluctuation occur-
red within two of the defined wind sectors. With
this fluctuation, it was necessary to set a scheme
which would permit data capture in the dispersion
path of this "new" wind sector. These configurations
are listed under Alternate Schemes - Category A. A
scheme of alternate configurations were determined
to supplement the study with data concerning off-
road concentration levels. These configurations are
listed under Alternate Schemes - Category B. See
Figure 1 and Table 1 for site locations, monitoring
configurations and time priorities used during the
Wisconsin and Western Avenue monitoring program.
Data Evaluation
Analysis of Compliance With Standards
During the entire 4-week monitoring program,
the 1-hour CO standard of 35 ppm was never exceeded.
However, the 8-hour standard of 9 ppm was exceeded
both at the intersection and at sites away from the
intersection. Table 2 lists occurrences of concen-
trations in excess of the 8-hour NAAQS monitored at
each of the intersection corners.* It must be noted
that this table does not list all the occurrences of
concentrations in excess of 9 ppm possible during
the monitoring period. Due to the wind variability
and the various monitoring configurations utilized,
all four intersection corners were not monitored
every day of the monitoring period. However, site 1
was monitored every day of the 21 days monitored.
An analysis of these 21 days indicated 14 days (67%)
with an 8-hour CO concentration in excess of 9 ppm.
Table 3 lists the chronological occurrences of
CO concentrations in excess of 9 ppm at monitoring
sites away from the intersection*. A comparison of
Tables 2 and 3 indicates that the occurrence of
concentrations greater than 9 ppm at the inter-
section are higher and more frequent than off the
intersection. Additionally, an analysis of running
8-hour averages indicated that concentrations greater
than 9 ppm were of a longer duration at the inter-
section than away from the intersection.
The data in Table 3 is divided into two cate-
gories: concentrations experienced at upwind and
downwind monitor locations. Since the intersection
is the major source of CO emissions in the vicinity
of the Wisconsin and Western crossing, any off
intersection concentrations in excess of 9 ppm were
expected to occur downwind of this source. The
upwind, off intersection averages greater than 9 ppm
are attributable to significant CO sources adjacent
to the upwind monitor. The one site where this was
distinctly evidenced is Site H. Site H is located
at the intersection of Western Avenue and 44th Street
(see Figure 1). The CO concentrations monitored at
H were approaching the level of Site N which was
across the intersection and just downwind of the
Western Avenue - Military Road intersection. Also
on April 29 and 30, the CO levels monitored at
Site H were higher than those at Site A, the south
corner of the intersection. The contributors to the
CO levels monitored at Site H are the vehicles
entering Western Avenue from 44th Street, the
*The data was compiled by calculating running 8-hour
averages and then proceeding sequentially through
the averages to an 8-hour average greater than
9 ppm. The next 7 hours of data were ignored when
determining the next occurrence of an 8-hour average
greater than 9 ppm.
18-2
-------
vehicles transversing Western Avenue and the vehi-
cles turning into the parking lot across Western
Avenue from Site H. The comparison of the contribu-
tion from these sources, as monitored at Site H,
with the coincident monitoring at Site A and N
indicates that the traffic patterns in the vicinity
of Western Avenue and 44th Street result in a signif-
icant contribution to CO levels in the vicinity of
the intersection of Wisconsin and Western Avenues.
This indicates that any evaluation of CO levels in
the vicinity of Wisconsin and Western must include
an examination of the significant concentrations
resulting from the secondary streets in the area.
Additionally, it indicates that the CO concentra-
tions experienced in the vicinity of the inter-
section are influenced by a number of "local" sources.
The significant local sources evidenced during
the monitoring program, in addition to Western
Avenue and 44th Street, were the intersection of
Western Avenue and Military Road, Willard Street and
Wisconsin Avenue and the Bus Terminal and Taxi
Station across Western Avenue from Washington Clinic.
TABLE 2
MONITORED 8-HOUR CO CONCENTRATIONS IN EXCESS
OF THE NAAQS AT THE INTERSECTION OF WISCONSIN
AND WESTERN AVE., NW, WASHINGTON, D. C.
Concentration
1
Hours
Date
(ppm)
Site
Monitored
April 29
13.36
1
(D)2
11-7
April 30
9.85
A
(U)
11-7
13.64
1
(D)
11-7
May 2
16.62
B
(D)
11-7
May 3
13.6
1
(D)
11-7
May 6
9.71
1
(D)
6-2
10.8
A
(D)
6-2
May 7
9.33
A
(D)
7-3
May 8
10.81
1
(D)
6-2
May 10
10.09
1
(U)
11-7
9.03
C
(U)
11-7
16.33
A
(D)
11-7
May 11
10.01
1
(U)
11-7
13.33
C
(D)
1-9
May 14
9.1
1
(D)
6-2
May 15
10.16
1
(D)
11-7
13.29
C
(D)
11-7
May 17
9.6
1
(D)
5-1
13.68
A
(D)
11-7
May 20
12.25
B
(D)
6-2
12.9
A
(D)
6-2
May 21
9.01
1
(U)
6-2
10.13
C
(D)
6-2
May 22
12.68
1
(D)
11-7
May 23
11.86
B
(U)
6-2
11.46
1
(D)
6-2
May 24
11.95
B
(U)
7-3
11.74
1
(D)
6-2
12.86
C
(D)
7-3
*See Figure 1
'Location with respect to the intersection and wind
direction. (D) - Downwind, (U) - Upwind.
TABLE 3
MONITORED 8-HOUR CO CONCENTRATIONS IN EXCESS
OF NAAQS
2
Downwind of the Intersection of
Wisconsin and Western Avenue
Date
Concentration
(ppm)
Site1
Hours
Monitored
April
29
9.9
N
11-7
April
30
11.14
S
11-7
May 1
9.8
U
7-3
May 2
14.88
J
11-7
May 3
11.48
N
11-7
May 8
11.83
E
6-2
2
Upwind of
the
Intersection
of
Wisconsin
and
Western Avenue
April
29
9.85
H
11-7
April
30
11.75
H
11-7
May 3
10.26
H
11-7
May 6
9.46
V
6-2
9.91
U
6-2
*See Figure 1
2
Location with respect to the intersection and moni-
tored wind direction.
Analysis of Maximum Concentrations
Having determined the relationship of monitored
CO concentrations to NAAQS, it is appropriate to
examine the magnitude and pattern of the highest CO
concentrations experienced during the program. The
highest 1-hour average CO concentrations ranged from
0.5 ppm, at a background site,* to 24.2 ppm at the
west corner of the intersection. The highest 1-hour
average concentrations monitored at each of the
monitoring sites were calculated. From this data a
histogram which compares the distribution of maximum
1-hour averages monitored at each site to the time
of occurrence was drawn (Figure 2). This chart
includes data from all the monitoring sites except
those inside buildings. This histogram indicates
the degree to which peak 1-hour concentrations are
contingent on traffic patterns. The two highest
frequencies of occurrence were monitored at the
midpoint of the morning and evening rush hours,
respectively. At the intersection, three of the
four corners registered peak 1-hour average concen-
trations at 5 PM. The occurrences at sites which
were monitored for three hours or less on any given
day are subtracted from the histogram (gray areas).
These values are subtracted from the histogram since
they do not contain an observation during peak
traffic, and, therefore, do not necessarily exhibit
the extreme value which could have occurred during
the peak emission time. The resulting histogram
closely resembles the diurnal variations which were
similarly exhibited in the traffic correlation
analysis.
*This background site was located approximately
1,600 feet away from the intersection. The monitor-
ing period at this site was composed pf three hours
of sampling. Each of the three hours experienced a
1-hour average concentration of 0.5 ppm.
4
18-2
-------
3hrs.
A similar tabulation and isopleths analysis
(Figure 4) were compiled for the 8-hour averages. The
highest 8-hour average CO concentrations ranged from
4.5 ppm, inside the clinic, to 16.62 ppm at the west
corner of the intersection. A similar trend of higher
values near the intersection is evident.
7 8
A M
u. ee
tt-83
*\ .5.5/
• tO. 48
10 II 12 I 2
Hour of the Day
Figure 2 Distribution of Maximum 1-hour Average CO
Concentrations Monitored At Each Site
Tabulated According to Time of Occurrence.
Figure 3 shows isopleths of the peak 1-hour
concentrations monitored at the S-foot level at
Sites I through Y. The higher CO levels resulting
from traffic at the intersection corners are exhibited
in these isopleths.
/
9 ppm
20 ppm
9 ppm
Figure 4 Maximum 8-Hour CO Concentrations Monitored
at Sites I through Y.
A
2000 2 8
o
o
1000 o 4
——- CO (ppm)
Traffic
Regression Line
CO* 2 06*.0024 Traffic
Correlation Coefficient
r* .82
6789 10 II 12 I 2345 6789
A M. pM
Hour of the Day
Figure S Average Diurnal Variation of CO Concentrations
at Site 1 and Total Traffic Transversing
the Intersection Monitored Between May 10
and May 24, 1974.
Figure 3 Maximum 1-Hour CO Concentrations Monitored
at Sites I through Y.
S
18-2
-------
Correlation of CO Concentrations to Traffic Volume
Emissions from vehicular traffic through the
intersection of Wisconsin and Western Avenues account
for a major portion of the CO concentrations moni-
tored in the vicinity of the intersection. An
examination of the monitoring data was performed to
determine the correlation of CO concentrations at the
intersection with the traffic transversing the
intersection. The traffic counts utilized for this
correlation were compiled by the District of Columbia
Department of Highways and Traffic (DCDHT).
The total traffic entering the intersection
exhibits periodic variation in volume as a function
of the time of day, as do the CO concentrations.
Figure 5 illustrates the diurnal variation observed
between May 10 and May 24, for each of the daily
monitoring periods for these two parameters.
Analysis of the Diurnal Variation of CO Concentrations
The CO concentrations monitored at the inter-
section display a periodic variation with the time of
day. Figures 6 and 7 illustrate the average of the
1-hour CO concentrations experienced during the 4-week
monitoring period. Both figures illustrate the
direct relationship exhibited between peak CO levels
and rush hour traffic volume.
The first diagram (Figure 6) illustrates a
comparison of the levels of CO monitored at the
trailer at the 5-foot (Site 1) and 30-foot (Site 2)
heights, and the CO monitored outside the clinic
(Site 5E), at an approximate 30-foot level. Curve 1
exhibits sharper fluctuations in concentrations
because it is monitoring closest to the "major source"
of emissions. Curve 2 is smoother in comparison to
curve 1 because the source of most of the emissions
is relatively distant and there is a greater degree
of mixing than at the lower height. At 500 feet away
from the intersection (curve 5E), the emissions from
the intersection have dispersed and mixed with the
background. It is evident from these graphs that the
CO concentrations at the three sites fluctuate with
traffic variability. The peaks of the three curves
coincide, but those of curves 2 and 3 exhibit a
slight time lag due to the relatively longer travel
times from the major CO sources to the respective
monitors.
A comparison of the CO levels at the 5-foot and
30-foot intersection heights and CO levels inside the
Clinic (curve 51) is illustrated in Figure 7. As in
the previous graph, there is a correlation between
traffic levels and the CO levels experienced at each
site. One consideration regarding the air monitor
inside the clinic is that the air intake for the
building was below ground level and parallel to
Military Road which, with the prevailing SSW winds
during the monitoring period, was not in the direct
path of the emissions transported from the intersec-
tion for the majority of the averaging periods.
A comparison of the graphs representing sampled
CO levels inside and outside the Clinic shows two
appreciable differences. The concentrations moni-
tored inside the Clinic exhibit a lower magnitude and
a time lag in fluctuation, with respect to the pat-
tern of outside levels. The ventilation system of the
building was primarily responsible for the differences
between inside and outside pollutant levels. The ven-
tilation system delays the response of inside concen-
trations to changes in the outside concentrations be-
cause of the time required for the mixing of incoming
air with the air already within the building. In
addition, the magnitude of the variation of inside
air quality levels was smaller than those outside
because of time response of the ventilation system to
abrupt change in levels. In the absence of a removal
process, however, we expect the long term average of
inside air to be close to that of the ambient air at
the ventilation intake.
Site / (5 ft level)
Site 2 (30ft level)
~~'— Site5F Outbids Clinic
(30ft level)
Hour of the Day
Figure 6 Average Diurnal Variation of CO Concentra-
tions Experienced During Coincident Moni-
toring at Sites 1, 2 and 5E.
Site / (5 ft level)
SiteS (30ft level)
~~ Site 5'! Inside Clinic
Hour of the Day
Figure 7 Average Diurnal Variation of CO Concentra-
tions Experienced During Coincident Moni-
toring at Sites 1, 2 and 51.
6
18-2
-------
5 »
J
I
— 1 i |
- At# Oft t)S'
1
o
1 '
H
1
N
"——
/
in
A
J_ 1
o iOO 4O0 700 9C
Upwind ¦
Figure 8 8-Hour Average CO Concentrations Monitored
on April 29 as a Function of Distance from
the Upwind Corner of the Intersection,
A.—tf W*4 109
Figure 9 8-Hour Average CO Concentrations Monitored
on April 30 as a Function of Distance from
the Upwind Corner of the Intersection.
4><>a«r * ¦>./ 'Smtm
An additional diurnal analysis compared the CO
levels monitored at the intersection with levels
measured at the Ecolyzer sites. This analysis indi-
cated that there was no general relationship between
concentrations monitored at the east corner of the
intersection and concentrations monitored away from
the intersection.
Analysis of the Spatial Variation of CO Concentrations
An analysis of CO concentrations, measured at
and away from the intersection, was performed to
determine the spatial distribution of concentrations.
The purpose of this section is to illustrate some of
the spatial patterns of CO concentrations evidenced
in the vicinity of the intersection. Five days of
8-hour averages were chosen as examples. The first
three days allow a comparison of CO level variations,
at the same monitoring locations, resulting from
changes in meteorological conditions and traffic
patterns. The remaining two examples permit an
examination of across-intersection build-up and
direct vs non-direct downwind dispersion of inter-
section generated CO concentrations.
The first three cases, April 29, 30 and May 14,
illustrate CO levels monitored at Sites H, A, 1, N
and S. The graphs of the three cases exhibit vari-
ations of a skewed bell-shape curve, each of which
illustrates the spatial variations of CO levels from
site to site (Figures 8, 9, and 10). The three
curves can be broken up into four sections, the
change in CO levels from H to A, A to 1, 1 to N, and
N to S. Each section exhibits a change in concen-
tration. The following is a discussion of the dif-
ferences in the change exhibited in each case.
The change in CO levels from H to A was negli-
gible on May 14. On April 29 and 30, concentrations
were higher at Site H than Site A. The May 14 moni-
toring was during a period of lower traffic volume,
there were less emissions from 44th Street traffic
and, therefore, lower concentration levels than on
April 29 or 30.
The build-up across the intersection of Wisconsin
and Western Avenues (Site A to Site 1) was similar on
April 30 and May 14 and only slightly greater on
April 29.
The drop-off from the peak level of concentra-
tions at Site 1 to Site N was similar on May 14 and
April 30. On both days, CO levels at the downwind
Site N drop off to those of Site A, upwind of the
intersection. April 29 exhibited an extended drop-
off so that the concentration levels monitored 390
feet upwind of the intersection were not achieved
until a distance of 700 feet downwind of the inter-
section. It is probable that this extended drop-off
was due to the greater wind speed as well as in-
creased traffic on Western Avenue downwind of the
intersection as compared to May 14 and April 30 wind
speed and traffic.
The increase in the concentration of CO from
Site N to Site S, which was similar on both April 30
and May 14 (see Figures 8 and 9), was attributable to
the traffic entering and leaving the taxi lane and
bus terminal concourse.
Figure 10 8-Hour Average CO Concentrations Monitored
on May 14 as a Function of Distance from
the Upwind Corner of the Intersection.
The next two examples. May 11 and May 10, allow
an examination of across intersection build-up, at
Wisconsin and Western, as a function of the vertical
distance from the upwind corner of the intersection.
The difference in build-up on these two days was
7
18-2
-------
that on May 11, the build-up from the upwind corner A
to Site 1 was sharp and on May 10, the build-up from „
the upwind corner C to Site 1 was gradual (Figures 11
and 12). On each day, the change in the concentra-
tion between the upwind corner and the downwind
corner were markedly different. This was due to the
greater wind variation exhibited on May 10 in compari-
son to May 11. On both days, Site 1 was 100 feet
from the upwind corner, but the expected similarities
in the change in the level of CO concentrations
between the upwind corner and Site 1 because of the
analagous horizontal distance were not exhibited.
To examine the rate of drop-off in CO levels as
a function of wind direction and distance from the
intersection, May 10 is presented as an example. On
May 10, monitors were set up at both Sites H and G,
which, though equidistant from Site A, experienced
different exposure to CO emissions. The average
monitored concentrations at these two sites differed
by 5 ppm (Figure 12). Site H monitored the dispersing
intersection emissions (the highest CO concentrations
due to the greatest traffic volume) as well as the
emissions from 44th Street, Western Avenue and the
parking lot across Western Avenue. Site G monitors
the dispersing intersection emissions and Wisconsin
Avenue emissions. Site H experienced higher CO
levels than Site G because hourly wind fluctuations
transported the emission from these multiple sources
to the Site H downwind location.
The examination of these five examples demon-
strates that there was no single pattern to the
build-up and dispersion of the CO concentrations in I •>
the vicinity of Wisconsin and Western Avenue. Vari-
ations in the meteorological parameters and traffic
patterns influence the change in concentration levels,
as did a monitor's distance from the intersection and
location with respect to wind direction and emission
source. Usually, CO concentrations peaked at the
intersection and decreased significantly within
300 feet of the downwind corner of the inter-
section.
>
- - - - ?
! ! | — '
i
ArtraQt Wind ftinea 5 8
•
1 ¦ 1
Dlitonc* (ft.)
Figure 11 8-Hour Average CO Concentrations Monitored
on May 11 as a Function of Distance from
the Upwind Corner of the Intersection.
\
j
| - Arrtaqr >¥mt iMtiiny, III'
| *¦¦>.> .;.««/ S <">l*
A
c
/ \
1
1- ,
Conclusions
The data obtained from this measurement program
adequately characterized the peak CO concentrations
experienced in the near field of the intersection
during April and May of 1974. These results indicate
that it is feasible to implement a program for the
evaluation of air quality at an at-grade urban inter-
section. The procedures employed in the design and
implementation of this measurement program are use-
ful as a guideline for developing urban intersection
monitoring programs. Further monitoring should be
conducted to characterize, as adequately as roadside
levels have been characterized, concentrations ex-
perienced in the vicinity of an urban intersection
at distances up to 100 feet off the roadways.
Figure 12 8-Hour Average CO Concentrations Monitored
on May 10 as a Function of Distance from
the Upwind Corner of the Intersection.
8
18-2
-------
OZONE IN RURAL AREAS
H. H. Westberg, K. J, Allwine,
R. A. Rasmussen, E. Robinson
Air Pollution Research Section
Department of Chemical Engineering
College of Engineering
Washington State University
Pullman, Washington
ABSTRACT
Ozone as an air pollutant initially was considered
a feature of congested urban areas of southern Califor-
nia. However, in the past few years, ozone concentra-
tions have been measured in excess of air quality
standards in many areas not previously identified as
regions affected by photochemical air pollutants.
Specifically, this includes rural midwest and east
coast areas of the United States. Possible sources
for ozone in rural areas include stratospheric injec-
tion, natural tropospheric photochemical reactions, and
the persistence of urban ozone as an airmass contami-
nant. Rural and urban measurements of ozone and other
relevant pollutants indicate that transport from urban
areas is the most reasonable mechanism for explaining
high ozone levels in rural areas. This paper will
integrate current field measurements with meteorology
to develop an urban-rural ozone model.
INTRODUCTION
The investigation of ozone as an atmospheric
constituent began in the early part of the 20th cen-
tury with the discovery that the cutoff of the solar
spectrum at around 3000A was due to the presence of
ozone in the atmosphere. The investigation of ozone
as a surface air pollutant was sparked primarily by
its participation in photochemical smog reactions and
its identification in the early 1950's as a major
component of urban air pollution in the southern
California area.
The United States Environmental Protection Agency
has established an hourly average of 80 ppb not to be
exceeded for more than one hour in any one year. In
recent years, violations of this standard have not been
confined to large metropolitan areas but have been
reported commonly in rural areas well away from major
urban areas. The two primary contributors to rural
ozone long have been considered to be stratospheric
injections and natural photochemical processes. How-
ever, neither of these, or even the combination of the
two, is expected to produce ozone levels over 80 ppb
except possibly on rare occasions. A third source--
namely transport from distant urban areas—recently has
been considered as a potential source of ozone in re-
mote regions. Inthe following discussion we will
examine the concentrations and sources of naturally
produced ozone as well as the evidence for significant
advection of ozone from urban areas into rural areas.
In this way we believe we can establish a relation-
ship of ozone concentrations in rural areas to ozone
precursors and ozone generation by photochemical
reactions in upwind urban areas.
OZONE IN URBAN ATMOSPHERES
Although we do not intend to dwell on urban-area
ozone behavior, it is necessary to discuss, in a general
manner, photochemical ozone production 1n industrialized
areas in order to document this particular source.
Since ozone was one of the initial materials which could
be measured quantitatively in urban atmospheres, a large
data base exists from which a basic understanding of
ground-level ozone patterns has evolved. In general,
a distinct diurnal pattern exists, with essentially
zero concentrations during nighttime hours and a rapid
rise during the daytime as photochemical processes be-
come important. A combination of urban-area nitric
oxide emissions plus shallow atmospheric mixing depths
is responsible primarily for the low nighttime levels
recorded at urban ground stations. However, the fact
that Og levels measured on the ground are low doesn't
mean ozone concentrations at 1000' above the ground are
also near zero. In fact, quite the opposite is true.
Aircraft soundings during nighttime hours generally
show a dramatic increase in 03 as the aircraft ascends
above the nighttime radiation inversion.
Investigation of urban ozone shows a distinct
diurnal pattern. Concentrations are essentially zero
during the nighttime hours, when no photochemical
reactions are occurring, followed by a rapid rise
of ozone concentrations during the daytime as photo-
chemical processes become important. Urban and rural
ozone concentrations were studied in Ohio during 1974
by WSU field parties. The research was done at Canton,
Ohio and at Wooster, Ohio, about 40 miles southwest
of Canton. Wooster is a rural site as far as its
local exposure is concerned, but because of the large
urban areas around it the larger area of Ohio must be
considered as a semi-urban area in almost all loca-
tions. Typically, a strong diurnal cycle in ozone
concentrations occurred at both the urban and the
rural site. There are differences, however, between
individual days and between sites that show the effects
of advection and transport in these areas. Some days
show local generation of photochemical oxidant. On
these days, concentrations were essentially zero during
the early morning hours until photochemical activity
began, about 9 a.m. Peak concentrations occurred about
noon but continued to rise slowly during the afternoon
hours and then fell off rapidly about sundown. On some
days, however, at Canton there were early morning
maximum concentrations observed. This is undoubtedly
an indication of ozone transport into the area and
vertical mixing of ozone from the layers aloft down
to the surface measuring station. Thus, in the case
of Canton, Ohio, this urban area is affected not only
by its midday photochemical reactions Involving local
sources, but also by advected ozone from some unidenti-
fied but more distant source.
RURAL OZONE MEASUREMENTS
It has long been recognized that the transport of
ozone or oxidant on the same day over distances of
several tens of miles has been seen in southern
California and other areas over a period of many
years.
Longer-distance transport of 0^, however, now has
become important in the Midwest and east coast areas
of the U.S. Here, 03 concentrations in rural areas
18-3
-------
were observed at levels that were significantly above
the 80 ppb air quality standard. These concentrations
thus required attention in order to determine the
nature of controlled regulations that might be applied
to bring the observed levels down to levels that would
meet the air quality standards.
WSU researchers measured rural ozone at Whiteface,
New York during 1974 and examined several instances of
long-distance transport of ozone. On July 24, the
research showed a wave of ozone moving into the study
site between 3 and 4 a.m., accompanied in this case by
elevated Freon-11 concentrations which confirmed urban
pollutant contributions. In this situation it is
suspected that the upwind source was the highly urban-
ized areas of New York City and its environments be-
cause of the fact that the wind trajectories showed the
sampled air mass had moved across the Mew York City-
Long Island area some 18 to 24 hours prior to when
the measurements were made. This rural New York study
shows that photochemical reactions can occur in rural
atmospheres and produce ozone concentrations from
apparently local emissions or natural components of the
rural atmosphere as well as from urban transport.
Rural ozone photochemistry was studied in more
detail in a measurement program at Elkton, a rural area
in southwestern Missouri. On the basis of Freon-11
concentrations, we believe that the air masses at this
site had been unaffected by urban pollutants. On
these days, ozone concentrations in the afternoon
frequently reached levels of about 50 to 60 ppb, rising
from nighttime minimum concentrations of about 20 ppb.
Low concentrations during the nighttime cycle are
believed to be indicative of a balance in the reactions
between stratospheric 03 and naturally emitted nitrogen
oxides. The natural ozone cycle shown for Elkton,
Missouri 1s not unique to this area, but has been
observed in measurements in a number of locations—in
particular in the Northwest, in both Oregon and
Washington. The natural or rural ozone cycle contains
components both of stratospheric ozone and natural
photochemical production.
Although strong inferences can be drawn as to the
nature of long-distance urban transport from surface
measurements such as in Ohio and New York, ozone
profile data through the surface layers of the atmo-
sphere provide a more definitive examination of the
ozone transport mechanism. In Canton, Ohio a combina-
tion of both surface measurements and aircraft 03
profile data was obtained for the city. The influx
of ozone aloft clearly was illustrated by high 03
concentrations aloft over the city when surface con-
centrations were low. The effect of transport in the
Canton, Ohio sector also was shown by the fact that
ozone sometimes would appear in high concentrations at
midnight or even later. Because of the wind patterns
existing over the Canton area at the time these measure-
ments were taken, it is considered highly likely that
the ozone observed aloft was a result of emissions
from an upwind city area and the photochemical forma-
tion of ozone during a previous period.
AN OZONE TRANSPORT MODEL
phase of the cycle has a sink at the ground surface.
In addition, reactions with gases of natural origin and
surface reactions on dust particles provide sink mecha-
nisms. We estimate that the natural 03 cycle may mova
through a concentration range from about 10 ppb to
maximum concentrations of 60 to 70 ppb.
Urban subcycle begins with a decrease to zero as a
result of reactions with large amounts of reactive
gaseous materials. A rapid increase to a peak value
follows as a result of photochemical activity in the
urban area. Where concentrations are most favorable
for large ozone development, 500 ppb 0, 1s a realistic
maximum value.
The third subcycle of the 03 rr,udel is the advec-
tion or long-distance transport of urban 0^ and 03
precursors. This phase has been recognized fully only
in the last few years. Here, downwind urban plume con-
centrations of 03 are depleted gradually by atmospheric
mixing and by destructive reactions involving gaseous
and solid surface removal mechanisms. Where the large
mass of ozone is protected from surface reactions,
such as during passage over water or over an area where
ground-level inversions limit surface contact, the
ozone may persist for periods of time greater than 24
hours.
A meteorological model of the advection of rela-
tively high urban ozone concentrations has been devel-
oped as a result of measurements during the Canton
investigations. This model relates the pollutant
levels to high-pressure areas moving across the ob-
serving network. Concentrations on the leading edge of
the high-pressure system were relatively low—essen-
tially background. As the high-pressure system moved
across the station, ozone concentrations increased
rapidly, apparently as a result of local pollutant
emissions. The advection transport phase of the high-
pressure system model occurs as the station comes under
the influence of the backside of the high with
southerly winds. The increased wind flow with the
approach of the following trough typically cleans out
local-emitted pollutants and also brings in 03 gener-
ated in source upwind and accumulated by the travel of
the high-pressure system.
CONCLUSIONS
The major concern at this time is for the source
of the observed elevated rural 03 concentrations. We
believe that these are due primarily to the transport
of urban pollutants in sufficient concentrations and
from a sufficient number of sources so that migratory
high-pressure systems achieve a widespread level of air
mass contamination. Thus, it is the urban areas that
are the sources of the rural pollutants, not vice versa.
The fact that both rural and urban areas have similar
day-to-day cycles of pollutant concentrations, as shown
by Wooster and Canton, Ohio, is an indication that
these air mass systems are large-scale systems and not
just local weather patterns.
Our 03 transport model has 3 subcycles—natural,
urban and rural. The natural cycle begins with strato-
spheric transport through the troposphere Into the
surface layers of the atmosphere. Ozone concentrations
In the surface layer levels of 30 to 50 ppb seem to be
encountered commonly as a result of stratospheric
transport. There also seems to be a natural photo-
chemical background input of about 20 ppb. The natural
2
18-3
-------
THE USE OF FISH TO CONTINUOUSLY MONITOR
AN INDUSTRIAL EFFLUENT
by
Garson F. Westlake, William H. van der Schalie,
John Cairns, Jr. and Kenneth L. Dickson
3iology Department and Center for Environmental Studies
Virginia Polytechnic Institute and State University
Blackshurg, Virginia 24061
Summary
In order to detect developing toxic conditions in
industrial waste discharges, a fully automated biolog-
ical monitoring system using sublethal responses of
fish has been developed. The installation and opera-
tion of this system at an industrial site are discuss-
ed. Data gathered continuously from the fish at the
industrial site are analyzed by a small computer which
automatically produces alarms at the site should toxic
conditions appear to be developing.
Introduction
The present level of surveillance of industrial
waste discharges is too irregular to detect many damag-
ing spills from point sources. Even when a spill is
detected, it is often too late to take any meaningful
corrective action. Chemical-physical monitoring fre-
quently does not enable one to predict the toxicity of
complex waste effluents. However, a biological moni-
toring system is able to integrate these effects and
provide an early warning of changes in toxicity.
A system using biological techniques to monitor
industrial waste waters has been under development in
1 O O
this laboratory for a number of years. ' In
laboratory simulation tests, the latter system has
been successful in detecting sublethal concentrations
of materials ranging from pure chemicals such as
zinc3 and phenol4 to complex waste effluents ("pink
water" and nitroglycerine waste) from a munitions
5 £
plant. ' A new and fully automated version of this
earlier biological monitoring system is now being
tested at an Industrial location plant on the New
River in Virginia. The implementation and operation
of this system is described below and the possible
application of such systems to other industrial
settings is discussed.
System Description
In the demonstration study at on-site
location itwas necessary to modify the laboratory-
designed biomonitoring system to make it compatible
with the industrial setting. Due to limitation on
space 1n the waste treatment plant area the biomoni-
toring system had to be made as compact as possible.
We found that a modular design allowed us to best
utilize the space as well as permitting us to check
out the equipment 1n the laboratory prior to installa-
tion at the site. The system (which is described
below) was designed to be inexpensive to construct
and relatively simple and trouble-free in terms of
mechanical and electronic complexity.
Modules
Figure 1 illustrates one of the three modules
installed in the plant. Each consists of four repli-
cate chambers in which a single tank 1s installed.
All electronics and plumbing is dlsconnectable so that
the system can be moved easily. The modules each
measure 203 * 84 x 78 cm. Each of the four chambers
open separately and is supplied with water and drains
and a fluorescent light attached to the top. Photo-
period is controlled by an automatic timer. A pro-
portion of the tested effluent is supplied to the
modules by means of a dilution system of original
design in the end of each module.
Figure 2 illustrates the dilution system. It is
much simpler than the normal dilution system used in
the laboratory because of the continuous supply of
effluent. The dilutor consists of a number of
chambers. Into chamber A the effluent is sent so that
it overflows into the drain, chamber C. Similarly,
the reference water flows into chamber B. The amount
flowing out of each chamber, A and B, is regulated by
means of two adjustable weirs. A baffling chamber D
insures mixing of the two. Constant heads are main-
tained in chambers A, B, and D so that even though
variations in the flow occur the portion is always
constant. Out of chamber D a number of lines run to
the tanks. The tanks in one of the three modules act
as controls, being supplied with unchlorinated well
water only. The second module is exposed to a
dilution of the same treated waste that flows into the
river. Untreated waste is tested in a third module to
give early warning of a possible increase in the
toxicity of the waste effluent.
Monitor Tanks
The monitor tanks themselves are specially de-
signed 40 liter fiberglass tubs (fig. 3). The bottom
slopes to a stand pipe in the corner which is
enclosed in a sheath that allows the water to withdraw
from the bottom of the tank. This combination makes
the monitor tanks self-cleaning even with daily feed-
ing. The automatic feeders currently being usedhold
a two week supply of food; fish are fed once a day.
During testing a single bluegill sunfish (Lepomis
macrochirus) is placed in each of the tanks. Respira-
tion is recorded by means of stainless steel elec-
trodes attached to each end of the tank. These elec-
trodes pick up minute changes in electrical potential
(1-2 microvolts) produced by the fish each time it
opens and closes its gill opercula. The number of
peaks in this signal per unit time represents the
respiration rate of the fish. A section of four inch
plastic pipe (not shown) supplied as a refuge for the
fish helps to improve the signal by minimizing move-
ment and changes in position of the f 1 sh.
Automated Recording Apparatus
The data are recorded by means of a microproces-
sor located on the site itself. Figure 4 1s a block
diagram of the entire system. The small signals pro-
duced by the fish and picked up by the electrodes are
amplified by means of an amplifier specially designed
for the purpose. Figure 5 shows the circuit diagram.
It essentially consists of two gain stages and two
low pass filters. The gain is approximately
107; the filters allow the operation 1n an industrial
setting which 1s high 1n electrical noise.
18-4
-------
The amplifiers, one for each tank, are multiplexed
through an analogue multiplexer and digitized by means
of an analogue-to-digital converter. Under the control
of the microprocessor digital samples are taken from
the signals at regular intervals from all of the tanks.
The microprocessor then computes the breathing rate by
means of a program and accumulates breathing rates for
half-hour intervals. At the end of each half-hour
interval, the mini-computer in the laboratory gathers
the respiration rates from the microprocessor over a
dedicated phone line. By a statistical technique
3
described previously , 1t is determined whether each
of the fish is behaving normally or abnormally. A
decision is automatically made by the computer based
on previously established criteria whether an alarm is
necessary. If so, a signal is sent back to the micro-
processor which in turn lights an alarm light in the
guard house in the plant. This guard house is manned
around the clock and the followup prucedure can be
initiated from there.
Follow-up to an Alarm Situation
After the light in the plant's guard house has
automatically been lit, the guard on duty calls the
person charged with the follow-up responsibility. This
person has in his possession a video-terminal with
which he can call to the central mini-computer over any
regular phone line. He is then able to determine which
- fish responded abnormally by examining the data just
collected or by viewing a trace of the fish's respir-
ation signal.
If a true alarm situation is found to exist, the
follow-up person can immediately start an automatic
refrigerated sampler which is tied into the waste
supply line at the industrial site. Sample analysis
may then be carried out as soon as possible to
determine the cause of the abnormal response of the
fish.
Results and Discussion
The biological monitoring system outlined above
has been installed and is currently being tested and
evaluated. At this time, enough data have not been
accumulated at the industrial site to allow a deter-
mination of the system's response characteristics to
developing toxic conditions. However, something can
be said about the types of materials to which the
system will and will not respond.
The waste materials entering the monitoring
system must of course be diluted considerably before
the fish are exposed to them; about a one percent
dilution is used at the industrial site. Thus, the
fish are not likely to be sensitive to fluctuations
in certain waste characteristics such as temperature,
pH, and dissolved oxygen. These parameters, however,
are already continuously monitored at the site by
conventional physical-chemical sensors which, like the
biological monitor, are tied to an alarm system in the
guard house. In this way, the biological monitor is
used to complement, rather than totally replace,
currently utilized physical-chemical sensors.
In many situations, in fact, biological monitor-
ing could prove to be more useful than presently
available continuous monitoring physical-chemical
systems, particularly where industrial effluents are
complex and subject to considerable fluctuation in
composition. It may be possible to optimize the use
of biological monitors in varied situations by using a
particular species of fish having appropriate response
characteristics to the principal components of the
waste being monitored. Another possibility being
explored in our lab is the use of invertebrates such
as the crayfish in a continuous monitoring set-up.
Whatever final form biological monitoring systems
take, they are a potentially valuable asset in the
field of water quality monitoring and control.
References
1. Cairns, J., Jr., K. L. Dickson, R. E. Sparks, and
W. T. Waller. 1970. A preliminary report on
rapid biological information systems for water
pollution control. J. Water Pollut. Contr. Fed.,
42:685-703.
2. Cairns, J., Jr. and R. E. Sparks. 1971. The use
of bluegill breathing to detect zinc. II. S.
Environmental Protection Agency, Water Pollution
Control Research Series, 18050 EDQ 12/71.
3. Cairns, J., Jr., J. W. Hall, E. L. Morgan,
R. E. Sparks, W. T. Waller and G. F. Westlake.
1973. The development of an automated biological
monitoring system for water quality. Office of
Water Resources Research Bulletin. Virginia Water
Resources Research Center, Virginia Polytechnic
Institute and State University. Bulletin 59.
4. Westlake, G. F., E. L. Morgan and J. Cairns, Jr.
1972. The use of a biological monitor to detect
phenol. Amer. Zoo!., 12:713.
5. Westlake, G. F., W. H. van der Schalie,
J. Cairns, Jr. and K. L. Dickson, (unpublished).
Responses of biosensors to toxic spills. I. TNT.
6. Westlake, G. F. , W. H. van der Schalie,
J. Cairns, Jr. and K. L. Dickson, (unpublished).
Responses of biosensors to toxic spills. II.
Nitroglycerine.
drain
line
1. An Effluent Monitoring Module, Showing Diluter and
One Monitoring Tank.
2
18-4
-------
chamber D
standpipe
chamber A
chamber C
chamber B
mixing
baffle
toxicant
to
tanks
dilution
water
drain
drain
2. Diagram of Effluent Diluter. (See text for explan-
ation).
VIRGINIA TECH
magnetic
tape
[teletype |
IBM
'370
computer
REMOTE INDUSTRIAL
SITE
DEC
pdp 8/c
minicomputer
leased telephone lino
Control Logic
microprocessor
converter
analog
temperature
conductivity
upstream effluent
water
4. Schematic Diagram of Monitoring System.
feeder
3. A Monitor Tank, Showing Path of Water
-------
SOURCES AND SAMPLING OF POLLUTANTS FOR GEOTHERMAL STEAM AREAS
Leonard A. Cavanagh and Ronald E. Ruff
Stanford Research Institute
Menlo Park, California 94025
Summary
Stanford Research Institute recently contracted
with a research group consisting of Pacific Gas and
Electric Company, the major utility of Northern Cali-
fornia, Union Oil Company, Burmah Oil and Gas, and
Pacific Environmental Corporation--major steam suppliers
in the Geysers geothermal area--to undertake a study to
characterize hydrogen sulfide (H2S) concentrations in
the area surrounding the geothermal development. Con-
centrations of H2S that exceed the Ambient Air Quality
Standards of the California Air Resources Board (0.03
ppm)* have been observed in the area of interest. The
data base will include two years of continuous monitor-
ing data from an eight-station network.
without combustion of fuel to produce steam, there are
no emissions of the familiar air pollutants: oxides of
nitrogen, oxides of sulfur, carbon monoxide, unburned
hydrocarbons, and particulates. Second, geothermal
emissions are nonuniformly distributed over a large area
whereas the fuel-fired plants are well characterized by
their stack locations and parameters.
The natural steam from a geothermal well is pre-
dominantly water vapor; however, the steam contains a
noncondensable gas portion, rock dust, and trace quanti-
ties of minerals. At the Geysers, the noncondensable
gas portion of the steam averages about 1 percent by
weight, with boron being the major trace element present.
The composition of the noncondensable gases contained in
the geothermal steam at the Geysers is shown in Table 1.
Background
Commercial utilization of geothermal energy to
produce electrical power has been economically feasible
for about 15 to 20 years, particularly in areas where
geothermal steam may be tapped. The energy crisis has
initiated considerable interest and activity in explor-
ing the potential of geothermal energy. Geothermal
areas that can provide dry steam from subterranean steam
reservoirs are considerably more rare than those areas
where hot water may be tapped as an energy source.
Historically, the first and until recently the largest
geothermal steam power plant was located near Lardarello,
Italy, producing 365 megawatts (MW) . The other major
geothermal power plant outside of the United States is
Wairakei, on the North Island of New Zealand, producing
192 MW from energy extraction of hot water.2'^ Within
the United States the only commercial production of
electrical power from geothermal energy is in the Gey-
sers area in Northern California. Other areas in the
United States, such as the Imperial Valley in Southern
California, are at present under consideration as poten-
tial sources of geothermal energy.
The Geysers is located in rugged mountainous ter-
rain about 80 miles north of San Francisco. Currently,
11 power plants are in operation in this area providing
492 MW. The first of these plants went on-line in 1960
and the complex today is the largest power generation
facility powered by geothermal energy in the world.
The steam used to drive the generators of a geo-
thermal plant Is obtained from wells drilled into reser-
voirs of steam or hot water lying hundreds to several
thousand feet below the surface. Early wells at the
Geysers were drilled adjacent to natural steam vents to
depths of 400 to 1000 feet. Currently, deeper steam
zones are being tapped with wells ranging in depth from
2000 to 9000 feet. A typical new well will produce
steam at a rate of 200,000 lb./hr. The output from
about 10 wells is required to operate a power plant of
100 MW capacity.
From the point of view of pollutant emission, some
obvious and some less obvious differences exist between
fuel-fired power plants and geothermal plants. First,
Table 1
NONCONDENSABLE GASES IN THE
GEOTHERMAL STEAM AT THE GEYSERS
Gas
Percent by Weight
Low
High
Design
Carbon,dioxide
0.0884%
1.90%
0.79%
Hydrogen sulfide
0.0005
0.160
0.05
Methane
0.0056
0.032
0.05
Ammonia
0.0056
0.106
0.07
Nitrogen
0.0016
0.064
0.03
Hydrogen
0.0018
0.019
0.01
Ethane
0.0003
0.002
—
Total
0.120%
2.19%
1.00%
Source: Reference 4.
The individual sources of steam emissions are many
and varied within a geothermal region. In the Geysers,
natural steam vents and fumaroles are numerous and the
magnitude of these natural emissions is difficult to
characterize. New wells and idle wells are other
sources of emissions. New wells are allowed to vent
directly to the atmosphere during the final drilling
phase for a period of one or two weeks to loosen and
blow off particulate materials. After the wells are
tested to ascertain production capacity, the output is
constricted to provide a controlled continuous vent to
the atmosphere until the well is brought on-line to a
power plant. In the case of the deeper wells, the
continuous vent rate of geothermal steam is 3000 to
4000 lb./hr.5
The steam from producing wells passes through a
centrifugal separator to remove condensed water and
particulate material before transmission to the power
plant. The steam vented at this separator is, in gen-
eral,' about 3000 to 4000 lb./hr. The insulated steam
lines that connect the individual wells to a 36-inch
main steam line leading to the power plant are frequently
up to a mile long.
The turbines typically operate at 100 lb./sq. in.
18-5
-------
at 355°F. The steam from the turbine exhausts to a
direct contact cascade condenser at a pressure of 4
Inches of mercury absolute. Steam ejectors purge the
noncondensable gases from the condenser unit. The con-
densed steam and cooling water are then pumped to
cooling towers. Under all operating conditions the
cooling tower evaporation rate is less than the tur-
bine steam flow into the cycle. Therefore, there is a
surplus of water representing about 20 percent of the
steam flow to the turbine. The surplus water is re-
turned to the steam field by means of reinjection wells.
Treatment of H2S Sources
From the viewpoint of monitoring and modeling, the
emissions from a geothermal field must be treated as a
large diffuse source. The rule of thumb for well
spacing is 20 acres per operating well. With about 10
wells required to supply each power plant, this repre-
sents an area source of over 200 acres emitting 30,000
to 40,000 lb,/hr. of steam from the centrifugal separa-
tors over and above the 2,000,000 lb./hr. supplied to
power plants of the latest design.
The emission sources of H2S in the Geysers can be
summarized in general terms as follows:
• Power units
• Producing wells
• Idle wells
• Well drilling
4800 lb. H2S/hr.
200 lb. H2S/hr.
100 lb. H2S/hr.
60 lb. H2S/hr.
Two factors dominate the transport of the HjS with-
in the area of interest:
• Concentrations are the highest in the summer
period when synoptic scale gradient winds are
light and local valley circulations predominate.
• Critical conditions arise when the gradient wind,
forces just exceed the drainage flow and cause
the H2S to be transported across the ridge line
into the populated area to the east.
A careful analysis of the terrain features shows
that any air flow west to east under the influence of
synoptic wind gradients will be channeled through a
limited number of key cols or saddles. This is shown in
Table 2.
Table 2
INSTRUMENT COMPLEMENT OF MONITORING STATIONS
The spacing for power plants 1b largely governed by
steam reserves and pressure drop in the steam mani-
folding from individual wells. Currently, power plants
of 100 to 110 MW capacity spaced about one mile apart
appear to be optimum.
Proven reserves in the Geysers can provide 1000 MW,
or double the present capacity. Exploratory wells indi-
cate that the steam field extends considerably beyond
the areas currently supplying operating units. Esti-
mates of the total capacity of the Geysers for power
production range as high as 2000 MW.®
Layout of Monitoring Network
Within this framework, SRI with the assistance of
Ball Brothers Research Corporation, will install and
operate a monitoring network to characterize the H2S
distribution within the area. Figure 1 shows the loca-
tion of the Geysers geothermal area. Figure 2 shows
the area where monitoring stations will be sited for
two years of operation.
The transport and diffusion of H2S in the rugged
terrain of the Geysers area is highly complex with com-
plete characterization on a microscale level requiring
unreasonably large resources. Figure 2 shows the topo-
graph y of the Geysers area. A major ridge line running
nearly north and south forms the border between Lake and
Sonoma counties. IMs ridge separates a populated re-
sort area to the east from an unpopulated area immedi-
ately to the west. A series of large parallel canyons
lead to the northwest from the ridge line. These can-
yons are about two miles from ridge to ridge and about
2000 feet deep. The power plants, shown in Figure 2,
vary in altitude from about 1400 feet to about 3200
feet. Due east of the Geysers complex is Cobb Mountain
at an altitude of 4700 feet.
Measurement
Instrument Parameter
Sites
123456789
Houston Atlas H2S
XXX X XX
825 R
Tracer 270HA H2S, S02,
X X
total S.
Anemometer Wind speed and
XXXXXXXXX
direction
Thermometer Temperature
XXXXXXXX
Hygrometer Dew point
XXXXXXXXX
Remote Rain Gauge Precipitation
XXXXXXXX
Acoustic Radar Temperature dis-
X
continuities
aloft—mixing
layer depth
Selection of appropriate sites is based on the ob-
jectives that must be met in the program, balanced by
certain practical constraints. The important purposes
and requirements include:
• Characterize H2S concentrations over the area
of interest
Define population exposure to $2®
Define meteorological conditions In the area
Define processes affecting transport and trans-
formation o£ H2S
• Define accessibility for maintenance, power and
communication.
Sites 1 and 2 are located in cole near the crest
of the Mayacmas Mountains north and south of Cobb Moun-
tain. This ridge effectively separates the main geo-
thermal area from the populated resort areas to the
east. Wind and H2S measurements at these sites will
define H2S transport out of the geothermal area and also
the path of travel around Cobb Mountain. Temperature
measurements at the ridge will provide a basis for
inferring the pressure of stable layers.
Site 3 provides a measure of the air movement to
the north from the Geysers area. This site will help
define H2S transport during periods of southwesterly
winds that sometimes precede winter storm systems.
Site 4 will help define the H2S concentration in a
populated resort and housing area. This site is located
3200 feet on the eastern side of the Cobb Valley--a
2
18-5
-------
relatively narrow valley running northwest to south-
east. Since Site 4 is located at a similar elevation to
Site 1, the data obtained here can be used to evaluate
whether the airflow passes over Cobb Valley without
following the underlying terrain. In such areas the H2S
may impinge on the mountain causing higher concentra-
tions than at the lower elevations in the valley. Pre-
liminary studies by PG&E and the odor complaint pattern
suggests this as a possibility.
Under conditions where the air has followed an ele-
vated trajectory, the air mass is largely isolated from
surface emission sources for much of the distance be-
tween the geothermal area and the monitor. For this
reason, a gas chromatographic sulfur analyzer, providing
measurements of H2S, SOj, and total sulfur, located at
Site 4 will generate less ambiguous data about transfor-
mation processes than it would at other potential loca-
tions .
Sites 5 and 6 will provide data that can be used
to determine the air flow patterns around Cobb Mountain
and the valley at its base. Site 6 is located where
nearby geothermal development is underway; it can pro-
vide a data base that can be used to evaluate the
effects on air quality of future developments. Sites 5
and 6 are both located near population centers and the
data generated can be used to evaluate population ex-
posure .
Site 7 is located on the floor of Cobb Valley and
can provide information about airflow following a tra-
jectory close to the underlying terrain. Also, in con-
junction with Sites 1 and 4 it will help define the
vertical diffusion of the trajectory over the col. The
acoustic sounder will be located at Site 7 to monitor
atmospheric stratification just above the valley floor
up to elevations above the ridge line. The relation-
ships between observed stratification above Site 7 and
that in the Geysers area will be established during the
special studies when two acoustic sounders will operate
simultaneously in these locations.
Site 8 will normally provide a measure of the H£S
concentrations entering the Geysers area before ex-
posure to the significant sources within the area.
During the rare periods of east winds, this site will
provide a measure of transport toward the west from the
Geysers area.
Site 9 contains meteorological sensors only, and
will provide measurements of wind speed and direction
and dew point. This site will be located on the ridge
near Units 7 and 8 to provide measurements of air flow
near the larger sources.
In addition to the monitoring network, an array of
10 Colortec H2S detectors are sited around the perimeter
of the network up to 20 miles from the Geysers. The
Colortec detector consists of a chemically treated pad
mounted on a paper card that provides a color indication
proportional in intensity to the H2S dosage during the
exposure period.
Data Acquisition and Analysis
The data acquisition and flow cycle is illustrated
in Figure 2. The primary transfer mechanism from the
automatic stations is accomplished by a radio telemetry
link to the combined monitoring/central data acquisition
facility at Site 4. The minicomputer in the central
facility checks and reduces the data to engineering
units and merges the data from all stations onto the
same tape. This tape is transferred to Menlo Park
where it is added to the data base resident on a larger
computer. From there, it is analyzed and reported to
PG&E. Other elements in the data flow include a back-
up recording system at each automatic station, tape
casettes from the special meteorological station (Site
9), and field-reduced Colortec cards. In addition,
segments of the data acquired automatically are trans-
mitted over telephone lines to SRI on demand from a
teletype port.
A requirement for 90 percent data validity dictates
a central data acquisition system collecting telemetered
data from the remote stations. The alternative of
manually retrieving data from each station, reducing and
analyzing it, and later merging this with data from
other stations is prohibitive costwise. The increase
in manpower for the additional station visitation and
data processing time is much more costly than the hard-
ware expense associated with the telemetered system.
The telemetered system also offers several techni-
cal advantages over a manual one. First, all data are
objectively reduced by one computer before display,
thereby reducing interpretational errors. Next, all
data can be reviewed simultaneously, thereby facilita-
ting identification of readings that are sifcpicious.
In addition to facilitating such identification, the
telemetered system, by virtue of its immediate display
capability, greatly reduces the time element in alerting
field personnel to problems. In fact, the central com-
puter is programmed to detect certain types of errors
and to send out alarm messages to that effect. With a
manual system, such errors would not be detected until
a routine visit by field personnel.
A radio link between central and remote stations
Is the logical approach to telemetry in this applica-
tion. A telephone linkage becomes impractical because
of the unavailability of telephone lines at some of the
preferred monitoring sites. The HjS and meteorological
data are transmitted every six minutes to the central
facility where they are converted to engineering units
based on the most recent calibration data.
The calibration data for H2S, and any other
optional automatic chemical analyzers, are transmitted
automatically on a daily basis. The central computer
checks these data for zero and span drift as compared
to the manufacturer1s published specification and the
instrument's previous performance history.
After the central station automatically reduces
the data to engineering units and runs initial validity
checks, the data are transferred to an industry stan-
dard magnetic tape unit with segments displayed on the
local teletype unit. Similarly, segments of the data
will be displayed and reviewed daily on the Kenlo Park-
based teletype unit. The magnetic tape represents the
central facility's final product. Every few days this
is transferred to the overall data base resident on a
CDC 6400 computer system.
The CDC 6400 facility is equipped with software to
retrieve and analyze the data for reporting purposes.
Reports will contain summaries of data validity achieved
and statistics from each Instrument and device (in-
cluding Colortec cards, special meteorological sta-
tion, and acoustic sounder); reports will also contain
3
18-5
-------
WHISPERING
PINES
ANDERSON
SPRINGS
Sacramento
Geysers
San Francisco
Au torn* tic MjS-SOj- rS'WueoiolOBiral Station*
Q Automatic H?S'M«t*
-------
computer-genetated lsoplethe of H2S concentrations
superimposed on objectively calculated wind streamlines.
Streamlines will be broken into categories that can be
well identified to occur frequently.
The analysis also includes a feasibility study on
the adaptation of an air quality model to the Geysers
area. The type of model most appropriate to this appli-
cation is a "Gaussian puff" formulation that is capable
of representing complex wind fields and topographic
features. A model with most of the requisite features
has been developed for the Tennessee Valley Authority
(TVA) Kingston Steam Plant.7 This model both predicts
pollutant concentrations at selected points In the area
and presents control alternatives to plant operating
personnel when adverse meteorological conditions occur
(i.e., those conditions that lead to violation of
federal air quality standards).
In adapting the TVA model to the Geysers area, the
major modification will be in upgrading the model to
include more complex sources, meteorological and topo-
graphical features specific to the area, and empirical
treatment of the chemical transformations. The identity
of the source (generating unit) is retained in the model
calculations such that the air quality calculation at
any point in space and time consists of both total con-
centration and breakdown of the contribution by each
major source.. The topography of the Kingston are a is of
similar complexity to that of the Geysers—being
characterized by numerous ridges and valleys. This
leads to similar complexities In estimating the wind
fields and other meteorological parameters.
References
1. Swanson, R.N, and M.L. Mooney, "Hydrogen Sulfide
Survey Geysera Area Report," Meteorological
Office, Gas Control Department, Pacific Gas and
Electric Company (January 1973),
2. Finney, J,P., "The Geysers Geotherroal Power Plant,"
Chem. Eng. Prog. 68 (7), 83-86 (1972).
3. Axtmann, R.C., "Emission Control of Gas Efluents
from Geothermal Power Plants," Environmental
Letters 6 (2), 135-146 (1975).
4. Finney, J.P., "Design and Operation of the Geysers
Power Plant," Geothermal Energy: Resources,
Production, Stimulation, P. Kruger and C. Otte,
ed., p. 149 (Stanford University Press, Stanford,
California, 1973) .
5. Frye, G.A., Private Communication, Burmah Oil and
Gas Company.
6. Weinberg, C., Private Consnunication, Pacific Gas and
Electric Company.
7. Ruff, R.E., F.L. Ludwlg, L.S. Gasiorek, A. Chaplin,
W.S. Meisel, S. Miller, and H.J. Payne, "An SO2
Emission Limitation Program for the Kingston
Steam Plant," Final Report, SRI Project 4044, to
be released.
Acknowl edgment 8
The authors wish to acknowledge contributions from
R.T.H, Collis and F.L. Ludwig, SRI, and C.P. Edwarda,
Ball Brothers Research Corporation.
5
-------
ESTIMATION OF POINT SOURCE
EMISSION STRENGTHS WITH AIRCRAFT
J. E. Cunningham, W. F. Harris, J. W. Key, C. D. Wolbach
Texas Air Control Board
8520 Shoal Creek Boulevard
Austin, Texas 78758
It is shown that emission source strengths can be determined
by aerial sampling to good precision. Furthermore, the procedure
is more cost effective and more versatile than standard sampling
procedures. The procedure will provide a data basis upon which
organizations involved in sampling air pollutants can more
efficiently allocate limited resources to achieve their particular
goal.
Introduction
A knowledge of point source emissions is needed for a wide
variety of reasons by both regulatory agencies and industry. The
accuracy of measurement needed for these emissions varies from
exact measurement of mass emission rate for determination of
compliance with statutes to an estimation of emissions for use in
approximating the pollutant burden in an area. The most accurate
knowledge of source emissions is obtained by stack sampling
techniques, e.g. EPA Method 5. Although this method can give
results with a relatively high degree of accuracy, it is quite slow,
labor intensive, expensive in both capital investment and personnel
support for the sampling team and its equipment as well as the
sampling facilities that must be supplied by the source, and the
accuracy and precision of the data are dependent on the degree of
training and experience of the sampling team.
To overcome some of these problems, the collection of air
pollution samples from aircraft is becoming widely used and is
gaining rapidly in acceptance, diversity, accuracy, and application.
The activities encompassed by aerial sampling range from simple
spotting of polluting activities for further investigation from the
ground to very sophisticated detection and tracking by very high
altitude surveillance aircraft or by satellites. Most of this effort
is directed towards collection of data on ambient pollutant dis-
persion patterns, or tracking of a particular pollutant to determine
its ultimate fate. Some use has been made of aerial sampling to
verify and/or update and modify, dispersion models. The staff
of the Texas Aii Control Board has conducted an evaluation of
aerial sampling as a means of determining point source emissions.
This study has included determination of instrumentation adequate
for the pollutants to be measured, development of appropriate
methodology for flying the aircraft, recognition and measurement
of other parameters necessary for the calculation of emission
rates, and development of mathematical equations for reducing
the data in order to calculate emission rates. This paper describes
the general approach adopted and developed for SO2, and some
representative data and correlations. Economics of aerial sampling
are mentioned. Further evaluation of methods including other
pollutants is continuing, and the results will be reported at a
later date.
Hie Use of an Aircraft
For Source Statu* Determinations
There are many reasons that the ability of an aircraft to
proceed directly to a pollution source and collect air pollution
samples proves advantageous to a source monitoring and
sampling system based on ground sampling. Some of these reasons
are quite obvious, and some are less readily discernible. Among
these reasons aie: the ability of the aircraft to move quickly to
separate and remote areas and its freedom from the restrictive
limitations of road accessibility and observance of property lines.
Also, since the aircraft is able to move from place to place
quickly, and testing requires no time-consuming preparation or
setup at the site, this sampling can be accomplished very rapidly.
This rapid screening, then, could allow the examination of fifteen
to twenty sources per week instead of the one-to-three sources
per week with current ground-based methods. The savings in
time, money, and manpower realized in conducting an equivalent
amount of sampling by aircraft, as opposed to ground-based
methods is a considerable one, since aerial sampling would
result in the sampling of many more sources per week for an
equivalent capital and personnel commitment. This ability of the
aircraft, then, to sample a source quickly no matter how isolated
it is and to make a rapid determination as to whether or not the
emission levels were of sufficient concern to warrant further
study is a very important reason for utilizing an aircraft for
source sampling. If further data is needed or if existing regula-
tions do not permit complete acceptance of the data acquired
by aerial sampling, this screening will allow the more conven-
tional sampling methods to be applied to those sources which
are clearly in need of further study. From the standpoint of a
control agency enforcing regulations, the total compliance level
would also be much greater because the surprise value of the
aircraft which can come in, take a sample, and move on, no matter
how isolated the location, would make polluting activities of a
clandestine nature much less feasible. Everyone would know
that they were subject to being observed, measured, and, if
necessary, taken to court at any time with no prior warning.
This may not be necessary in most instances, but it will certainly
be a great deterrent to those few who feel that pollution laws
can easily be circumvented. Some of the other reasons for using
the aircraft for source sampling involve the sample quality
factors associated with freedom of movement, i.e. the collection
of samples from the plume downwind of the source with no
concern about the accessibility of the property downwind of the
source perimeter. This means that rough fields, lack of roads,
swamps, or forests pose no problems to the collection of the
sample. Neither is any problem caused by inability to contact
property owners or by their reluctance to allow entry to their
property. This freedom arises not only from the aircraft's
ability to move freely over these terrestrial restraints, but also
from the fact that the aircraft is operating in public domain.
This last factor should help to reinforce the legal standing of
the samples collected since it should not be necessary to show
that someone lives in the area and could be harmed by the
pollution, because the pollution presents its hazard to those
legitimately flying through the area. The freedom of movement
without regard to perimeter accessibility relieves the samples of
deficiencies caused by an inability to site the sampling in the
plume, or a shift in plume position during the sampling period.
This applies equally whether the siting or plume shifting
problems are in the horizontal or in the vertical plane.
Several factors of importance to regulatory agencies are
associated with the aircraft's speed and mobility. One is that
sample location and time selection can be scheduled for the
benefit of the agency and not at the convenience of the source
to be sampled. This means not only that the samples can be
collected where and when they are needed, but that no advance
notice is needed, so samples are truly representative of normal
operating conditions and not biased by special attention to
operating parameters or deliberate alteration of operations
for the sampling period. The actual alteration of operations
during sampling may indeed be a rare occurrence, but the
elimination of this possibility would greatly improve confidence
in the data collected.
Theory
An approximation of the emission rate of a pollutant
emitted by a point source may be made through the use of two
assumptions: that the distribution of pollutants in a plume is
approximately gaussian, and that the instantaneous sum of the
pollutant being swept through a cross-section of a plume is
1
18-6
-------
equivalent to the instantaneous emission from the source. With
the concentration of pollutant at the. centerline of the plume
measured (C„ax) the concentration for any point x in the plume
at a distance rrom the centerline can therefore be calculated by
c = CmaxeXp("2'3026 r2/r2max> (1)
Since the distribution has been assumed to be gaussian, those
points of equal concentration will exist in discs perpendicular to
the axis of the plume. The volume of air being swept through
such a disc is expressed by
dV = 2*vrdr (2)
where v is the wind velocity. Combining the air volume and
concentration calculations for these discs and integrating across
the plume gives an emission equation
E = 1.23 kv Cmax r^max (3)
where r is the horizontal half width and k is a constant to
correct for an eliptical cross-section. Further corrections are
necessary for non-standard conditions and to express the results
in consistent units.
The parameters needed for this calculation are pollutant
concentration; plume horizontal and vertical distance measure-
ments; wind speed; and for gaseous pollutants, molecular weight,
ambient temperature, and ambient pressure. In practice, the plume
width is determined as the product of chart speed and the linear
measurement of peak width for a recorded peak chosen as being
the centerline of the plume.
The plume is arbitrarily defined as', that portion of chart
recording where deflection from the baseline is greater than 10%
of the total chart deflection. This biases any errors in favor of
the source. Any other errors induced from the theory will also
tend to be in favor of the source. This 10% value, then, is used
as the point where the width measurement is made on the chart
peaks to determine the width of the peak. This may be done by
taking a graphical 10% of the peak height, but due to nonlinear
response in the instruments this value is in fact determined by a
mathematical computation which determines the exact mathemati-
cal 90% value, or 10% value. This mathematical value is then used
to determine the peak width. This value generally tends to be
higher than the graphical value, so using the graphical value would
bias the calculations against the source, and tend to give higher
results for the emission rate. Systemic errors in this sampling
system will tend to give lower results, i.e. lower calculated
emission rates. For example, not traversing the centerline at the
altitude of the centerline will give a lower maximum concentration
reading, failure to obtain the top and bottom of the plume
measurements correctly will tend to cause the assumption that
the plume is of lesser depth, which will also give lower values.
It has been found that, if looping commences, the values assigned
for the top and bottom of the plume for the depth tend to be
less than the actual value and the calculated emission rates
are much lower than the real values. Another problem which will
cause lower results is injection of the plume into upper layers
where the wind increases. This commonly happens, e.g. at
2000 feet the wind is greater than at 1000 feet and at 3000 feet
the wind is greater than at 2000 feet. This leads to the plume
being dissipated at different rates, and the determination of the
correct wind speed becomes quite difficult under these conditions.
All of these conditions will, of course, give lower values to the
emission rates calculated. Systemic errors induced by the instru-
mentation also tend to give reduced emission rates. Chart speeds
will normally become slower in the event of a malfunction and
this will result in the peaks appearing narrower as though the
plume were not as wide. Any detector or recorder response time
lags will tend to cause the response not to reach its maximum
during the traverse. The area under the peak or the peak height
itself will then be not as great. This is, of course, accentuated
by a higher air speed. Other errors in the system or in the sample
pickup will involve, for instance, an attempt to take the sample
from an air source which has been designed to supply clean air to
the cabin, i.e. it has filtering or too long a residence time in the
system so that it tends to spread out the peak and give a lesser
peak height. This results in a greater part of the sample being
in the 10% of the peak which is rejected as part of the design of
the system. The most severe errors introduced to the system
tend to be those associated with wind speed related problems. The
wind speed being one of the inputs to the formula for calculating
emissions, it is a critical parameter and also does not lend itself
to measurement as easily as the other values do. Wind sp> ed is
determined as an average, and under conditions where the wind
speed is not consistent (such as very light and variable winds or
strong gusty winds) average wind speed is not truly representative
of the instantaneous wind speed which has acted upon the part of
the layer which is being sampled. Light and variable conditions
normally result in quite low readings because the part of the
plume being sampled is not a completely representative part of
the plume. Part of the instantaneous emission at the time the
sample was emitted will be carried in a different direction and will
not be in the line of traverse of the plume or the assumed plume.
Under strong gusty conditions it becomes almost impossible to
calculate the emission rates because the pollutant is concentrated
in such a way that one pass at an altitude will reveal a very high
concentration and the pass immediately thereafter, at the same
altitude will reveal a very low concentration at that same altitude.
This is caused by the pollutant being acted upon by winds of
different intensity and tends to aggregate the pollutant in waves.
It also makes it very difficult to determine the top and bottom
of the plume. This is one condition that may lead to the assump-
tion of some figures which are entirely too high and cause
erroneously high results, so sampling under strong gusty wind
conditions is not completely satisfactory.
Tracking
One important application of aerial sampling to air quality
is the determination of the true source of a specific pollutant
found (to exist) in an area. The staff of the Texas Air Control
Board has developed a 'tracking' method, which allows the
patrolling of a suspected polluted area by an instrumented
aircraft, and when a specific pollutant is detected by the instru-
ments, the plume can be accurately tracked upwind to its source.
This method involves homing on the source by the same
bracketting techniques used to intercept and track the aural radio
ranges which were used for aerial navigation prior to todays VOR's.
Patrolling is done on a crosswind heading at an altitude 500 feet
to 2000 feet AGL but below any inversion layer, and with the
instruments set to the most sensitive practical detection range.
When a pollutant plume is detected by a peak in the trace on a
strip chart recorder, for example, the course is reversed to
determine precisely the boundaries of the plume.
A 135° turn is then made upwind with the instrument
operator notifying the pilot when the plume is entered and
exited as determined by the instrument response.
Each time the aircraft exits the plume the pilot executes an
upwind turn, regularly reducing the angle, typically by 15° , so as
to come more nearly onto an upwind heading. As the source is
approached, the plume is narrowed and the change in the angle of
turn must become greater or the plume will be crossed so
quickly that it becomes quite difficult to track and may be lost.
The application of this technique is useful both in the case of
determining the origin of citizen complaints when that source is
unknown and in locating the major contributors to the pollutant
burden in an area. An example of a technique associated with
tracking and of the versatility of aerial sampling occurred in
March, 1975. While conducting tests at a power plant, a dense
persistent black plume was observed near the sampling site. When
the plume proved to be a continuing one, an investigation was
made. Tracking was not necessary in this particular instance
since the source was isolated, and the plume had a Ringelmann
reading of 2 some twenty miles downwind from the source.
After locating the source, which proved to be open burning
2
18-6
-------
of oil leaking from storage tanks, its position was fixed by deter-
mining the bearing, i.e. the omni radials, from the two nearest
VOR's. Color slide photographs of the source were taken with a
camera equipped with a telephoto lens. Some of these photos
were taken from approximately the cardinal compass directions
at an oblique enough angle to show surrounding roads and terrain
features. Using the data on location provided by the radials from
the two VOR's, at the site of the burning, it was possible to
determine the position of the source on the appropriate Sectional
Aeronautical Chart. Using the latitude and longitude data and the
terrain data from the Sectional Chart, and using the photographs
for recognition of fine detail, it was possible to identify positively
the site of the burning on the appropriate U. S. Geographical
Survey 7.5-minute series topographic map.
With the site identified and the information provided by the
aerial sampling crew, it was possible for the office in charge of
the area to determine that the activity was conducted in violation
of the law. Detection of violations such as this are of obvious
value in themselves, since it allows determination of the responsi-
bility for emissions. A perhaps even greater benefit from such
aerial sampling operations, however, is the dissuasion to the
commission of such acts, provided by the knowledge that positive
detection can be accomplished quickly and with relative ease.
Experimental
Various aircraft have been used as platforms for aerial
sampling by the Texas Air Control Board. The majority of
flights, however, have been in Cessna 182's. All aircraft are
operated on short term leases, thus no modifications to the
aircraft were made. The samples were obtained by running the
shortest possible length of Teflon tubing through an outside vent
in the aircraft. It should be noted, however, that although the
cabin fresh air vents are an obvious source of outside air, they
normally have filters, blind water traps and unnecessarily long
air paths causing excessive residence times, and are generally
not the best choice for collecting a representative sample.
Experience has shown that, although this simple sample pickup
system seems to be acceptable for gaseous pollutants, a better
designed system that more nearly approaches an isokinetic
sample is. necessary foe particulates. A three man crew consisting
of pilot, navigator, and instrument operator is preferred; with
the navigator maintaining a written and verbal log, and coordinat-
ing between pilot and instrument operator. Successful sampling
has been accomplished, however, using an experienced two man
crew provided with an, intercom system and taping al! intercom
activity for later data reduction.
Instruments giving real time data, with as little lag as possible
are desired. Instruments used to date include those for total
sulfur, by flame photometric detection, particulate mass by a
quartz crystal micro balance technique and carbon monoxide
by potentiometric reduction. Instrument data were recorded on
strip chart recorders. Other parameters, e.g. air speed, outside
air temp, altitude, etc., were recorded on the voice record tape.
Equipment was operated on internal supplies, i.e. batteries and
necessary gases; or external supplies including power from the
aircraft.
Since the method is still in the developmental stage, no
attempt has been made to develop a comprehensive quality
assurance program with a determination of stability or drift
rates for the various instruments. Instead, the instruments are
calibrated in the laboratory prior to flights.
Besides the pollutant data other parameters needed include;
air speed, or indicated air speed, air temperature, the altitude of
a specific pass, (the top and bottom centerline of the plume being
determined from the altitude of the traverses taken). Since the
wind speed is an important parameter in the equation for calcula-
tion of emissions, wind speed at the plume centerline altitude is
determined by making runs parallel with the wind downwind and
upwind between geographical fixes, timing these runs, and then
calculating the average wind speed from the air speed and the
elapsed times for the upwind and downwind runs. The pollutant
measurements are made by flying at a given altitude perpendicular
to the plume, normally over a fixed geographic reference point at
a minimal air speed. For aircraft currently in use, 80 miles per
hour indicated air speed has been found to be the optimum air
speed considering the need to have the speed as slow as possible
in order to give the best instrument response and still maintain an
adequate margin of safety for flight operations. The same air
speed normally is maintained for both pollutant measurement
runs and wind speed runs. It is an extremely important part of
the pilot technique that the altitude and wind speed be held
constant during all measurement runs. Since the traverses of the
plume should be perpendicular to the wind the heading for these
plume traverses is normally maintained at a slight angle into the
wind in order that the true course may be perpendicular to the
plume. This also is an important aspect of the pilot's responsibility
in maintaining the correct crab angle to produce this true cross
plume track. A determination of the base and top of the plume has
been arrived at by different methods. One method requires flying
parallel such that a horizontal line projected into the plume will
touch either the top or the bottom of the plume. Another method
tried has been a constant-rate climb or descent through the plume,
noting the times or the altitudes when instrumentation shows
entrance and exit through the plume. The most common determina-
tion of plume base and top, however, has been through perpendicular
traverses of the plume at altitude increments until it has been deter-
mined at which altitudes a minimal indication is observed on the
strip chart recorder.
Normally the same indication for the base and the top is
required to insure that both have been observed. This method is
perhaps more time consuming than other methods, but is felt to
be more accurate and is also easier to use in the event that there
is some slight looping of the plume. The usual pattern for
determining emission rates consists of flying a pattern of passes,
starting above or below the level of the plume, at even 100 feet
intervals until the complete depth of the plume has been traversed
for measurements. Following this series of passes perpendicular
to the plume, a wind speed run downwind and then back upwind
is normally made at the apparent centerline of the plume as
determined from the previous traverses. This run is normally made
over a period of two to three minutes downwind and a corresponding
upwind time in order to obtain a reasonable average of the wind
speed.
Results
Aerial sampling operations for SOo have been conducted at
plants whose average emissions ranged from 400 pounds/hour to
24,000 pounds/hour. The standard deviations have ranged from
two to six percent of the average emissions for those samples.
Such testing, when accomplished while simultaneous stack sampling
was in progress, has produced results which agree with in-stack
measurements.
A summary of the statistical data for the correlation between
aerial sampling and simultaneous stack sampling by an independent
contractor is giver, in Figure \. Commensurate results have been
obtained when aerial sampling is compared to stack sampling not
conducted concurrently.
Economics of Aerial Sampling
The economic considerations of the use of aircraft to
determine point source emissions at first seem quite forbidding,
since the common view exists that an aircraft is expensive and,
in the eyes of many, a frivolous toy. When studied closely,
however, aerial sampling as a method of determining point source
emissions becomes much more feasible. It is, in fact, economically
quite sound, since a study will show that aerial sampling is really
the less expensive system to operate. The factors to be considered
in the economics of aerial sampling versus stack sampling include:
the initial cost of the aircraft, its operational costs, the costs for
the ground transportation for standard sampling means (including
initial capital outlay and their operational costs), the cost of
personnel expenses for employment and training of crews to
conduct the various types of sampling, the equipment costs, the
time required to collect an aerial sample, the time required to
collect a stack sample, the per diem cost of supporting an aerial
sampling or a stack sampling crew in the field, the full time
chemist in the laboratory for analysis of the samples collected by
the crew. This also brings in the added cost of maintenance of a
laboratory necessary for analysis of these samples. The actual cost
to set up the laboratory for analysis will exceed the initial Cost
3 l»-6
-------
differential of the aircraft purchase. It is obvious that the total
cost of supporting a laboratory is not chargeable to a single opera-
tion, but the proportion of the cost of a laboratory such as used by
the Texas Air Control Board for analysis of stack samples far exceeds
the original cost of the aircraft versus the original cost of the stack
sampling crew equipment.
A breakdown of the capital outlay for aerial sampling and
stack sampling is given in Figure 2. TACB aerial sampling crews
have been able to collect data on from one to six sources in a
single day. There is also little or no per diem incurred for main-
taining a crew in the field since most trips can be completed in
less than eight hours. The normal stack sampling trip, however,
involves sending three or more people out for an entire week to
sample one or two sources and ties up the two vehicles needed to
transport the crew and their sampling equipment. A comparison
based on current operating and per diem cost for supporting crews
in the field, and collecting six samples by aircraft with a crew
of two staying one night, as opposed to a stack sample with a
three man crew staying out for the week and collecting two samples
is given in Figure 3. TTiis is based on direct costs and does not take
into consideration the additional administrative costs of employ-
ment.
Assuming fifty sources are sampled per year, and that capital
costs are amortized over three years, gives a net cost per sample
of $516 for aerial sampling and $890 per sample for stack sampling.
If, however, the aircraft and vehicles were sold at market value
after three years the net cost per sample would become $436
for aerial sampling and $878 per sample for stack sampling.
Aerial Sarnies
Aerial Sampling Mean
Standard Deviation
Range Around Mean + ~
Precision (Mean Error) +
Stack Samples
Stack Sampling Mean
Standard Deviation
Precision (Mean Error) +
Difference Stack Sampling-Aerial Sampling
% Difference
Aerial Samples Within + of Stack Sample
Average
Figure 1
SC>2 Emissions
20,138 lbs/hr
99 5 lbs/hr
%%
406 lbs/hr
20,129 lbs/hr
200 lbs/hr
11S lbs/hr
9 lbs/hr
0.04%
12
22,662 lbs/hr
1,355 lbs/hr
9%
391 lbs/hr
22,618 lbs/hr
200 lbs/hr
115 lbs/hr
44 lbs/hr
0.19%
9%
NOTE: Data has been weighted for differences in power loading but has not taken'account of the three hour sampling time
for stack sampling versus 20 minutes time involved with aircraft sampling.
Aerial Sampling
Aircraft
SO? Analyzer
Particulate Analyzer
Recorder
Miscellaneous Accessories
Data Reduction
Figure 2
Capital Costs for Sampling Equipment
Stack Sampling
$16,000
4,500
5,000
750
250
800
$27,300
Bubble Top Van
Crew Van
Stack Sampling Equipment
Miscellaneous Accessories
Laboratory, including glassware,
balance, data handling
$6,500
5,000
8,600
200
5,000
$25,300
Figure 3
Operational Costs for Sampling
Aerial Sampling Stack Sampling
Direct Operating Costs
5.6 hrs flight time
$100.80
464 mi. @ 8 mpg +• 514 mi. @ 6 mpg
$ 84.76
Per Diem Costs
2 meals @ 2.SO
5.00
3 crew members @ $77/4 days
231.00
Crew Salary
2 crew @ $56/day
112.00
2 crew members @ $56/day + 1 crew
member @ $40/day
152.00
Analysis and Data
2 days @ $56/day
112.00
3 days analysis @ $56/day
168.00
Reduction
Preparation
1 day report <® $56
56.00
Gases and Chemicals
5.00
30.00
$334.80 $721.76
4
18-6
-------
DEVELOPMENT OF A REFERENCE METHOD
FOR THE MEASUREMENT OF PLUTONIUM IN SOIL
E. W. Bretthauer and P. B. Hahn
Office of Research and Development
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Las Vegas, Nevada 89114
This paper describes the need for and the consi-
derations involved, the approaches used and the prob-
lems encountered in developing and proposing a refer-
ence method for the measurement of plutonium in soil.
The proposed method is described and its choice over
other methods is justified. Results of a single lab-
oratory evaluation and a preliminary interlaboratory
evaluation of the method are presented.
Introduction
During the early part of 1976 the U.S. Environ-
mental Protection Agency (EPA) intends to develop
guidelines for cleanup, restoration and occupancy of
areas contaminated by plutonium. Although the most
important pathway for plutonium to the human receptor
is via inhalation of particles resulting largely from
resuspension from existing contaminated land surfaces,
the guidelines,when promulgated, will probably be
based, at least in part, on plutonium-in-soil concen-
trations. The reasons for this emphasis are (1) there
is a correlation between air concentrations and soil
concentrations, and (2) soil concentrations are rela-
tively stable as a function of time, while air con-
centrations can fluctuate widely according to existing
meteorological conditions. In order to establish
accurate soil concentrations, a reference sampling and
analytical method for monitoring plutonium in soil will
be necessary.
EPA policy is to select a single reference method
for a single environmental pollutant. A method so
designated must have acceptable accuracy and precision
performance characteristics which have been scientifi-
cally and statistically validated by multiple-
laboratory collaborative tests under a variety of anti-
cipated user conditions. The method must be such that
it can be readily implemented by prospective user lab-
oratories. This precludes designation of very expens-
ive, sophisticated methods as reference methods even
though they may be the most accurate methods available.
Occasionally there is an extreme urgency to prom-
ulgate a standard and the EPA may specify a "tentative
method". A tentative method is chosen as the most
reliable method in the opinion of the EPA but is unvali-
dated by collaborative tests. It is replaced by a
fully-validated reference method as soon as it becomes
available.
A reference method must be published in a concise,
easy-to-follow format. The document must indicate the
principle, applicability, range, sensitivity, inter-
ferences, accuracy and precision of the method. It
must describe the apparatus and reagents employed and
present specific procedures for sampling, analysis,
calibration and calculating results. It must contain
all pertinent references to the literature and special
precautions for handling and disposing of any hazardous
materials.
The general procedure for producing a reference
method is as follows1:
1. The sensitivity, accuracy and precision levels are
defined to meet the requirements of the standard pro-
mulgated by the EPA.
2. Based on past experience, generally applicable
user criteria for the selection of acceDtable methods
are recommended by a panel of experts representing the
ultimate users.
3. An intensive single laboratory evaluation is per-
formed to determine how candidate methods meet the
criteria established by the Agency and the users.
4. Those methods which best meet the criteria are
then collaboratively tested by user laboratories.
5. The single method which best meets Agency and
user criteria is then designated as a reference method.
Criteri a
Since an environmental plutonium standard for
soil has not yet been promulgated, exact criteria for
sensitivity, accuracy and precision for a plutonium-
in-soil reference method were unavailable. Efforts to
develop a method were begun in spite of the lack of
these criteria as there was widespread, well-documented
uncertainty in the scientific community as to whether
any of the available methods could meet certain of the
anticipated Agency and user criteria. If no method
were available, substantial lead time would be neces-
sary to develop such a method. It was anticipated that
the method would have to be applicable to soils con-
taining background to highly-elevated levels of
plutonium,and efforts to evaluate current methodology
were begun on that assumption.
The candidate method would have to be reliable,
exhibiting reasonable accuracy and precision, and
applicable to the wide variety of soil types found 1n
the U.S. The method would also have to be capable of
accurate analysis of all chemical and physical forms
of plutonium known to exist 1n the environment. A
necessary reguirement was that if the method should
not be applicable to a given sample type it would be
unequivocally apparent either during the course of
the analysis or from the results obtained.,
Choice of Method
The first task was to oerform a critical review
of current!- a"dilable methods, two approaches were
employed. T
-------
Commission (AEC) laboratories, two EPA laboratories,
and three private laboratories. These laboratories
perform at least 80% of all plutonium-in-soi 1 analyses
conducted in the United States. Scientists from other
organizations having expertise in the area also at-
tended, as well as selected personnel representing
elements of agencies administratively responsible for
Plutonium analysis of soil, i.e., two senior scientists
from EPA's Office of Radiation Programs, EPA quality
assurance officals, and officals from certain AEC field-
operations offices. There was near-unanimous agreement
by the attendees on certain criteria. Those criteria
were as follows:
1. A 10-gram sample size would be adequate for the
sensitivity required.
2. An extremely important consideration in the
analysis of soil for plutonium is the decomposition
of the sample and the extraction of plutonium from
the soil matrix for subsequent purification and
counting. Certain forms of plutonium, especially
those which have been fired at high temperatures, are
known to be extremely refractory and consequently are
difficult to extract from a soil matrix. The method
must be applicable to such forms of plutonium.
3. Available methodology for the analysis of pluto-
nium in soils can be classified into three major groups
according to sample decomposition techniques: acid
leaching, total dissolution with nitric and hydro-
fluoric acids, and total decomposition by sequential
potassium fluoride and potassium pyrosulfate fusions.
Acid leaching techniques were considered unacceptable
because of the large quantities of residue which gen-
erally remain after such treatment. Total dissolution
of soils with nitric and hydrofluoric acids is all too
often not truly total and small quantities of residue
or turbidity remain after 10 grams of soil have been
treated. The fusion technique, however, is known to
dissolve the most refractory and intractable pluto-
nium compounds while at the same time completely dis-
solving the soil matrix. This allows for the complete
exchange between the plutonium in the sample and the
plutonium-236 tracer used to trace the plutonium
recovery through the analysis.
Based on these criteria it was decided that the
fusion iifithod had the widest applicability. A method
developed by the Atomic Energy Commission Health and
Safety Laboratory, Idaho Falls, ID, which is extremely
2
well documented in the literature was chosen for an
intensive single laboratory evaluation. In this method
a known quantity of plutonium-236 tracer is added to
10 grams of soil which is decomposed completely by
sequential potassium fluoride and potassium pyrosulfate
fusions. The fused sample is dissolved in dilute
hydrochloric acid and the plutonium is isolated and
~Effective June 29, 1975, the National Environ-
mental Research Center-Las Vegas (NERC-LV) was
designated the Environmental Monitoring and
Support Laboratory-Las Vegas (EMSL-LV). This
Laboratory is one of three Environmental Moni-
toring and Support Laboratories of the Office
of Monitoring and Technical Support in the U.S.
Environmental Protection Agency's Office of
Research and Development.
i
Since January 19, 1975, the Energy Research
and Development Administration.
separated from uranium by co-precipitation with barium
sulfate. The barium sulfate is dissolved in perchloric
acid, the solution is adjusted to approximately 2M in
aluminum nitrate and the plutonium is reduced'to the
tetravalent state with sodium nitrite and extracted
into Aliquat-336 (methyl tricaprylyl ammonium chloride).
The trivalent actinides are separated by scrubbing the
organic extract with nitric acid and thorium is sepa-
rated by scrubbing with hydrochloric acid. Plutonium
is then stripped with a perchloric-oxalic acid solution
and electrodeposited onto stainless steel disks.
Plutonium-2 36, piutonium-238 and piutonium-239 plus
plutonium-240 are determined by alpha spectrometry.
3
The proposed reference method employs these sample
decomposition and plutonium separation techniques
coupled with an electrodeposition from an ammonium
4
sulfate solution .
Single Laboratory Evaluation
The proposed method was extensively tested by
analyzinq soil samples containinq various known levels
of plutonium. The soil samples were prepared by
mixing a calibrated amount of plutonium-239 with
approximately 200 grains of wet -200 mesh soil, drying
under a heat lamp and then muffling at approximately
700°C for several hours to convert the plutonium to a
refractory oxide. The spiked soil was pulverized to a
-200 mesh and weighed to determine the final plutonium
concentration. Losses were compensated for by col-
lecting and analyzinq all of the residues for plu-
tonium. The spiked soil was then blended with un-
spiked soil to prepare soils containing various known
5
levels of plutonium ,
Several laboratories which expressed interest in
the collaborative testing of the method agreed to ana-
lyze a practice sample by the proposed method and pro-
vide results and comments on both the method and the
document describing the method. Table I presents
the results of the plutonium-239 analyses performed
at this laboratory on several soils containinq known
auantities of plutonium. Table II presents the results
from outside laboratories attempting the analysis for
the first time on a soil sample containing 5.77 +
0.02 dpm/q plutonium-239 and 0,089 ± 0.001 dpm/a
pi utoni um-238.
The results obtained at this laboratory demon-
strate the ability of the tentative method to achieve
a single-laboratory accuracy and precision approaching
that of counting statistics alone when samples are
analyzed which contain 1-10 dpm/g of plutonium-239.
Plutonium-236 recoveries were generally above 80% pro-
viding nearly optimum sensitivity and precision. Com-
plete decomposition of the soil matrix was achieved
during the fusions for every sample analyzed. This
included several other types of soil which were also
analyzed by the method. Similar analytical results
have been documented in the literature for the analysis
of soils at the 30- and 0.5-dpm/g levels using a method
similar in all respects except for the electro-
deposition'. Another study also provides detailed
information on decontamination factors from possible
actinide interferences, i.e., greater than 300 for
thorium and greater than 5,000 for the trivalent
2
actinides .
The preliminary results from the practice
analyses performed by the outside laboratories were
also encouraging. The excellent comments provided by
these laboratories allowed for document revision in
terms of concise rewording of several critical steps
which should improve future performance. Although
2
19-1
-------
low and inconsistent yields were obtained by two of
the laboratories, this was understandable as a cer-
tain amount of experience by the analyst is necessary
before optimum results can be expected. A similar
situation existed at our laboratory when we first
attempted the method. The internal precision exhi-
bited by each of the laboratories was excellent,
especially in view of the fact they were provided
with only enough sample to perform triplicate analyses.
Differences among three of the laboratories and this
laboratory appeared minor, being on the order of 5%.
Such could possibly be reduced to even a lower level
by an interlaboratory calibration pf the plutonium-
236 tracer. Some outside laboratories questioned
whether the method would be applicable to soils con-
taining large quantities of calcium. We experienced
no problem in this regard. A simple modification
allowed for the analysis of a soil containing 80%
CaCOg with plutonium recoveries on the order of 60%.
TABLE I.
Results of single laboratory evaluation.
TABLE II.
Results of interlaboratory evaluation.
Analytical Result
Known Value
0.99 ± 0.01
Approximate
239Pu Level
(dpm/g)
1.01
1.01
0.99
0.99
0.01
0.01
0.01
0.02
0.97 ± 0.01
0.97
0.93
0.98
0.98
0.97
0.96
1.03
0.96
0.92
0.03
0.03
0.02
0.01
0.02
0.01
0.03
0.04
0.04
Plutonium
Recovery
(%)
88
61b
83
88
81
87
79
93
88
92
92
91
68
88
91
92C
95°
93
89
57d
64d
1.01 ± 0.02 <2
1.03 ± 0.03
0.99 ± 0.03
0.98 ± 0.03
1.08 ± 0.04 5
0.94 ± 0.04
a 1 a counting uncertainty
b Partial spi'11 of sample
c Soil contained 2-3% CaCO^
d Soil contained 80% CaC03
There is one very important potential weakness
in the method: protactinium-231 quantitatively fol-
lows plutonium through the analysis and it can, at
times, be difficult to resolve the protactinium-231
peaks from plutonium-239 and plutonium-240 peaks by
alpha spectroscopy. The effect Is most acute when
analyzing for lower levels of plutonium (<0.2 dpm/g)
and apparently negligible at the higher levels (>1
dpm/g). The collaborative testing of the method should
provide the necessary information to decide whether an
additional step to separate protactinium 1s necessary.
Results
Plutonium
2 3<3pu
2 38pu
Recovery
Laboratory
(dom/q)
(dpm/g)
(%)
A
5.81
0.072
54
5.65
0.082
52
B
5.45a
0.090
2
5.50a
0.070
33
5.34a
0.074
53
C
5.44
<0.6
22
5.51
<0.4
52
5.53
<0.4
37
D
4.92b
0.08
69
5.02b,C
0.08
65
Expected
5.77
0.089
a Used own electrodeposition technique in
sulfate media
b Used plutonium-242 tracer and own electro-
deposition technique in chloride media
c Used own adaptation of the fusion method
The method is presently undergoing a collaborative
test. To date eleven laboratories are participating
by analyzing, in triplicate, four soils containing
known rlutonium-239 and plutonium-238 levels ranging
from ..01 to 10 dpm/g. Results of the test should be
available by January 1976.
Conclusions
The authors are confident that this method is the
best available for the Duroose intended; a reference
method to evaluate other methods for equivalency by
standardizing reference materials necessary to perform
such a task. The method has disadvantages and may
have to be modified in the future in view of the
protactinium-231 problem. It certainly cannot be used
by small laboratories having no perchloric hood. It
does however, have the advantages which we feel most
necessary for a reference method. It employs the best
technique known for completely extracting the sought
for substance from the sample matrix, and it contains
built-in mechanisms to critically evaluate the results
obtained by using the method, i.e..visual observation
of whether the sample has been completely decomposed,
extent of recovery and resolution of the tracer peak
in the alpha spectrum. In terms of cost, the fusion
method for the analysis of plutonium 1n soils does
not appear significantly more expensive than total
dissolution methods. The cost differential atone
large commercial laboratory is only 10%.
In retrospect, perhaps more effort could have been
spent in the single-laboratory evaluation, especially
in terms of more intensive evaluations of other methods
or Improvements to this proposed method. Even now,
preliminary Investigations at this laboratory have
implied that a combination of a potassium fluoride
fusion and an anion exchange purification may be feasi-
ble for the analysis of plutonium 1n glass-fiber air
filters and perhaps soils. Such a method would be
much more available to many smaller laboratories. But
time was also a controlling factor due to the urgent
critical need for a validated method.
3
19-1
-------
References
1. U.S. Environmental Protection Agency, "Measure-
ment methods standardization strategy document,"
Office of Monitoring Systems, Quality Assurance
Division, Environmental Protection Agency,
Washington, DC 20460. (September 1973)
2. Sill, C.W., K.W. Puphal, and F.D. Hindman,
"Simultaneous determination of alpha-emitting
nuclides of radium through californium in soil,"
Anal. Chem., Volume 46, No. 12, pp. 1725-1737.
(October 1974)
3. Bretthauer, E.W., P.B. Hahn, P.R. Altringer,
A.J. Cummings and N.F. Mathews, eds., "Tentative
method for the analysis of piutorn*urn-239 and
plutonium-238 in soil," Office of Research and
Development, Environmental Monitoring and Support
Laboratory, Environmental Protection Agency,
Las Vegas, NV 89114. (In Press)
4. Talvitie, N.A., "Electrodeposition of actinides
for alpha spectrometric determination," Anal. Chem.,
Volume 44, No. 2, pp. 280-283. (February 19727^
5. Sill, C.W., and F.D. Hindman, "Preparation and
testing of standard soils containing known quantities
of radionuclides," Anal. Chem., Volume 46,
pp. 113-118. (January 1974)
4
19-1
-------
PROBLEMS IN ENVIRONMENTAL SURVEILLANCE
AROUND NUCLEAR FACILITIES
B. Robinson, W. H. Westendorf, atvd C. A. Phillips
Monsanto Research Corporation
Mound Laboratory*
Miamisburg, Ohio 45342
INTRODUCTION
A vast technology has been built up over the
past 35 yr in the field of nuclear, chemical,
and mechanical engineering to cope with the
problems of containment of large quantities
of radioactive material in process and to
limit the quantity of process effluents to
precise small concentrations. This tech-
nology involves more than in-plant contain-
ment of the radioactive material in use; it
also includes monitoring to ensure that
effluents have indeed been adequately con-
trolled. Our environmental monitoring pro-
grams, however, are beset with problems.
For example, intercomparison of data between
facilities is virtually impossible because
techniques, frequency, and relative locations
for collection of air, water, biological,
sediment, and soil samples vary widely from
one facility to another. Detection capabil-
ities between facilities are also very differ-
ent. The biggest problem is that most facil-
ities simply do not do representative
sampling. A review of current environmental
surveillance practices at 19 major facilities
by D. H. Denham and D. A. Waite [l] shows
that the rationale for air sampler placement
at 38% of the sites surveyed is available
electrical power and not the physics of
particulate collection.
Often it seems the belief is that the collec-
tion of larger and larger samples will
produce a more representative sample and in
this way overcome other sampling problems.
But this is not the answer, and problems in
intercomparison of data still exist.
Obtaining accurate, reliable, environmental
surveillance data from sampling and analysis
is at best very difficult, primarily because,
for the most part, representative samples
are not collected in the field. For years,
greater emphasis has been placed on analyt-
ical techniques than on sampling techniques.
Analysis of samples is important; however,
many of the analytical methods are well
established, and are reliable as long as
correct aliquoting and sampling are done.
Too often, nonrepresentative samples are
brought into the laboratory for rigorous,
tedious analysis by time-consuming methods.
A review of current guides on sampling and
analysis reveals that most of the guides are
general and give practically no guidance at
all; those which are specific are limited to
one radionuclide or one medium. It would be
refreshing to see publications on methodology
for sampling much as one sees the publication
of a new method for analysis of various ele-
ments, compounds, and isotopes.
PROGRAM PLANNING
The first priority for any facility in which
there are radioactive contaminants is to
establish an environmental surveillance pro-
gram plan. This planning stage often reveals
an imbalance in emphasis or even inadequacies
in the program.
Sampling and Analysis are the heart of any
environmental program. The location, frequency,
and the media to be sampled should be planned
on a calendar year basis for a smooth program.
A miminum of 5% of blanks and duplicates
should be analyzed for quality control pur-
poses. Additionally, on an annual basis,
selected samples of each type of media ana-
lyzed should be split and sent to an independ-
ent laboratory for a comparison. Many nation-
ally and internationally sponsored laboratory
intercomparisons can be participated in, such
as those sponsored by the Environmental Pro-
tection Agency, the American Industrial
Hygiene Association, and the International
Atomic Energy Agency.
The results obtained from sampling and analysis
and diffusion modeling are used for dose
estimation.
The Facilities Section of the environmental
surveillance program should see that control
*Mound Laboratory is operated by Monsanto Research Corporation for the U. S. Energy Research
and Development Administration under Contract No. E-33-1-GEN-53.
1 19-2
-------
facilities, buildings, processes, and sampling
and analytical devices are designed for maxi-
mum performance. Older facilities should have
a plan for upgrading with phases and target
dates written into the program. Release goals
more stringent than the standards should be
set and possibly lowered each year.
The Research Section not only trains environ-
mental surveillance personnel, but also trains
personnel handling the pollutant, emphasizing
the importance of minimizing releases.
The last section suggested for an overall
environmental surveillance program is Communi-
cation and, because the idea of having a
Communication Section as part of an environ-
mental surveillance program is rather new, it
will be discussed in more detail.
One should document a plan at the beginning
of the year as much as one plans the frequency
for sampling and analysis. This plan would
constitute the routine program and lay the
groundwork for the special radiological inci-
dences, accidental releases, etc. Your plan
should be directed to your three audiences:
your employes, the general public, and govern-
ment officials.
Employes can be reached through the plant news-
paper, orientation lectures, bulletin boards,
etc.
An important channel to the general public is
the news media. Contact and inform the public
through news media open house; news releases;
environmental brochures; city trade shows; a
speakers' bureau; and environmental displays
in banks, malls, etc,
Government can be subdivided into technical -
local, state, and federal health agencies -
and nontechnical - local, state, and federal
officials. A good program routinely communi-
cates with both groups through such devices as:
a technical monitoring report, a nontechnical
monitoring summary, open house for health and
other government officials, and finally, a
well-planned contact program.
The purpose of a contact program is to ident-
ify in a systematic manner off-site contacts
to be made as a part of the effort to communi-
cate proper information pertaining to programs
for environmental control, emergency planning,
and assistance to governmental agencies in the
areas of environmental control.
With a minimum effort, one can achieve tremen-
dous benefits; for instance, acceptance as a
responsible citizen, confidence in your envi-
ronmental surveillance and safety program, and
a reputation for responding promptly and
honestly in the event of an accident that in-
volves a possible public hazard.
ENVIRONMENTAL SURVEILLANCE GUIDELINES
Some typical guides which are presently avail-
able on environmental surveillance are briefly
discussed below. The first, "A Manual for
Environmental Radiological Surveillance at
ERDA Installations" by Corley, et al. , [2] is
still in draft form, but it is the most exten-
sive ERDA (AEC) publication written on radio-
logical surveillance. The 150-page manual
covers topics on purpose, standards and
criteria, surveillance, program design, sampl-
ing and measurement techniques. quality
assurance, data analysis and statistical treat-
ment, dose estimation (individual and public),
and reporting. The only deficiency in this
guide is that specific sampling methodology is
often left to judgement, and although individual
judgement is always needed, specific methodo-
logy can be given for a specific purpose in
most cases.
The first really comprehensive guide was the
25-page document, "Environmental Radioactivity
Surveillance Guide", [3] issued by the Office
of Radiation Programs of U.S. EPA in 1972.
This guide recommends methods for conducting
a minimum level of environmental surveillance
outside the plant site boundary of nuclear
power facilities. It covers topics on environ-
mental surveillance protocol, sampling and
analysis, and dose estimations. It is a good
guide for sample collection, frequency for
sampling, and analysis; it references many
analytical methods. Again, this guide could
give more details on methodology for sampling.
The ICRP Publication #7 on "Principles of
Environmental Monitoring Related to the Handl-
ing of Radioactive Materials" [4] is a general
guide and includes objectives and factors
affecting the design of routine and emergency
surveys - emergency surveys are not covered in
the two previously discussed manuals.
Regulatory Guide 4.x., "Measurements of Radio-
nuclides in the Environment - Sampling and
Analysis of Plutonium in Soil", [57 is an
excellent guide because it gives detailed
procedures for sampling, preparation, aliquot-
ing, and analysis of soil. However, the guide
is specific to plutonium in soil only, although
parts could be applicable to other radio-
nuclides .
A proposed American National Standard, ANSI
N13.10, [6] "Specification and Performance of
On-Site Instrumentation for Continuously
Monitoring Radioactivity in Effluents", applies
to continuous installed instrumentation mon-
itors that measure normal releases, detect in-
advertent releases, and show general trends.
Recommendations are given for the selection of
instrumentation specific to the continuous
monitoring and qualification of radioactivity
released to the environment. This standard
2
19-2
-------
specifies detection capabilities, physical and
operating limits, reliability, and calibration
requirements and sets forth minimum perform-
ance requirements for effluent monitoring
instrumentation.
Regulatory Guide 1.21, "Measuring, Evaluating,
and Reporting Radioactivity in Solid Wastes
and Releases of Radioactive Materials in
Liquid and Gaseous Effluents from Light Water-
Cooled Nuclear Power Plants", [7J describes
programs acceptable to the regulatory staff
for measuring, reporting, and evaluating re-
leases of radioactive materials in liquid and
gaseous effluents and guidelines for classi-
fying and reporting the categories and curie
content of solid wastes. This is a general
guide with emphasis on monitoring of effluents.
"Procedure Manual", [8] published by the ERDA
Health and Safety Laboratory, is probably the
best overall manual because it includes many
analytical methods in addition to giving de-
tailed sampling techniques for air and soil.
Other guides include:
1. "Handbook on Environmental Radiation
Measurements" [9] which is being prepared
by Subcommittee 35 of NCRP,
2. IAEA Publication 16. [10]
3. ICRP Publication 10. [11]
4. ERDA Manual Chapter 0829, "Quality
Assurance". [12]
In conclusion, the various published guides
are in two categories, those which expound
general philosophy and those which are very
detailed but cover only one radionuclide or
medium.
SAMPLING AND ANALYSIS
Many things can be done simply and cheaply to
obtain representative samples and thus make
data between facilities comparable. The
problems include samples not taken after com-
plete mixing, "grab" samples being collected,
sample aliquots not being large enough to
obtain enough radioactivity to count, respir-
able fractions not determined for air sampl-
ing, etc.
Any environmental surveillance program will
involve sampling and analysis of air, water,
soil, sediments, and biological specimens.
Air Sampling First, let us consider air.
Since air is a primary exposure pathway to
man from radionuclides released to the atmos-
phere, environmental air is sampled to eval-
uate potential exposures from inhaled radio-
nuclides. Some airborne radionuclides,
notably tritium, may enter the body by tran-
spiration through the skin. It is of concern
that nuclear facilities release both the gas
and oxide form of tritium, but monitor almost
exclusively for the oxide form even though a
cheap, sensitive method for gas sampling has
been reported by Sheehan, [13 J
Off-site air sampling station placement should
be based on at least a general diffusion
model. Diffusion model placement requires
somewhat specific placement of air-sampling
stations. This is a problem area in itself
because the setup of an air sampling network
can require negotiating each plot of land and
installation of electricity. This can be ex-
tremely costly and time consuming, especially
in highly commercial or residential areas. A
rather simple solution to this problem can be
to integrate the air sampling program with
that of local health agencies. This integra-
tion will involve a contract with the agency
for setup and maintenance of air sampling
stations. This agreement has the following
advantages: 1) health officials have no
problem in placing stations near or at schools,
parks, and public buildings; 2) there is shar-
ing of data and samples because the agency's
needs generally require nondestructive testing
(weighing of total airborne particulates col-
lected) , and thus the filter paper and col-
lected particulates can be returned to the
laboratory for radiochemical analysis; and 3)
there is establishment of rapport and credi-
bility.
What are the problems in comparisons of air
sampling data? First of all, grab samples are
still collected at some facilities, and the
data from these are obviously biased high
because samples are collected during operat-
ing hours and thus are not comparable to
continuous sampling data. Grab-sample and
low-volume continuous sampling data are not
comparable to continuous high-volume air
sampling because of the large difference in
sensitivity. Height of air sampler placement
is critical; samplers placed too close to the
ground will include the radioactivity of
larger particulates in the nonrespirable
range, and air samplers placed too high will
not include the respirable material which is
resuspended and available for inhalation.
Even though this may not be realistic, air
samples should perhaps also correlate with
the breathing rate of a standard man; that
is, sampling rate during the 8-hr work day
should be twice that of the remaining 16 hr.
Respirable fractions are not determined.
Recently, EPA, working with several investi-
gators, perfected a dichotomous collection
device, based on virtual impaction principles,
which separates and collects particles in the
respirable and nonrespirable ranges. [141
After the problems of getting a representative
sample are solved, then collection of a large
sample is In order.
19-2
-------
We should be able to assess inhalation expo-
sure by sampling large volumes of air to
determine impacts above background. In a
good air sampling program for transuranics,
for example, each air sampler will sample
at 1.1 m-Vmin, and approximately 45,000 m^
of air is composited monthly for analysis.
In perspective, this is the air breathed by
approximately 85 "standard" men.
Water Sampling Because most liquid dis-
charges are to streams, rivers, or lakes,
samples of these should be collected. There
is little probability that ground water will
accumulate radioactivity (except for tritium)
from liquid effluent discharges to surface
bodies of water.
To accurately define the potential for
problems, one must first define a water
sample and to do that, these questions must
be answered: Are suspended solids a part of
the water sample? If not, what pore size
filter paper should be used to define a water
sample? Is the radionuclide in equilibrium
with the water and sediment phases?
This last question is asked because equilib-
rium studies by Rogers [15] on the distribu-
tion of plutonium between sediment and water
phases in natural water indicate a plutonium
ratio of 1/100,000 between the water and
sediment phases at equilibrium conditions.
Therefore, one should know the concentration
of the radionuclide of interest in both the
water and sediment because of the large ex-
pected difference in the behavior of ionic
plutonium species and plutonium sorbed onto
sediment.
What are the problems in comparing water
sampling data? First of all, some facilities
are collecting grab samples; these types of
samples have at least two basic problems: 1)
they are not representative and 2) they are
not proportional. Water samplers should be
located far enough downstream from the source
that good mixing has taken place.
This brings us to one of the trade-offs which
must be considered - especially for water -
since the sensitivity of the analysis is a
function of the size of sample collected; one
can design the program to demonstrate compli-
ance to the standard by analyzing a large
number of small volume samples or one can
design the program to analyze a few very large
volume samples with a sensitivity far below
the standard and down to baseline or back-
ground levels.
Any stream monitoring program, especially
effluent stream monitoring, with the purpose
of determining total effluent, must include
radioactivity carried off in suspended sedi-
ment and bedload.
Biological Sampling When biological data are
collected to determine exposure via ingestions,
there is always the question of whether or
not the radioactivity attached to the surface
should be included. Does the location in
which the sample is collected represent that
location and should the whole or part be
analyzed, especially for organs? How large
should the sample be? Which vegetables should
be analyzed?
Soil Sampling Now we come to soil. Many
reactor and plutonium handling facilities are
now routinely sampling soil around their
facilities. Soil sampling is only a supple-
ment to routine day-to-day air sampling which
gives much more information with respect to
the time when the contamination occurred.
The exact area chosen is one where wind and
water erosion have not been excessive nor
where sediment accumulation may have occurred
which would produce low and high values in-
stead of average values of deposition. This
preferred area would generally be a large
flat open area with vegetation. A separate
analysis of a contaminant whose concentration
is well-known can be used to determine whether
the sample is representative. Other problems
associated with the comparison of soil data
are the following. Although representative
samples are generally collected, they may not
be directly comparable in concentration per
gram because samples are collected at different
depths at various sites. A 30-cm core is
generally deep enough for collection of samples
for total inventory; however, profile studies
should be performed to verify this. Surface
samples, less than 5 cm in depth, are difficult
to collect quantitatively, because plugs which
do not go below the grass roots mass are
difficult to quantitatively remove.
After the sample is dried and mixed, the
particles are gently ground. The size soil
to include in the aliquot is not standardized.
Some locations analyze only those fractions
less than 100 mesh or 35 mesh, whereas others
grind all parts of the sample including the
rocks to <100 mesh to include in the analysis.
After the soil sample is prepared for aliquot-
ing, it should be riffled. It should be noted
that as the aliquot size decreases, the
standard deviation increases.
Additional studies indicate that the particle
size plays a large role in the precision of
the data obtained from the various samples.
For example, when the data from 10-g soil
samples are compared to data from 100-g
samples, although the total recoverable
plutonium-238 content is within 80-90% in
each case, precision of the latter is two to
ten times better than the former; suffice it
to say that many additional 10-g samples have
to be analyzed in order that the confidence
level of the analyses becomes acceptable.
4
19-2
-------
Such large samples are not as necessary for
plutonium-239 as they are for plutonium-238
(for example, a counting level of 0.5 dis/mln/
cm^, assuming particles of 0.09m in diameter,
will represent approximately 3,000 particles
of plutonium-239 and only 10 particles of
plutonium-238).
In the case of very discrete particles, ali-
quoting them becomes a problem in that the
number of discrete particles sampled may not
be statistically enough to obtain smooth data
after several replicates are analyzed.
Sediment Sampling The sampling and analysis
of silt is also a useful way to determine a
possible buildup of contaminants (from sedi-
mentation) released via an aqueous discharge
to a waterway.
Sediment samples are collected at, above, and
below the facility discharge area. This
downstream sample should be taken in that
part of the stream where the flow rate is the
greatest. Samples should also be taken down-
stream in an area which favors sedimentation,
such as the inner bank of a bend, at dams,
and at stream widening locations.
A sediment sampling program should be designed
to assess four areas: 1) accumulated levels
in surrounding ponds and lakes that are not
free-running, 2) distribution in the receiving
waterway, 3) distribution on the banks of the
waterway, and 4) "very surface" concentration
available to man.
The most precise, quantitative method of
sampling to assess these areas is to collect
core samples the total depth of the accumula-
tion. It is sometimes necessary to drive the
sampling tube beyond the depth of the sediment
accumulation to "plug" the end of the sampling
tube and prevent loss of the sample after
removal.
Care should be exercised that the sampling
tube has a diameter large enough to prevent
significant smearing and early plugging of the
tube (see Figure 1) so that the material below
the imaginary cylinder to be extracted is not
pushed aside and thus not sampled. Compaction
is an additional problem.
The "very surface" concentrations in the sedi-
ment are measured in most cases by collecting
and analyzing the solids suspended in the
natural water with in situ vigorous agitation
of the water near the sediment interface in
the sampling tube.
During the course of an investigation of
plutonium In sediments, [16] samples were
collected by Mound Laboratory and the U.S.
Environmental Protection Agency. The samples
were split two ways for analysis by each
laboratory. In the first case, unhomogenized
samples were split in the field and analyzed
by each respective laboratory. These results
are shown in Table I and as expected indicate
wide variances (error of -1167») in the data
obtained by the two laboratories. When the
field samples were dried and homogenized in
the laboratory, the data obtained is smoothed
out and the error obtained is less than 20%.
These data are shown in Table II. Thus,
again, the data show that large errors can be
encountered in sampling and aliquoting.
Figure 1
5
19-2
-------
Table I
INTERLABORATORY COMPARISON OF MOUND
AND EPA RESULTS ON FIELD-SPLIT SAMPLES*
Mound EPA Difference
Code
(nCi/g)
(nCi/s)
(7.)
EPA-17
0.284
±
0.035
0.047
-83.5
EPA-18
0.280
±
0.035
0.060
-78.6
EPA-1
0.165
±
0.023
0.230
+39.4
EPA-6
0.0052
±
0.0017
0.0038
-26.9
EPA-20
0.0011
±
0.0005
0.0019
+72.7
EPA-14
0.0009
±
0.0004
0.00044
-51.1
EPA-15
0.0009
±
0.0004
0.00096
EPA-3
0.0001
±
0.0001
0.00039
+390.0
EPA-13
0.0001
±
0.0001
0.00010
EPA-7
<0.0001
0.00044
>440.0
EPA-12
<0.0001
0.00019
>19.0
EPA-2
<0.0001
0.00012
>2.0
Average -116.0
*These samples were collected by the U. S. EPA near Mound
Laboratory. The samples were split between EPA and Mound in
the field without homogenization.
Table II
INTERLABORATORY COMPARISON OF MOUND
AND EPA 238pu DATA FROM SOIL AND SEDIMENT
SAMPLES HOMOGENIZED IN THE LABORATORY
Mound EPA Difference
Code
(nCi/k)
(TlCi/R)
(%)
EA1
0.0001 ± 0.0001***
0.00011*
0.0
EB1
<0.0001***
0.00012*
0.0
EC1
0.0029 ± 0.0011
0.0048
+65.5
EDI
0.0009 ± 0.0004
0.0011
+22.2
EE1
0.425 ± 0.0024***
0.440
+3.5
EF1
1.03 ±0.05*
1.13*
+9.7
EG1
0.0098 ± 0.0027
0.0108
+10.2
EH1
0.0238 ± 0.0053
0.026
+9.2
Ell
<0.0001
0.00098
0.0
EJ1
0.0010 ± 0.0005
0.0011*
+10.0
FA1
0.0094 ± 0.0026
0.0085**
-9.6
FE1
0.0138 ± 0.0025*
0.0181
+31.2
GA1
0.0004 ± 0.0002
0.00048
HA1
0.0047 ± 0.0016
0.0051
+8.5
IA1
0.0020 ± 0.0008
0.0025
+25.0
JAl
0.0007 ± 0.0004
0.0007
0.0
KA1
0.0309 ± 0.0065
0.0027
-91.3
LAI
0.0096 ± 0.0027
0.0109
+13.5
CE1
0.0302 ± 0.0064
0.024
-20.5
QE1
1.00 ± 0.09
0.920
-8.0
Average 16.9
*Mean of duplicates.
**Mea«i of triplicates.
***Mean of quadruplicates.
6
19-2
-------
REFERENCES
1. D. H. Denham, D. A. Waite, Summary of Selected AEC Contractor Environmental Surveillance
Techniques and Capabilities. BNWL-1817 (Draft), Battelie-Northwest, Richland, Wash.
(April 1974).
2. J. P. Corley, D. H. Denham, D. E. Michaels, and D. A. Waite, A Guide for Environmental
Surveillance at ERDA Installations. May 2, 1975.
3. Environmental Radioactivity Surveillance Guide. U. S. Environmental Protection Agency
Office of Radiation Program, ORP/SID 72-2, June 1972.
4. Principles of Environmental Monitoring Related to the Handling of Radioactive Materials.
ICRP Publication 7, International Commission on Radiological Protection, September 13
1965.
5. Measurement of Radionuclides in the Environment - Sampling and Analysis of Plutonium in
Soil. Regulatory Guide 4.x, Nuclear Regulatory Commission, March 1974.
6. Specification and Performance of On-Site Instrumentation for Continuously Monitoring
Radioactivity in Effluents. ANSI N13.10, Atomic Industrial Forum, Inc., Institute of
Electrical and Electronics Engineers, Inc., October 1973.
7. Measuring. Evaluating and Reporting Radioactivity in Solid Wastes and Releases of
Radioactive Materials in Liquid and Gaseous Effluents from Light Water-Cooled Nuclear
Power Plants. Regulatory Guide 1.21, June 1974.
8. John H. Harley, HASL Procedures Manual. HASL-300, ERDA Health and Safety Laboratory, 1972.
9. Handbook on Environmental Radiation Measurements, being prepared by Sub-Committee 35 of
NCRP.
10. IAEA Publication 16.
11. ICRP Publication 10.
12. Quality Assurance. ERDA Manual Chapter 0820, ERDA 0820-01 CONS., December 12, 1973,
13. W. E, Sheehan, M. L. Curtis, and D. C. Carter, Development of a Low Cost Versatile Method
for Measurement of HTO and HT in Air. MLM-2205, Mound Laboratory, February 14, 1975.
14. R. K. Stevens and T. G. Dzubay, "Recent Developments in Air Particulate Monitoring"
IEEE Transactions on Nuclear Science. Vol. NS No. 2, April 1975. '
15. D. R. Rogers, Mound Laboratory Environmental Plutonium Study. MLM-2249, Mound Laboratory,
September 15, 1975.
16. B. Robinson, D. R. Rogers, W. H. Westendorf, and H. A. Black, Mound Laboratory Plutonium-
238 Study - Off-Site Analytical Data: May-December 1974. MLM-MU-74-72-0001, Mound
Laboratory.
7
19-2
-------
THE NUCLEAR REGULATORY COMMISSION CONFIRMATORY
MEASUREMENT PROGRAM FOR ENVIRONMENTAL
AND EFFLUENT MEASUREMENTS
Bernard H. Weiss
Office of Inspection and Enforcement
U. S. Nuclear Regulatory Commission
Washington, D.C. 20555
Summary
The assessment of actual or potential exposure of man
from radioactive materials or radiation is the moat
important aspect of licensee effluent monitoring.
Assuring the public that such measurements are valid
relates directly to the NRC's responsibility for the
protection of the health and safety of the public.
Although NRC relies on licensee effluent and envir-
onmental data, a program utilizing confirmatory
measurements has been established to assure that
licensees and/or their contractors are capable of
making these measurements with reasonable accuracy.
The primary objective of this program is to be able to
relate licensee measurements to the national reference
standard, the National Bureau of Standards. A formal
system has been devised to establish traceability
between NBS and the NRC reference laboratory. The
program and results of that effort are discussed. The
NRC reference laboratory, in turn, tests licensees
with the use of split, duplicate or simulated samples.
The technical problems encountered, the criteria used
for Judging acceptability and the experience are
reviewed.
Introduction
The operation of nuclear power plants and other
nuclear facilities always entails the release of some
quantities of radioactive materials to the environment.
The experience of the nuclear industry has been that
these releases have generally been a small fraction of
regulatory limits.^ It has been the Commission policy
and a regulatory requirement that the design and
operation of licensed activities should maintain
radioactive effluent releases "as low as practica-
ble". 2 In order to provide guidance to applicants
and licensees of light water cooled nuclear reactors
on how to meet this concept, the NRC issued an amend-
ment to its regulations which provides numerical
guides for design objectives and limiting conditions
for operation to meet the criterion "as low as practi-
cable" for radioactive material effluents.3 These
design objectives are based, in a large part, on the
experience of industry. Adherence to these objectives
provides reasonable assurance that members of the
public will not receive a radiation dose from a
nuclear reactor in excess of a few percent of the
radiation protection guides established by the Federal
Radiation Council. The Commission is considering "as
low as practicable" design objective criterion for
other facilities in the nuclear fuel cycle.
In order to assess the actual or potential exposure of
man to radioactive materials or radiation, the Com-
mission requires licensees to conduct monitoring
programs.^'5 Effluent characterization programs are
usually able to measure the facility releases fairly
accurately since the licensee is able to measure
effluent streams containing radioactivity at much
higher concentrations than those found In the environ-
ment. Suitable dose models are available to then
estimate the potential or actual dose to the offsite
population. Use of environmental measurements to
estimate population dose generally has a much larger
uncertainty than when effluent data are used. This is
because the dose for individual pathways is much small-
er than the "as low as practicable" design objectives
and the measurement sensitivities are not always ade-
quate to detect these much smaller radioactivity
concentrations. In addition, the final step of the
pathway may be difficult to characterize or sample.
Consequently, the assessment of dose by use of effluent
measurements is considered the primary method. Envi-
ronmental measurements are important as a backup to
the effluent program and to determine if buildup of
radioactive material is occurring in the environment.
It Is our belief that part of NRC's responsibility for
the protection of the health and safety of the public
relates directly to assuring the public that licensee
environmental and effluent measurements are valid.
Acceptance of licensee results without some corrabora-
tive evidence of analytical accuracy will not be
beneficial to any party. Consequently, the Commission
is developing a comprehensive program which will test
the capability of analytical laboratories performing
environmental and effluent measurements and relate
these results to the national reference standard, the
National Bureau of Standards.
The Confirmatory Measurement Program
Traceability to NBS
A formal system has been devised to establish trace-
ability between NBS and the NRC reference laboratory,
Health Services Laboratory (HSL), Idaho Falls, Idaho.
Traceability to NBS involves more than merely the use
of an NBS source to calibrate a measurement system.
Testing by NBS must be part of such a program and the
laboratory desiring traceability must state the tests
needed and the degree of traceability desired. Table 1
shows the tests and acceptability which is the basis
for the NBS-NRC program which requires testing at least
once per year and, preferably, each six months.
TABLE 1
Test Acceptability*
Gamma Radiation**
1. <0.100 Mev 8X
2. 0.100-0.300 Mev 8%
3. 0.300-0.700 Mev 52
4. 0.700-1.100 Mev 5%
5. >1.100 Mev 5%
Beta Radiation 5%
Alpha Radiation 5%
Tritium 5%
NBS Testing of HSL Standards***HSL-6
-------
It is anticipated that other government agencies,
professional and industrial groups and private
companies, which for technical and quality control
purposes intercompare measurements with a user at a
lower level, will also participate directly with NBS
in establishing traceabllity to the NBS measurements
system. In this way, a scheme can be established
whereby essentially all the radiation measurements
made in the United States can be related to the
National Bureau of Standards. Such a scheme is shown
graphically in Figure 1.
FIGURE 1
XRACEABILITY
FIRST LEVEL LABORATORIES TRACEABLE TO NBS
NATIONAL BUREAU OF STANDARDS
NATIONAL MEASUREMENTS SYSTEM
SECOND LEVEL USERS RELATABLE TO NBS THROUGH TRACEABLE
LABORATORIES
Effluent Measurements
With regard to effluent measurements, in particular
those from nuclear power plants, NRC has incorporated
a "confirmatory measurements system" into its routine
inspection program of nuclear power plant effluents.
The objective for confirmatory measurements is to
evaluate the licensee's ability to measure radio-
activity in effluents accurately under usual operating
conditions. The measurement intercomparisone associ-
ated with this program normally include four
measurements for LWRs which present diverse problems
in both the type of measurement and sampling methodol-
ogy. The samples used are:
1. Liquid waste tank samples
2. Gaseous waste (offgas or gas holdup)
3. Particulate filters
4. Charcoal cartridges
In general, an attempt is made to obtain samples, which
have identical physical and chemical composition, for
each participating laboratory. Ideally, it would be
best for both laboratories to analyze the same sample.
This can be accomplished for particulate filters and
charcoal cartridges but is not practical for the other
two categories above. For liquid waste tank samples,
a large sample is taken from a well mixed storage tank
and then shaken vigorously before obtaining aliquots
for the participating laboratories. This apparently
straightforward method of sample splitting has demon-
strated some inadequacies. The most significant pro-
blem identified has been in retrieving the sample from
the sample container for counting in an appropriate
geometry. Fractionation of radionuclides between
filtrate, filterable material and especially large
percentages of activity on the walls of the sample
container has posed serious analytical problems. Since
we are interested in the measurement of the activity
that goes into the sample container rather than what
may pour out, a procedure has been devised to assure
that essentially all the activity is retrieved in
stable solutions. HSL evaluated several sample treat-
ments for these liquid wastes including the addition
of carriers, filtration at the time of collection and
the addition of ashless filter paper. The latter has
been found to be the most satisfactory alternative in
considering both the reduction of activity on the
container walls and the ease of use in the field.
Paper pulp added to the sample provides sufficient
surface area, compared to the container wall, for most
radioactivity to deposit on the pulp. After filtering
the paper pulp, the filtered material can be dissolved
and reconstituted with the filtrate.
Offgas samples are not split from a larger sample but
rather taken in ways which minimize the possibility
of obtaining poorly split samples. For BWRs, repeti-
tive samples can be obtained within a short time. In
most PWRs, the sampling devices can be lined up in
series and the offgas purged through the samplers.
Consideration has been given to collecting samples
in a manner similar to the liquid waste tank samples,
i.e., taking aliquots from a large sample. There is
some difficulty in this approach because of the awk-
wardness of such a procedure and the high radiation
levels encountered with some offgases.
In addition to splitting actual effluent samples, NRC
has devised split and duplicative samples to further
assess the analytic capability of licensees, parti-
cularly prior to startup, before there is any radio-
active waste. Inspectors have available a charcoal
cartridge, (a type used extensively by LWR licensees)
which is spiked with barium-133 to assess the calibra-
tion of the instrumentation. Barlum-133 has a gamma
energy very close to that of iodine-131 and a very
Bimilar radiation intensity. In addition, particu-
late filters, spiked with several long lived nuclides
usually identified in gaseous effluents, and enclosed
in plastic, are available. For liquid effluents, the
NRC reference laboratory has developed a simulated
liquid waste containing several gamma nuclides plus
strontium 89, 90 and tritium. Spiked gas samples are
now being developed by the NRC reference laboratory.
After measurements from the NRC reference laboratory
and the licensee are obtained, a comparison must be
made to determine if significant discrepancies exist.
In considering the question of what agreement is
between a set of measurements, NRC has had to consider:
1. The accuracy needed to characterize effluent
releases;
2. the state-of-the-art with respect to the
analytical procedures used; and
3. the experience of various laboratories in
assessing effluent samples.
Although temptation is always strong to base a test
for comparison on established statistical tests,
neither the population considered (two) nor the
accuracy required by the regulatory agency could
justify the application of such a test. Consequently,
an empirical relationship was established using the
above considerations. This test utilizes the concept
that as a measurement approaches the limits of
detection, the relative difference between values
which "agree" can increase. This relationship of
the measurement value to the limit of detection is
expressed as resolution and is defined as:
Resolution- ^ Laboratory Value
NRC Value Uncertainty (lo)
After the resolution is calculated along with the
ratio of the licensee value to the NRC value, one
can determine whether the value is in agreement,
possible agreement or disagreement from Table 2.
19-3
-------
It should be noted that there are two "possible agree-
ment" categories, A and B. This results from NRC
experience in intercomparing measurements. The "A"
group is for those measurements where the licensees
have had limited difficulty, e.g., tritium measurements
and iodine on charcoal adsorbers. "B" is reserved for
the more difficult analyses, i.e., strontium 89 and 90
determinations and measurement of low energy gamma
ray emitting nuclides.
TABLE 2
Resolution
Licensee value
NRC lab value
Possible
Possible
Agreement
Agreement A
Agreement B
<4
0.4 - 2.5
0.3 - 3.0
no comparison
4-7
0.5 - 2.0
0.4 - 2.5
0.3 - 3.0
8-15
0.6 - 1.66
0.5 - 2.0
0.4 - 2.5
16 - 50
0.75-1.33
0.6 - 1.66
0.5 - 2.0
51 - 200
0.80-1.25
0.75-1.33
0.6 - 1.66
>200
0.85-1.18
0.80-1.25
0.75 - 1.33
State generally collect samples from the same number
of locations as the licensee. Where split samples are
not feasible, the State operates side-by-side sampling
stations. This latter technique is used principally
for air particulates and direct radiation monitors.
Results from the split and side-by-slde samples can
then be compared and evaluated on a long term basis.
With regard to simulated environmental radioactivity
samples, NRC is not planning to develop its own pro-
gram. Most laboratories doing analyses of environ-
mental media belong to the Environmental Protection
Agency's Environmental Radioactivity Laboratory Inter-
comparison Studies Program. NRC did not feel that
it was necessary to duplicate the EPA program which
has been well recognized as doing commendable and
meaningful work in this area. Discussions have been
held between the two agencies and there has been
agreement to cooperate more, closely in the area of
radioactivity measurements quality control by con-
sidering NRC needs in the EPA cross-check program.
NRC licensees will be encouraged to have all
laboratories analyzing environmental samples for
radioactivity to participate in the EPA program and
take some corrective action if a significant dis-
crepancy with the known EPA value occurs.
If results indicate a disagreement, the next step is
generally to retest as soon as practicable. Where
the disagreement persists and a violation of a techni-
cal specification could have occurred, the licensee
will be cited for violation of NRC regulations. When
the values in disagreement are consistently lower than
the NRC reference laboratory and a small fraction of
license specifications, the licensee will be encouraged
to use an appropriate correction factor for his
measurements until the discrepancy can be resolved.
It should also be understood that as part of the
regular NRC inspection program, the data and infor-
mation developed by the Inspector is made public.
Consequently, the results of all intercomparisons are
public information and are made part of the public
record after it is determined that any discrepancies
are not the result of calculaticmal errors or sample
splitting.
Environmental Measurements
Confirming individual environmental measurements in the
same manner as used for effluents is not a workable
procedure. The radioactivity content of environmental
media due to nuclear facilities is generally below or
very close to the analytical sensitivity of the pro-
cedures utilized. Consequently, meaningful inter-
comparisons are not possible. In order to obtain a
reasonable estimate of the accuracy of a licensee* s
program there are two alternatives: 1) intercompari-
sons of long term measurements of environmental trends
by comparing repetitive measurements of split or side-
by-side sampling; and 2) intercomparison of spiked and
simulated environmental media samples where the
radioactivity content is well known and sufficient
for accurate analytical analysis. In a sense, the
first alternative evaluates the overall environmental
program performance while the latter procedure consi-
ders the licensee's analytical capability. NRC is
using both of these techniques to check the validity
of licensee environmental results.
Long term repetitive measurements of environmental
trends will be accomplished with the assistance of
State agencies. NRC has a collaborative program
whereby State environmental surveillance offices
periodically split environmental samples with licensees,
e.g., milk, water, fish, etc. These splits are not at
the same frequency as that of licensees nor does the
Program Summary
The Nuclear Regulatory Commission is conducting an
expanding program to provide reasonable assurance
as to the accuracy of the results of radioactive
effluent and environmental measurement results from
nuclear facilities. Intercomparisons of the analyti-
cal results of split and duplicate samples are made,
where feasible, to determine the licensee's accuracy
in characterizing the radioactivity releases from his
facility and the impact on the environment. Where
this technique is not possible, spiked or simulated
samples are utilized to determine the licensee's
capability to prepare and measure these artificial
samples.
References
1. U. S. Nuclear Regulatory Commission, Summary of
Radioactivity Released in Effluents From Nuclear
Power Plants During 1973, NUREG-75/001, January
1975.
2. Title 10, Chapter 1, Code of Federal Regulations,
Part 20, Section 20.1.
3. Nuclear Regulatory Commission, Radioactive Material
In Light-Water Cooled Nuclear Power Reactor
Effluents, 40 F.R. 19439 - 19443, May 5, 1975.
4. U. S. Atomic Energy Commission, Regulatory Guide
1.21, Measuring, Evaluating and Reporting Radio-
activity in Solid Wastes and Releases of Radio-
active Materials in Liquid and Gaseous Effluents
from Light-Water Cooled Nuclear Power Plants,
Revision 1, June 1974.
5. U. S. Nuclear Regulatory Commission, Regulatory
Guide 4.1, Programs for Monitoring Radioactivity
in the Environs of Nuclear Power Plants, Revision
1, April 1975.
6. U. S. Environmental Protection Agency, Environ-
mental Radioactivity Laboratory Intercomparison
Studies Program 1975, EPA-680/4-75-002b, May
1975.
19-3
-------
INSTRUMENTATION FOR OFF-SITE REACTOR PLUME STUDIES*
R. C. Ragaini, D. E. Jones,' and G. W. Huckabay
Lawrence Livermore Laboratory, University of California
Livermore, California 94550
Summary
As part of the Lawrence Livermore Laboratory program to
assess the environmental impact of various types of electric
power plants, methods are being investigated for the real-time
analysis of reactor plume isotopic exposure rates. The Bio-
Medieal and Environmental Research Division mobile laboratory
that was developed for terrestrial radioactivity measurements has
been modified for atmospheric studies. The existing Ge(Li)
spectrometer and high-pressure argon-ion chamber-detector sys-
tems have been augmented by the addition ot wind speed and
direction measuring capability and a second, remotely operable,
ion chamber. The mobile laboratory and techniques for plume
measurements are discussed.
Introduction
In recent years the generation of electricity and the sizes of
electrical power plants have increased at unprecendented rates.
This growth intensifies the need for a better understanding of
health and environmental impacts of power plant emissions; thus
we may address the long-term issues of site selection and alter-
natives for both emission, control and production methods.
Our nuclear power plant studies were centered on techniques
and feasibility of measuring individual exposure rates of major
radioactive components of plume in real time. As a prelude to
the plume studies, in situ measurements of terrestrial radionu-
clides were performed tor several sites tor each ot five preoper-
ational power plants. These included the three types bf power
reactors in U.S. operation: Boiling Water (BWR), Pressurized
Water (PWR), and High Temperature Gas Cooled (HTGR). The
five power stations studied were: Fort Calhoun, Nebraska; Fort
St. Vrain, Colorado; Cooper Nuclear, Nebraska; Rancho Seco,
California; and Diablo Canyon, California. Some preliminary
plume measurements were made near an early BWR and the
Lawrence Livermore Laboratory (LLL) Pool Type Reactor for
technique development.
Our recent activities have included three major areas:
(1) documentation of the preoperational data and incorporation
of the data into a computerized data bank for selective retrieval
and additions. Special emphasis has been given to the highly var-
iable man-made nuclides already present at the sites and the
ability to selectively retrieve information in several categories,
e.g. by reactor, nuclide, date, exposure rate, type of site, etc.
(2) We looked at relatively simple methods of estimating plume
component exposure rates that might be useful for real-time
field work and planned to compare these methods with more
elaborate computer methods used by others for measured field
data. (3) A cooperative measurements program has been planned
to collect data that may be used to determine the effect of a
gas holdup system uppn off-site and on-site exposure rates. This
program will be a joint effort between Pacific Gas and Electric
(PG&E), Health and Safety Laboratory (HASL), and the LLL
BioMed and Atmospheric Physics Divisions, The data collected
will also be used for the program in (2) above.
Terrestrial Measurements
The preoperational nuclear power plant measurements were
carried out at approximately 15 sites for each of the five reac-
tors. In general the sites were chosen close to the reactor's TLD
sites with additional consideration given to the likelihood of
future disruption for planting, construction, or other major soil
disturbance. Total external gamma exposure rates ranged from
about 6 nR/hr for a near sea-level altitude with low natural
radionuclide soil content to greater than 18 /xR/hr for higher
elevations with high natural radionuclide soil content. Man-made
radionuclides included 137Cs, 95Zr, 144Ce, and 125Sb, although
'3'Cs was the only one with high enough concentrations to be
found consistently. The '^^Cs exposure rates were highly vari-
able and generally less than 10% of that from the natural ter-
testrial sources. Figure 1 shows the soil profiles for '3^Cs at
three permanent pasture sites and Fig. 2 shows soil profiles for
sites that have been altered since the cesium deposition.
Although these exposure levels from '^Cs were very low,
they were not negligible compared with the small increases above
Site
LLL-K
£•
-o
CD
Site
Site
0.01
Soil depth — cm
Fig. 1. Depth distribution of Cs at three permanent pasture
sites around Cooper Nuclear Generating Station. The
slopes were 0.33 ± 0.01, 0.34 ± 0.01, and 0.13
± 0.01 cm"' for sites LLL-K, 52, and 33, respectively.
The solid lines are least-squares fits to an exponential
function.
*This work was performed under the auspices of the U.S. Energy Research & Development Administration,
tCurrent address: Dosimetry Branch, Health Services Laboratory, USERDA, Idaho Falls, Idaho 83401.
-1- 19-4
-------
ation and the relatively short distance of the reactor from LLL,
it was the logical site for a preliminary field study.
Experimental
The Humbolt Bay Held experiment was designed to help
understand the practical problems and instrumental requirements
of real-time plume measurements with a mobile system. The
BioMed mobile laboratory- that was used for the in situ terres-
trial measurements was also used in this study. Figure 3 shows
the major features of this laboratory. In addition to its "down-
looking" Ge(Li) spectrometer and high-pressure argon-ion chamber,
a limited meteoroligical capability was added. This addition
included standard commercial wind speed and direction trans-
ducers that input data into electronic modules chosen for com-
patibility with the other instrumentation. Wind speed was taken
in the conventional way by counting sensor rotations in a
NIMBIN scaler module. Wind direction was recorded on a
scaler by sending the potentiometric sensor a reference voltage
with the position-determined output sent to a voltage-to-
frequency converter.
Table 1. External gamma exposure rates (fiR/hr).
Sites
Total
Terrestrial
Fallout
Rancho Seco
IS
5.8- 9.8
2.9- 7.1
0.09-0.27
Fort St. Vrain
17
14.2-18.7
9.6-15.3
0.09-0.47
Fort Calhoun
19
9.2-10.9
O
00
0.03-0.75
Cooper
17
9.4-10.6
6.1- 7.6
0.16-0.70
Diablo Cunyort
15
5.8-11.6
2.1- 9.3
0.02-0.26
2 6 10 14 18 22 26
Soil depth (cm)
Fig. 2. Depth distribution of '^Cs at two sites where an expo-
nential distribution was not found. The solid lines are
least-squares fits to an exponential function.
the natural background of current interest. Furthermore, the
variability does not allow for estimating levels of '3?Cs without
individual site measurement.
Many analyses of man-made radionuclides assume a surface
deposition for "new" sources and an exponential distribution for
older activity.'-- Our extensive soil sampling data show these
assumptions to be an oversimplification in many eases. We are
examining the results to answer the following questions: (1) What
was the range of errors associated with assumed distributions
when compared with the actual measurements? (2) Can the man-
made radionuclides be better quantified through application of
known information about each site and the most likely period
of deposition without extensive soil sampling? (3) Can we mod-
ify the data analysis routines to improve the results based upon
items (1) and (2) above?
This preoperational phase of the power plant studies has
been completed. A first summary paper was presented at the
December 1974 IEEE Nuclear Science Symposium at Washing-
ton, D.C.3 Tables I, II, and 111 show some of the results from
that paper. The first two show the overall character and ranges
of the external exposure rates. Table III shows the variability
in the preoperational '37q. distributions for the relatively simple
case of the permanent pasture sites.
The next phase was to study the applicability of similar field
techniques to real-time radioactive plume measurements. The
BioMed Division and PG&E have previously cooperated in marine
biological measurements at the latter's Humbolt Bay BWR that
has been operating since 1963. With this past history of cooper-
Table II.
Total external
gamma exposure
rates (pR/hr)*
Utilities
LLL ion
exposure time
Utilities
Power plant
chamber^
(mo)
TLD exposures
Rancho Seco
L*
00
00
3
7.5-12.8
Fort St. Vrain
14.2-18.7
1
21.6-27.0
Fort Calhoun
9.2-10.9
3
9.2-14.9
Cooper
9.4-10.6
3
6.3- 8.1
Diablo Canyon
5.8-11.6
1
11.8-15.9
%
TLD exposures and ion-chamber readings are not necessarily
at the same sites.
t Ion chamber measurements taken for 1 hr at each site.
Table III. '^7Cs external gamma exposure rates (nR/hr), permanent
pasture sites.
Exposure
Slope Concentration rate
Power station No. (cm"') (nCi/m^) (jiR/hr)
Fort Calhoun
GM-L
0.42
94
±
3
0.47
± 0.01
Nebraska
GM-G
0.14
140
5
0.47
* 0.02
Fort St. Vrain
F14
0.39
98
±
4
0.47
± 0.02
Colorado
A3 2
0.37
72
±
3
0.33
± 0.01
Cooper Nuclear
33
0.13
143
±
5
0.45
± 0.02
Nebraska
LLL-K
0.33
157
±
4
0.70
± 0.02
52
0.34
122
±
2
0.56
± 0.01
Rancho Seco
11
0.29
48
±
2
0.18
± 0.01
California
8
0.64
53
±
3
0.27
t 0.01
Diablo Canyon
2
0.15
56
1
3
0.19
± 0.01
California
8
0.12
47
±
4
0.15
± 0.01
7
0.092
56
±
4
0.17
± 0.01
-2-
19-4
-------
Fig. 3. The Van-mounted mobile laboratory. Shown in the
figure are the Ge(Li) spectrometer, pressurized argon-
ion chamber, wind measuring devices, and electronic
support equipment.
The resulting pulse train was counted in a scaler identical to
those used for the wind speed and the ion chamber. These
integrated, timed counts were printed 011 a teleprinter and
recorded 011 paper tape. The sensor was orientated toward the
source, e.g., the power station, with the relative angle (to north)
recorded. This method kept the ambiguous angular region, where
oscillation rates go from maximum to minimum within a few
degrees, to wind directions such that emissions were blowing
away from the detector. The primary wind data came from the
reactor meteorological tower, but the availability of local data in
a more quantitative format was of assistance to operators during
the measurements. It also added to the usually limited meteoro-
logical information that was available for analysis.
For this first plume study, data were taken for 24 hr for
3 days, including times with plume both present and absent.
Ge(Li) spectra were taken hourly for a 50-min counting time.
This gave 10 min for examination of the data and for storage
011 the seven-track magnetic tape. As each spectrum count was
started, the sealer-timer system was started to record the ion-
chamber exposure rates and wind data. These data were printed
and punched on paper tape every 30 s with the data taken for
20 s. The scalers were manually stopped after each spectrometer
period so that the paper tape could be labeled with the proper
spectra numbers during the 10-min setup and storage periods.
For this first study the data were all input to the LLL Octo-
pus computer system for analysis after return from the field.
Site Selection
One of the major problems of environmental monitoring has
been the selection of suitable sites for representative measure-
ments. This is especially true if the data were to be used for
modeling and characterization of more general conditions for the
surrounding vicinity. In the preoperational studies the selection
of relatively fiat and undisturbed sites that were expected to
remain so was often very difficult. Similarly, "good" sites for
meaningful meteorological data from a mobile system were rare.
Figure 4 shows a map of the Humboldt Bay area. The prevailing
winds were generally directed inland, but the nearby hills caused
a splitting effect with a southerly fraction and another fraction
going up the narrow Elk River valley. With stronger winds and
110 inversion conditions, the wind flow can pass over Humboldt
Hill. To test any comprehensive model one would need a num-
ber of sites operating simultaneously and arranged in an arc over
the hill. The implementation of such an experiment would be
difficult even if one had the required number of mobile moni-
tors because suitable sites are extremely limited. Most of the
hill was covered by a housing development.
After consultation with reactor personnel and examination of
several areas, we selected a site at Fields Landing adjacent to
the bay and approximately one mile due south of the reactor.
Open area surrounded this location for distances equal to several
heights of any obstructions on all sides of the site. Thus the
site wind data should be characteristic of the immediate vicinity.
The dominant wind pattern at Fields Landing should be roughly
south because of the effect of the hills to the east.
Meteorological data were available from only three sources,
the reactor tower, the mobile lab tower, and the National Weather
Service (NWS) station atop a building in Eureka. The reactor
tower data were manually transcribed from strip charts for 5-min
intervals and the NWS data were taken from hourly summary
sheets. For real-time measurements only the mobile station data
would have been readily available for immediate calculations.
Results
Some typical exposure rate data are shown in Figs. 5 and 6
and Tables IV and V. Figure 5 and Table IV show the wind,
ion chamber, and spectrometer results for the Fields Landing site
with minimal plume contributions to the exposure rates. Figure 6
and Table V show the corresponding data for an hourly period
of maximum plume contributions. Computer codes were written
to read and summarize the raw data into computer files in the
LLL Octopus system. Figures 5 and 6 were prepared by a com-
puter code that was written to present the data for convenient
study of the temporal behavior l'rom the large amount of data
generated in such studies.
The wind vectors were vector averages over the 10-min
periods shown and were plotted to allow easy selection of
source and detector transport conditions for analysis. Unfortu-
nately, during the Humboldt Bay measurements a slow weather
front was present that was characterized by low variable winds
and unstable transport. This resulted in a meandering "plume"
condition that made quantitative estimates difficult so that the
evaluation of the Ge(Li) method was limited to general obser-
vations. The large change in efficiency of the "down-looking
diode" for airborne radionuclides further complicates the anal-
ysis. The major plume isotopes were readily observable as shown
in the center spectrum of Fig. 7. The top spectrum was taken
of a sample from the off-gas ejector before the release point
and the bottom shows the natural radionuclides present during
an upwind condition for the site of the center spectrum.
The estimates of the individual isotopic exposure rates used
the spectral data and simple line or plane source models. The
sum of these generally lies within a factor of two to five of the
ion chamber values for the 50-min counting times. These differ-
ences were not random as can be seen in Fig. 8. We assume the
major problem was inadequate handling of the wide range in
efficiencies of the Ge(Li) spectrometer for the ill-defined diffuse
plume source. No specific patterns have been discerned in the
handling of the various energies from these simplistic models.
The calculated isotopic exposure rates did not change markedly
with large changes in the assumed heights of the simple plume
models.
The isotopic calculations were performed using the same
computer code that was used for the terrestrial data. The code
was written by Anspaugh with the method described in Ref. 2.
Inputs to the code include tables of counting rates in the pho-
topeaks per unit soil activity (Nf/S) and dose conversion factors
for each of the isotopes. Using the measured overhead efficiencies
and the assumed source geometries, we calculated Nf/S tables
where the S was the activity per unit area for the overhead
plane source model. To calculate dose conversion factors we used
the buildup factors of the form (1 + k/ix), where k was equal
to (ju - /uu)/na.4 Because the efficiencies enter the calculation
only through the Nf/S values, the dose rates were overestimated
if the detector "sees" more activity surrounding it at larger
angles. The measured efficiency for the down-looking Ge(Li)
¦3- 19-4
-------
Humboldt Bay
Power Plant
Humboldt
Bay
LLL
Van Site
Fields Landing
L ^
One-half mile
Scale
Fig. 4. Map of the area surrounding the Humboldt Bay Reactor near Eureka, California.
FIELDS LANDING,
TH AVO(UR/HR)• 7.29 SIGMA- 0.7M
NAT. HEATHER HINDI 4 MPH, E
IW LOWEST VISIfl. • 10.00 MILES
ins YIS. OBSTRUCT. BLANK
SPECTRA THI396
NHS TEMPERATURE • -0
NHS SUNSHINE -0 M1N
NHS SKY COVERtO-IOI -0
KLT 1
ft I SMC
mmm
n
sfTT
TIME (MIN PAST MIDNITE),
741002
Fig. 5.
FIELDS LANDING.
TW AUG» 25.20 SIGMA- 6.7
NAT. WEATHER WIND: 4 MPH. E
NWS LOWEST OISIB. = 1.25 MILES
NWS OIS. OBSTRUCT. FOGGY
SPECTRA TW1372
NWS TEMPERATURE = 46
NWS SUNSHINE -O MIN
NWS SKY C0UER<0-10> -G
/
/
^7
Computer generated summary of climate and exposure
rate data for SO min in the absence of a reactor plume.
Van wind and exposure data are averaged from 20-s
samples repeated at 30-s intervals. Reactor meteorological
and National Weather Service data were taken from strip
charts and summary sheets respectively. Ge(Li) spectra
were taken simultaneously with each data set above and
repeated at hourly intervals.
-
HiSlitl ji [
"iijiiiilililHIi
I
TIf1E
-------
Table IV. Isotopic exposure rates from Ge(Li) spectrometer with minimal contributions. Plume com-
ponents are based on 1-m high plane source model.
Spectrum No. 1396
Fields Landing Count 38 started 0700 hr Feb. 17, 1974 Ion chamber 7.29 ± 0.79 nR/hr
External exposure rate
Radio-
nuclide
nCi/kg
Sigma
fiR/hr
Sigma
5.765E + 00
2.202E -01
1.032E + 00
3.941 E -02
3.410E- 01
2.127E - 02
9.617E -01
5.999E - 02
4.439E- 01
2.705E - 02
8.079E - 01
4.923E - 02
4.894E - 02
1.508E - 02
3.374E - 01
1.040E - 01
7.307E - 02
1.302E -02
1.47 IE + 00
2.621E - 01
5.260E - 02
2.028E - 02
3.617E - 01
1.395E - 01
9.954E - 02
3.198E -02
1.085E + 00
3.486E - 01
1.987E - 02
6.722E - 03
4.664E - 01
1.578E - 01
40
232'
238
19
Th90
C_U92
lKt36
1^36
Rb37
>38Xe54
138*55
Subtotal external exposure rate resulting from naturally
occurring radionuclides
Total external exposure rate
2.802E + 00
6.523E + 00
8.704E - 02
5.029E - 01
Table V. Isotopic exposure rates from Ge(Li) spectrometer with significant plume contributions. Plume
components are based upon a 1-m high plane source model.
Spectrum No. 1372
Fields Landing Count 14 started 0700 hr Feb. 10, 1974 Ion chamber 25.20 mR/hr
Radio- External exposure rate
nuclide
nCi/kg
MCi/m2
Sigma
jiR/hr
Sigma
4°K,9
5.793E + 00
3.031E -01
1.037E + 00
5.425E - 02
232Th90
4.090E -01
3.127E -02
1.153E + 00
8.819E -02
23%2
8.332E -01
5.595E - 02
1.516E + 00
1.018E - 01
87*36
5.764E - 01
2.070E - 02
3.974E + 01
1.427E -01
88&36
8.296E - 01
2.843E - 02
1.670E + 01
5.722E - 01
88Rb37
5.052E -01
4.250E - 02
3.474E + 00
2.923E - 01
135Xe54
1.109E + 00
2.022E - 02
4.394E + 00
8.013E - 02
13SmXe54
2.382E - 01
1.186E-02
1.505E + 00
7.492E - 02
,38Xe54
6.226E - 01
2.981 E -02
6.786E + 00
3.249E - 01
138Cs55
5.336E - 01
1.707E - 02
1.2S2E + 01
4.007E - 01
85mKr36
1.802E - 01
3.414E - 02
4.927E - 01
9.335E - 02
Subtotal external exposure rate resulting from naturally
occurring radionuclides
3.707E + 00
1.452E - 01
Total external exposure rate
5.356E + 01
8.609E - 01
detector is shown in Fig. 9. The abscissa represents the overhead
angle as measured from the detector centerline. At Humboldt
Bay the detector was most likely immersed in a semi-infinite
cloud with the cloud contributing more counts from the horizon-
tal angles than assumed from the simple models. An up-looking
detector has recently been obtained that corrects the problem of
large efficiency changes with angle. Figure 10 shows the meas-
ured efficiencies for this new detector.
Conclusions
The data above have shown that a mobile system using a
Ge(Ii) spectrometer can be used to estimate the isotopic expo-
sure rates from a mixture of radionuclides. Some measurements
have also been made of the low levels of '"Ar from the 3-MW
LLL Pool Type reactor. Both measurements indicate the Ge(Ii)
system can detect the presence of airborne radionuclides at
lower levels than those detected by the exposure rates and stand-
ard deviations from the ion chamber as we used them. In addi-
tion the spectrometer system has proven to be very reliable and
efficient to use in the field. Verification of the ultimate sensi-
tivities and usefulness of the Ge(Li) system for plumes remains
to be done.
We have planned to continue these evaluations through a
much more detailed field study and through Monte Carlo calcu-
lations for some assumed cloud source distributions for an up-
looking detector.
References
1. H. L. Beck, J. De Campo, and G. Gogalak, "In-situ Ge(Li)
and NalfTl) gamma-ray spectrometry," USAEC Health and
Safety Laboratory, New York, Rept. HASL-258 (1972).
2. L. R. Anspaugh, P. L. Phelps, P. H. Gudiksen, C. L.
Lindeken, and G. W. Huckabay, "The in-situ measurement of
radionuclides in the environment with a Ge(Li) spectrometer,"
-5-
19-4
-------
l—'—i—1—i—1—i—1—i—T—i—'—i—»—I ' r
Ge(L1) Spectra of Terrestrial and BWR Off-Gas and Plume Radionuclides
138Xe,397
Kr,2196
Xe,2005 | r88Kr,2232
\ | |"138Xe, 2252
Rb,2569
c
-------
228.
V-3 high angle response
2615 keV
01
239 keV
Horizontal angle 1
Horizontal angle 2
2
Degrees from detector axis
Fig. 9. Response of down-looking Ge(Li) detector versus the
overhead angle measured from the detector centerline.
— 228
V-4 high angle response,
2615 keV
239 keV
0 10 20 30 40 50 60 70 80 90
Degrees from detector axis
Fig. 10. Response of up-looking Ge(Li) detector versus the over-
head angle measured from the detector centerline.
-7-
19-4
-------
NOTICE
"This report was prepared as an account of work
sponsored by the United States Government.
Neither the United States nor the United States
Energy Research & Development Administration,
nor any of their employees, nor any of their
contractors, subcontractor!, or their employees,
makes any warranty, express or implied, or
assumes any legal liability or responsibility for the
accuracy, completeness or usefulness of any
information, apparatus, product or process
disclosed, or represents that its use would not
infringe privately-owned rights."
JRH/rt
19-4
-------
DEVELOPMENT OF A NATIONAL BUREAU OF STANDARDS
ENVIRONMENTAL RADIOACTIVITY STANDARD: RIVER SEDIMENT
J. R. Noyce, J. M. R. Hutchinson, W. B. Mann, and P. A. Mullen
Radioactivity Section, Center for Radiation Research
National Bureau of Standards, Washington, D.C. 20234
Summary
The National Bureau of Standards has developed and
produced a radioactivity standard of a fresh-water sedi-
ment for use especially in environmental radioactivity
measurements around nuclear and coal-burning electric
power plants. The radioactivities of 28 nuclides in
the sediment have been measured, of which 10 are certi-
fied. The development and production of this new Stan-
dard Reference Material (SRM 4350) are discussed.
Introduction
The National Bureau of Standards fresh-water-sedi-
ment standard is a homogenized, freeze-dried sample of
fine particulate sediment in which 28 a-, x-, and
y-ray-emitting nuclides have been identified and meas-
ured. This Standard Reference Material (SRM) is an im-
portant step in filling a need for mineral materials
containing well-characterized radioactivities in chem-
ical forms and concentrations that result from natural
processes. Since the behavior of trace quantities of
radionuclides is not always well understood, teste by
chemists of their radionuclide assay procedures using
this material should be more reliable than similar tests
on, for example, spiked materials.
The problems encountered in the development of this
SRM were connected basically with the homogenization of
the material and the measurement of the radioactivities
in it.
In a mixture of particles of widely varying sizes
the smaller ones tend, upon shaking, to settle to the
bottom. If the radioactivity per gram of material
varied with particle size, such segregation would cause
variations in the radioactive concentration obtained for
subsamples of the material. We therefore considered It
essential to limit the range of particle sizes in the
sediment to be incorporated into the standard. By doing
so, we were able to achieve adequate homogenization
solely by mixing, thus preserving the mlneralogical
characteristics of the individual particles.
Based on the usual accuracy requirements of radio-
logical monitoring, and the normal minimum sizes for
test samples, we felt the need to show that quantities
of a sediment standard as small as 10 grams varied in
radioactivity per gram by not more than ±5% from one
sample to the next. For most of the radionuclides
present we were successful, but for some a-particle-
emltting nuclides departures from radioactive homoge-
neity could be as large as ± 10%.
Many of the radionuclides were assayed by more
than one method, which not only allowed us to compare
and average the results for a radionuclide but also, in
some cases, to compare analytical procedures.
A noval method will be described for the calibra-
tion of the Ge(Li)-detector system for photopeak effi-
ciency as a function of energy with bulk sediment
samples.
The purpose of this paper is to provide information
about this sediment standard and its development, so
that persons who wish to use it, or who wish to produce
a similar material, may be guided by the practical ex-
perience we have gained.
Experimental and Results
Preparation and Testing for Homogeneity
The sediment was collected in a slack water area of
a river that was about 10 meters deep. The accumulation
rate there was estimated to be an average of 20-to-25 cm
per year of fine-grained sediment. The initial sample
was a composite of many cores, each 15 cm in diameter by
61 cm in length, that were taken with a gravity-operated
corer.
Measurements of samples of the wet, unmixed material
showed that the total y-ray count rates per gram varied
by a factor of three. In addition, careful examination
of a later sample revealed a "hot" particle whose radio-
active content was identified as almost exclusively 60Co.
Its Y-ray count rate was equivalent to hundreds of grams
of "normal" sediment. An assessment of the extent of
radioactive homogeneity was performed at each step of the
preparation procedure by determining the gross y-ray
emission rates of weighed samples.
The initial sample was mixed wet in a large V-cone
blender. The addition of distilled water, approximately
10% by weight, was necessary to achieve good flow in the
blender. After mixing wet, the measured variation in
the gross y-ray emission rate per gram was no more than
The size distribution of the sediment particles was
determined by wet seiving, following the method described
by Guy.1 The results (see Table 1) indicated the material
consists almost entirely of silt and clay.1 The gross
¦y-ray count rates per gram were measured on each frac-
tional size range after drying, and the relative count
rates are shown in Figure 1. Also shown are the results
for corresponding ranges in size from a sample of dry-
sieved sediment. The 60Co activity concentration was
examined for each wet-sieved fraction and was found to
vary with particle size in the same manner as did the
gross Y-ray count rate. The likelihood of achieving
homogeneity appeared to be greatest by using only that
material for the SRM which passed through the smallest
mesh sieve, because the count rates per gram (dry) were
equal for both wet-sieved and dry-sieved sediment.
Half of the sediment was freeze dried and then re-
mixed. It was next mechanically shaken through a series
of 20-cm diameter, braes, D.S. Standard Sieves!2 97,
14, 25, 45, 80, 170, and 325. (The mesh openings are
2830, 1410, 707, 354, 177, 88, and 44 vim on a side, re-
spectively.) The undersides of the last two sieves were
brushed after every few minutes of shaking in order to
reduce clogging. The material passing the final sieve
was mixed for 10 hours, then packaged in 100-gram quan-
tities in numbered, screw-lid glass jars. The blender
had a so-called intensifier bar that was used during the
mixings prior to sieving.
The sieve shaker became quite warm during extended
periods of use. Layers of cardboard were placed between
it and the receiving pan to reduce heating of the sedi-
ment. Even so, the temperature of the final sieve sam-
ples roae at times to approximately 35 C, and It aearned
desirable to determine Whether such heating would change
the denBlty of the material through permanent loss of
water of hydration. A weighed sample, which had previ-
ously been kept cool while sieving, was therefore
1
19.5
-------
subjected to repeated cycles of heating in an oven at
40 C for times up to several days, then cooling to room
temperature and reweighing. The largest change in mass
that was observed was only 0.2%. However, heating at
110 C caused some water to be irreversibly lost from
the material.
Samples of prepared material were tested to see
whether segregation of particles according to size was
likely to occur during shipment to a purchaser. Con-
tainers of the sediment were shaken and vibrated, in
one instance for several days, and then the upper and
lower portions of their contents were compared for evi-
dence of motion-induced inhomogeneities. None was
found.
Extensive tests were conducted to determine the
degree of homogeneity of Y-ray-emitting nuclides in the
final material. Nineteen jars of sediment were randomly
selected3 from the 190 jars that had been prepared.
They were counted with their lids resting on the front
surface of a 5-inch Nal(Tl) detector, because the lids
were of more nearly uniform thickness than the jar
bottoms. The containers were tapped before counting
so that the samples occupied approximately equal volumes.
Each sample was measured over the energy range of 0.01
to 3.0 MeV, and one sample was counted 19 times. The
sample-to-sample inhomogeneity, defined as the differ-
ence between the range of the Y~ray-emission rates of
the 19 samples and the range of the repetitive measure-
ments on the single sample, was ±2.0% of the mean for
the 19 jars. These 19 samples were also transferred to
plastic bags and counted in a 4tty 8 x 8-inch Nal(Tl)-
detector system. The inhomogeneities, as reflected in
the Y-ray-emission rates, were determined over three
energy regions given in Table 2, and were each found to
be within approximately ±2% of the mean.
Other, smaller check samples with masses of 6 to
16 grams were prepared in plastic vials. Nineteen of
these subsamples were y-r&y counted in the 1-inch well
of a 5-inch Nal(Tl) crystal over the energy region of
0.01 to 3.0 MeV; one was counted 19 times. The net
contribution of inhomogeneities to the uncertainty of
the mean was ±2.2%.
A 5-inch NaX(Tl) crystal with a 2.5-inch, well was
used to make the final homogeneity check. All of the
100-gram samples were counted with the jars upside down
in this detector, and one jar was counted 19 times.
The count rates from each sample were compared in six
energy regions, and also over the entire y-ray spectrum.
A histogram showing this distribution for one of the
selected energy regions is given in Figure 2. Samples
giving outlying results'* were rejected. The net sample-
to-sample inhomogeneities in the y-ray-emission rates
of the accepted samples are listed in Table 2,
The prepared material was also checked for homo-
geneity of a- and 3-particle-emitting radionuclides
through radiochemical analyses carried out by several
laboratories using both 10- and 50-gram samples. The
inhomogeneities are estimated to be less than ±5% for
all the radionuclides measured except, possibly, up to
±10% for the transuranic elements.
Elemental and Mineralogical Characterization
A semiquantitative elemental analysis of mixed,
wet-sieved material that passed through a 325-mesh
sieve was performed with an Ebert-type (dc arc), 3.5-
meter emission spectrograph. The results are in Table
3. Samples of the 170-to-325-mesh and 80-to-170-mesh
fractions of the wet-sieved sediment were also analyzed
by this method. The results indicated, within the limi-
tations of the technique, that the finest material was
depleted in Al, Ba, Ca, Mg, Mn, Na, Pb, Sn, and Sr
compared to one or both of the coarser fractions. The
other elements that were detected in the finest material
were also found in the other two fractions, and in approxi-
mately equal concentrations.
Optical mineralogical analysis indicated that the
sediment consisted mostly of badly weathered mica and
iron-«ich clay, plus smaller amounts of quartz, diatoms,
highly altered feldspar and rutile. Particles down to
smaller than 1 ym were seen, but the most common sizes
were in a range of 5-to-25 ym. If the average particle
size had been larger, for example between 44 and 88 ym,
a known lower limit on particle size for the material
used in the standard would have been possible.
Radioactivity Measurements by NBS
Gamma-ray-emitting nuclides were measured nondes-
tructively by means of a 60 cm3 Ge(Li) detector within
an anticoincidence shield consisting of an annular Nal(Tl)
detector, with a Nal(Tl)-detector "plug", and by the
NBS 4Ty 8 x 8-inch NaI(Tl)-detector system. The a- and
fS-particle-emittlng nuclides were measured by radiochem-
ical techniques on weighed, randomly selected samples
of the material.
Certified, NBS-determined radioactivity values are
given in Table 4. Uncertified values that include other
NBS measurements (and measurements reported by other,
cooperating, laboratories) are in Table 5. Each value
in the Tables is a weighted mean of individual determi-
nations. The weights are the reciprocals of the vari-
ances of the individual values.
The stated uncertainty in each certified activity
is the linear sum of the random error, computed as three
times the standard error of the weighted mean, and the
estimated upper limit of conceivable systematic errors.
These uncertainties are for sample sizes of 10 grams or
greater. The results of the tests for homogeneity were
used in estimating the inhomogeneity component of the
systematic errors for y-ray-emitting nuclides. Also,
uncertainties in the following were estimated, where
applicable, for each value and linearly added: y-ray
photopeak areas, detector efficiencies, non-random
errors in the radiochemical procedures, sample position-
ing in the detectors and half-life values. The uncer-
tified NBS values were not certified because the syste-
matic errors and, or, random errors in the measurements
were unduly large.
Specific aspects of the radioactivity assay methods
for the sediment are considered below:
Gamma-Ray-Emitting Nuclides. A novel method was
used to develop an efficiency curve for the Ge(Li)-detec-
tor system for assaying y-ray-emitting nuclides in bulk
samples of the sediment, in that the NBS 4tty 8x8-
inch Nal(Tl)-detector system5* 6 was used to calibrate
suitably spiked samples of the sediment. A sample of
sediment, that had previously been slurried with a
solution of faoCo, dried, and assayed for Co activity
by sum-peak coincidence counting7, was used to obtain
points on the efficiency curve. This curve was then
extended over the range of y-ray energies contributed
by the 152Eu radioactivity in another spiked sample.
Confirmatory measurements on the curve were obtained
using 4 K in salts of potassium, and a sample of sedi-
ment spiked with 137Cs, using the same spiking method as
for 60Co. These samples were assayed by photopeak-effi-
ciency counting in the 8 x 8-inch crystals, using them
in the summing mode, with anticoincidence shielding.
Several known masses from 10 to 100 grams of each of
these spiked samples were y-ray counted in order to de-
termine the extent of self-absorption, which was found
to be negligible. The measured activities of in
19-5
-------
the potassium salts were close to the calculated values.
The procedure for spiking sediment samples was
as follows. Sufficient spiking solution - radioactivi-
ty and carrier in water - was added to the sediment in
an evaporating dish to yield a pudding-like consistency,
then the mixture was dried in air at 40 C. The dry
material was moved into a glove bag where it was lightly
ground with a mortar and pestle. Small portions were
shaken through a covered Buchner funnel, supported by a
ring stand, into an Erlenmeyer flask that was attached
to the funnel. Gamma-ray counting of equal, weighed
sub samples confirmed the radioactive homogeneity of the
spiked material.
The same 19 randomly selected jars of sediment
chosen for homogeneity testing were assayed with the
Ge(Li) detector system. Each was counted for at least
7 x 101* seconds, and its y-ray spectrum was stored on
magnetic tape. Both sample measurements and background
counts were made with no coincidence requirements and
also in the anticoincidence mode. Prior to analysis of
the photopeaks, the spectra from each set were added,
giving total counting times of greater than 1 x 106
seconds. The composite spectra are shown in Figures 3
and 4.
The locations and areas of the photopeaks in the
composite spectra,for samples and backgrounds, were
found independently by two computer programs that were
developed in the Radioactivity Section, one of which
uses the peak-fitting method of Hutchinson and Walker.8
Radionuclide id'ntifications were made, and net count
rates established. The activity of each nuclide was
calculated, using more than one Y-ray line whenever
feasible (Table 6).
Strontium-90 - Yttriua-90. The activity of
90Sr-90Y was determined from four 11-gram samples of the
prepared sediment. An outline of the radiochemical pro-
cedure is shown in Figure 5. The dissolution steps were
adapted from a procedure for the analysis of plutonium
in soil.9 The 6 particles of the separated 90Y were de-
tected with a 2ir, thin-window, gas-flow proportional
counter whose background was less than 0.5 count per
minute. Each sample was counted once or twice a day
for one to two weeks, and a weighted least-squares com-
puter program was used to fit the daily counting data
for each sample on a log-linear plot, as a function of
time.
Plutonium-239.240. The activity of 239+2
-------
In Figure 1, the gross y-ray count rates per gram
of dry-sieved sediment as a function of particle size
are nearly constant compared to those for the wet-
sieved sediment. These data, along with the particle
size distribution of wet-sieved sediment (Table 1), in-
dicate that the large clumps in the sediment were mostly
agglomerates of smaller particles, and these clumps were
not broken up effectively by dry sieving. For the wet-
sieved sediment, the count rate per gram is significantly
greater for the smallest size fraction than for the next
larger fraction. This is not surprising because the
surface-area-to-volume ratio of a particle increases as
its size decreases, and, in general, radioactivities are
absorbed more on or near the surfaces of soil and sedi-
ment particles than far inside them.*1 The trend of the
data is reversed, however, for still larger size frac-
tions. A reason for this apparent contradiction is that
the larger fractions of this sediment consist more and
more of organic material, which apparently has a higher
capacity for holding radioactivity than the mineral
components do.
Thus, the result for wet-sieved material in the
177-to-354-|j.m range apparently relates' to a very small
residue of organic particles of that size range. The
result in the same range of particle sizes for dry-
sieved material is for agglomerated clumps of small-
sized particles, so that the measured activity per gram
is essentially that of smaller particles.
The sediment SRM seems to be fulfilling its in-
tended purpose. In the first three months following
the announcement of its availability, we received
orders for nearly one-third of the total number of units
that were prepared. Purchasers include electric utility
companies, state and federal laboratories, universities
and research institutes, and commercial and industrial
laboratories.
There have been several requests for additional
amounts of this sediment standard. In response, the
remainder of the original material is being prepared
as before, but radionuclides in it may not be assayed.
Instead, it would be issued as a homogeneous sediment
standard. However, its radioactivity per gram should
be very nearly the same as that of SRM 4350.
There is also a need for sediment or soil stand-
dards with higher levels of activity than those in this
sediment, for nondestructive y-ray analysis. We plan
to produce a mineral matrix SRM that will be spiked with
several y-ray-emitting nuclides.
Acknowledgments
In addition to the four cooperating laboratories
(Table 7), the authors wish to thank the following
persons for their assistance in the development and prep-
aration of this environmental radioactivity standard:
Dr. B. Coursey, Mr. J. Harding, Dr. L. Lucas, and
Dr. F. Schima of the Radioactivity Section, NBS;
Dr. R. Conley, Georgia Kaolin Company, Mr. J. Dinnin,
U. S. Geological Survey, and Mr. C. Mabie, NBS, for
mineralogical analyses; Dr. D. Edgington, Argonne
National Laboratory, and the late Mr. J. Kies, Naval
Research Laboratory, for helpful discussions; Dr. J.
Harley, HASL, for arranging the freeze drying; the late
Mr. J. Weber, Jr., NBS, for spectrochemlcal analyses;
and Mr. W. Reed, of' the Office of Standard Reference.
Materials, NBS, for the provision of many facilities and
services.
References
1. Guy, H. P., "Laboratory theory and methods for
sediment analysis", Chap. C.1 of book 5, Laboratory
Analysis, in Techniques of Water-Resources Investi-
gations of the United States Geological Survey,
U.S.C1.S., Washington (1969).
2. ASTM, "Standard specification for wire-cloth sieves
for testing purposes", designation Ell-70, in 1974
Annual Book of ASTM Standards, American Society for
Testing and Materials, Philadelphia (1974).
3. Dixon, W. J., and F. J. Massey, Jr., Introduction
to Statistical Analysis, Table A-l, McGraw-Hill
New York (1957).
4. Natrella, M. G., Experimental Statistics, chap. 17,
NBS Handbook 91, National Bureau of Standards,
Washington (1963).
5. Hutchinson, J.M.R., J. L. Lantz, W. B. Mann, P. A.
Mullen, and R. H. Rodriguez-Pasques, "An anti-
coincidence shielded Nal(Ti) system at NBS", IEEE
Trans. Nucl. Sci., NS-19(1), 117 (1972).
6. Hutchinson, J.M.R., W. B. Mann and R. W. Perkins,
"Low-level radioactivity measurements", Nucl. Instr.
Meth. , U2, 305 (1973).
7. Hutchinson, J.M.R., W. B. Mann and P. A. Mullen,
"Sum-peak counting with two crystals", Nucl. Instr.
Meth., 112' 187 (1973).
8. Hutchinson, J.M.R., and D. H. Walker, "A simple and
accurate method of calibration by photopeak effi-
ciencies", Int. J. Appl. Radiat. Isotopes, 18^, 86
(1967).
9. Private communication, Jesse Meadows, Lawrence
Livermore Laboratory, Radiochemistry Division,
Livermore, California.
10. Livingston, H. D., D. R. Mann and V. T. Bowen,
"Analytical methods for transuranic elements in
seawater and marine sediments", proceedings of
Symposium on Analytical Methods in Oceanography,
Atlantic City, September 1974, to be published in
Advances in Chemistry series (American Chemical
Society, Washington).
11. Wong, K. M., "Radiochemical determination of pluto-
nium in sea water, sediments and marine organisms",
Anal. Chim. Acta, j>6, 355 (1971).
12. Latner, N., and C. G. Sanderson, "The HASL Ge(Li)-
Nal(TA) low level counting system", IEEE Trans.
Nucl. Sci., NS-19(1), 141 (1972).
13. Harley, J. H., ed., HASL Procedures Manual, USAEC
Report HASL-300, U. S. Atomic Energy Commission,
Washington (1972).
14. Sill, C. W., K. W. Puphal and F. D. Hindman,
"Simultaneous determination of alpha^emittlng
nuclides of radium through californium in soil",
Anal. Chem., 46(12), 1725 (1974).
15. Bojanowski, R., H. D. Livingston, D. L. Schneider
and D. R. Mann, "A procedure for analysis of
amerlclum in marine environmental samples", in
Reference Methods for Marine Radioactivity Studies
II. Ruthenium, Iodine. Silver and the Transuranic
Elements, IAEA Technical Report Series, Interna-
tional Atomic Energy Agency (1975).
4
19-5
-------
16. Healy, J. W., "Problems of plutonium measurement
in the environment", IEEE Trans. Nucl. Sci.,
NS-19(1) , 219 (1972).
17. Hawthorne, H. A., "Fission product cycles in an
agricultural system. I. Sample heterogeneity",
p. 711 in Radioactive Fallout from Nuclear Weapons
Tests, A. W. Klement, ed., U. S. Atomic Energy
Commission, Washington (1965).
18. Plato, P. A., "Distribution of cesium-137 and
naturally occurring radionuclides in sediments of
Lake Michigan", Radiation Data and Reports, 13(4) ,
181 (1972).
19. Harley, J. H., "Transuranium elements on land",
p. 1-104, U.S.E.R.D.A. Report HASL-291, U. S.
Energy Research and Development Administration,
Washington (1975).
20. U.S.A.E.C., Sampling and Analysis of Plutonium in
Soil, draft regulatory guide, Directorate of
Regulatory Standards, U. S. Atomic Energy Commis-
sion, Washington (March 1974).
21. Glenn, J. C., and R. 0. Van Atta, Relations Among
Radionuclide Content and Physical. Chemical, and
Mineral Characteristics of Columbia River Sediments,
open-file report, U. S. Geological Survey, Water
Resources Division, Portland, Oregon (1971).
Table 1.
Particle size distribution in wet-sieved
river sediment.
Table 2.
Inhomogeneities in v-ray emission rates of SRM 4350.
a. Nineteen randomly selected jars counted in the
4tty 8x8-inch Nal (T ^)-detector system.
Energy Region,
MeV
0.17 - 0.43
0.52 - 0.79
2.30 - 2.90
Principal
Radionuclide
1S3Eu
137Cs
bO
Co (sum peak)
Half of the
Difference in Range
(% of mean)
1.9
1.7
2.4
b. All jars counted in a 5-inch Nal(TX) detector
with a 2.5 inch well.
Energy Region. MeV
Half of the
Difference in Range
(% of mean)
.04 -
.160
1.8
.165 -
.285
1.9
.290 -
.410
2.2
.600 -
.870
2.4
.900 -
1.250
2.6
1.270 -
1.530
2.8
.01 -
3.00
1.9
U. S. Standard
Sieve Mesh No.
Sieve Opening,
urn
7
2830
14
1410
25
707
45
354
80
177
170
88
325
44
passing 325
< 44
Amount Remaining
(a) all organic material
(b) mostly organic material
(c) much organic material
0
0
< 0.1
(a)
< 0.1
(b)
0.2
(c)
4.3
7.7
as
Table 3.
sieved
sediment that nasaed
a No. 325
mesh sieve .
Element
Amount,
Amount,
Weight %
Element
Weight %
k&
0.03 - 0.3
Mn
0.01 - 0.1
Ba
0.001 - 0.01
Na
0.1 -1.0
Ca
0.03 - 0.3
Ni
0.001 - 0.01
Cr
0.003 - 0.03
Pb
0.003 - 0.03
Cu
0.01 - 0.1
Si
< 10.0
Fe
1.0 - 10.0
Sn
0.001 - 0.01
Ga
0.001 - 0.01
Sr
0.001 - 0.01
K
0.1 - 1.0
V
0.003 - 0.03
Mg
0.01 - 0.1
Zn
0.01 - 0.1
The sediment was also examined for the following elementss
Ag, As, Au, B, Be, Bi, Cd, Co, Ge, Hg, In, U, Mo, Nb,
P, Pt, Rb, Sb, Ti, W and Zr. They were not detected
within the limits of the method, 10 ppm by weight.
19-5
-------
Table 4.
Certified NBS activity values, in nuclear transformations per second per gram
on January 1. 1975. of radionuclides In SRM 4350.
Radionuclide
nts~Vl(1)
Uncertainty, %
Random Systematic Total
Methods Code
(2)
Mn
6 0
Co
6E>,
Zn
9°Sr+®°Y
Cs
162 Eu
1MEu
aaaAc
33 9+34 0,
Pu
5.4 X 10
2.1 X 10-3
1.4a X 10_1
1.3 X 10~3
1.0s X 10~2
l.Oo x 10-1
2.4 X 10-1
5.2 X 10~B
3.4 X 10~2
1.4 X 10'3
2.4
7.0
0.7
3.5
9.7
0.6
0.7
2.4
9.3
3.2
6.4
4.6
4.9
10.0
5.3
3.9
5.1
5.1
9.8
5.3
8.8
11.6
5.6
13.5
15.0
4.5
5.8
7.5
19.1
8.5
1, 2
1
1, 2
1
3a
1, 2
1
1
1
5b
(1) The gamma-ray abundances and half lives used in determining these values are from
Atomic Data and Nuclear Data Tables, Vol. 13, Nos. 2-3, February 1974,
(2) See Table 7 for explanation.
Table 5
Uncertified activity values. In nuclear transformations per second per gram on January 1. 1975.
of radionuclides in SRM 4350.
The names of the laboratories and the names of the analytical methods corresponding to the method
codes are given in Table 7.
Radio- _ _ Laboratory Method
nuclide nts lg 1 Initials Code
4.63 x 10_1
HASL
1(1)
6tMn
2.3 x 10"3
HASL
1
e6Fe
1.6
UH
3d
60Co
1.36 x 10_l
HASL
1
6BZn
1.2 x 10"a
HASL
1
90sr+9°y
1.1 X 10_a
HASL
6c
l8£Sb
3.6 X Hf3
NBS
1
137Cs
8.95 x 10"a
HASL
1
9.41 X 10"8
HASL
6c
lBaEu
1,60 X 10_1
HASL
1
6
19-5
-------
Table 5 - (continued)
Radio- _ _ Laboratory Method
nuclide nts 1g 1 Initials Code
164Eu
4.37 X 10_a
HASL
1
16faEu
1.4 X 10~3
NBS
1
1.4 x 10_a
NBS
1
aiapb
6 X 10"3
NBS
1
aiaBi
5 x 10~a
NBS
1
31*Pb
4.i X 10_a
NBS
1
aitBi
3.4 x 10"2
NBS
1
S36Ra
3.1 X 10"3
HASL
1<2)
aa8Th
3.95 X 10"a
HSL
4b
a3°Th
3.66 X 10"a
HSL
4b
aaaTh
3.1 X 10~a
HASL
i(2)
3.78 X 10-a
HSL
4b
aalPa
1.75 x 10~3
HSL
4b
a34u
4.9e X 10"a
HSL
4b
1.85 X 10"3
HSL
4b
33 au
4.2a X 10-3
HSL
4b
338Pu
7.6 X 10~5
HSL
4b
5.8 x 10-E
WHOI
5b
33i>-h34 0pu
1.4 X 10"3
HASL
3b
1.2v x 10~3
HSL
4b
1.45 X 10~3
WHOI
5b
*41Am
3.1 x 10~4
HSL
4b
3.19 X 10"4
WHOI
5b
a44Cm
< 8 X 10~6
HSL
4b
1.6 X io"B
WHOI
5b
(1) The gamma-ray abundances and half lives used by HASL are from the HASL Procedures
Manual, USAEC Report HA5L~300, 1973, and are, in some cases, different from values
taken from the Atomic Data and Nuclear Data Tables to calculate the certified
activities.
(2) Values are based on measurements made of the gamma-ray-emiseion rates of two sub-
sequent members in its radioactive decay series.
Table 6.
Energies of V-ray photopeaks used by NBS for the assay of radionuclides in SRM 4350.
Gamma rav energies in keV^
Nuclide
Gamma ray energies in keV^
Nuclide
40k
1460.85
154Eu
54Mn
834.83
155Eu
60Co
1173.21, 1332.48
208W
65Zn
1115.52
212Pb
125Sb
427.9, 463.4
212Bi
137Cs
661.64
214Pb
152Eu
344.31, 443.98, 778.87, 867.33,
21*Bi
964.01, 1212.94, 1408.02
228Ac
873.16, 996.29, 1004.75, 1274.49
86.54, 105.30
277.35, 583.14, 860.37
238.63
1620.56
295.2, 352.0
609.3, 1764
209.4, 338.4, 911.1
(1) Values taken from Atomic Data and Nuclear Data Tables, Vol. 13, Nos. 2-3, February 1974.
7
19-5
-------
Table 7.
Table 8.
Participating laboratories and analytical methods
employed in the assays of radionuclides in SRM 4350.
Comparison of concentrations of some long-lived radio-
nuclides in SRM 4350 with those reported for radioactive
fallout in other sediments and soils.
Laboratory
HASL Health and Safety Laboratory
Energy Research and Development Administration
New York, New York
(Mr. G. A. Welford) [l, 3b, 6c]
HSL Health Services Laboratory
Energy Research and Development Administration
Idaho Falls, Idaho
(Mr. C. W. Sill) [4b]
NBS Radioactivity Section
National Bureau of Standards
Washington, D. C. [1, 2, 3a, 5b]
UH Department of Radiochemistry
University of Helsinki
Helsinki, Finland
(Dr. J. K. Miettinen and Dr. T. Jaakkola) [5d]
WHOI Woods Hole Oceanographic Institution
Woods Hole, Massachusetts
(Dr. V. T. Bowen and Dr. H. D. Livingston) [5b]
Analytical Methods
1 Ge(Li) y-ray detector
2 Nal(Tl) Y-ray detector
3 Acid dissolution
4 Fluoride-pyrosulfate fusion
5 Acid leaching
6 Sodium carbonate fusion
b Alpha-particle spectrometry with surface-
barrier detector
c Plastic-phosphor beta-particle scintillation
detector
d Liquid scintillation counter
Nuclide
90Sr+90Y
pcig
-1
137
Cs
238
Pu
239+240
Pu
0.27
-0.4
2.7
1.4
0.0020
-0.0004
0.0015
0.038
-0.02
0.023
Sample Reference
SRM 4350
Utah soil 17
SRM 4350
Lake Michigan sediments 18
SRM 4350
Soil, U. S. average 19
Marine sediment,
Buzzards Bay, Mass. 10
SRM 4350
Soil, U. S. average 19
Marine sediment,
Buzzards Bay, Mass. 10
241
Am
140
0.0086 SRM 4350
¦0.006 Soil, U. S. average
0.0059 Marine sediment,
Buzzards Bay, Mass.
SIEVE MESH NUMBER
325 170 80 45
19
10
120
2
<
ce
to
a Thin-window beta-particle proportional counter _ |qq
LlI
I—
r>
z
CO
H
z
=>
O
O
60
< 20
(T)
T~
O WET SIEVED
A DRY SIEVED
T~
1
rr 40
¦f
OL-4.
Figure 1.
354
707
44 88 177
PARTICLE SIZE, /xm
Variation of Y-ray emission rates per gram
of sediment with particle size ranges.
8
19-5
-------
r~r
44 -
40
36
32
28
(/)
„„
-------
PRACTICAL PROBLEMS OF MONITORING FOR PLUTONIUM
IN THE ENVIRONMENT
By: H. C. Woodsum
Mestinghouse Environmental Systems Department
Pittsburgh, Pennsylvania
Abstract
Preoperational survey data obtained at the Westing-
house Recycle Fuels Plant (RFP) site near Anderson,
South Carolina, have shown significant variability
in the measurement of plutonium concentration in
fish. Environmental monitoring for Pu in soil, pre-
cipitation, terrestrial biota, ground and surface
water and other aquatic biota conducted in the RFP
preoperational survey also showed some variability,
but the largest discrepancies were observed in the
aquatic media. Based on this experience, it is con-
cluded that better sampling and analysis standards
and quality control guidelines for plutonium environ-
mental monitoring should be established.
Introduction
Based on Westinghouse experience, present technology of
monitoring plutonium effluents released to the environ-
ment are adequate in assessing the plutonium (Pu) reach-
ing the environment, and these effluent releases re-
sulting from normal Pu fuel operations are well below
established guidelines. However, because of the very
low level of these releases, difficulties have been en-
countered when measurements have been made to determine
the Pu levels in environmental samples (i.e., air,
water, vegetation and soils). At the present time there
are no well defined standards for monitoring Pu in the
environment. Thus, considerable variability in the
measurements may result. When Westinghouse Environ-
mental Systems Department (WESD) initially began work
on the Environmental Report for the W Recycle Fuels
Plant (RFP) whose construction is proposed on a site
near Anderson, South Carolina, a preoperational moni-
toring program was planned. Assistance in planning the
preoperational phase as well as the operational phase
of the monitoring program, as documented in the RFP en-
vironmental reportj was obtained through conversation
and correspondence with various governmental agency
personnel who had experience in this field. This moni-
toring program consists of collecting environmental
samples of air particulates, water, soil, vegetation
and certain biota (such as fish). Samples are analyzed
for isotopic Pu and gross a and B radiation on a peri-
odic basis. The first survey was performed in February
through June 1973. After some of the preliminary pre-
operational survey data for the RFP site had been
gathered and analyzed, it was obvious that present
techniques leave much to be desired in terms of ob-
taining accuracy and consistency of the data at these
environmental levels and conditions.
Although, the measured concentrations are well below
maximum permissible levels for continuous ingestion and
in many cases below detectable limits, variability was
observed and the data were not consistently reproduc-
ible. A major difficulty has to do with differences
between what can and has been done 1n the laboratory
and what is reasonable in routine environmental anal-
ysis. With careful control of environmental sampling
conditions as well as of radiochemical techniques,
fairly consistent data can and have been obtained at
environmental levels; that is, at concentration levels
encountered in environmental samples. In the absence
of recognized standards, the techniques used and the
results obtained vary from laboratory to laboratory and
from time to time at the same laboratory due to varia-
tions in sampling preparation and radiochemical analy-
sis techniques, differences in instrumentation and in-
strument use, and to insufficient quality control.
Another practical constraint is that environmental sam-
ples in the "real world" are contaminated by other ele-
ments which provide interference. Also the plutonium
itself may be very nonhomogeneously dispersed in the
environment media. These factors can lead to large var-
iations in measured plutonium levels even in samples
taken from the same locale.
The relatively high cost of isotopic analysis (minimum
of about $50 per sample) also provides practical limi-
tations on the ability to obtain good spatial distri-
butions. On the positive side, some steps are being
taken to improve the accuracy and consistency of moni-
toring plutonium in the environment. For example, a
monitoring guide, similar to that for Light Water Re-
actor plants, is being prepared for fuel processing and
reprocessing plants and regulatory standards for cer-
tain types of analyses have been prepared (for example
2
in soil) or are in preparation.
Specific examples of some of the WESD experience con-
cerning difficulties encountered Tn actual monitoring
operations are discussed further below.
Environmental Monitoring of Pu in Air Particulates
Plutonium has been and is being monitored in particu-
late form in air at "environmental levels" both in the
USA and elsewhere in the world utilizing an air blower
and filter to collect the particulates. Since pluton-
ium is a man-made element, the current airborne levels
are primarily a result of atmospheric nuclear weapons
testing. Because of recent moratoriums on atmospheric
weapons testing by the USA and Russia, the "background"
air concentration has been falling off since peaking in
the 1961 to 1963 time period. Current ambient levels
in the eastern USA are about 3 x 10 ^ to 1 x 10 pCi/m
for Pu-239 or about 1/2000 to 1/600 of the existing MPC
for inhalation of soluble Pu-239. Pu-238 levels are of
the order of 5 to 10 times lower, or ^ to^ooO of MPC
for inhalation of soluble Pu-238. It is difficult to
monitor plutonium concentrations in this range because
the levels are very low. The typical "minimum" detect-
able Pu alpha activity is 0.02 pCi per sample which
corresponds, for a 50 percent detection efficiency, to
0.005 counts per minute. Thus, a 1000 minute or M7
hour detection interval is required to obtain 5 counts.
Allowing for the Inherent alpha background requires
even longer counting times to obtain good statistics
(95 percent confidence) above background. Typical back-
ground levels for plutonium analysis can be as high as
0.04 counts per minute. Thus, practical considerations
often increase this minimum detectable level to 0.1 pCi
per sample. Using the upper limit estimate of the ex-
3
pected background, it is seen that at least 1000 m
1
19-6
-------
o
(35,000 ft ) of air must be sampled to obtain meaning-
ful data. To obtain significant levels of the Pu-238
background, this air volume sample would have to be in-
creased by a factor of 5 to 10.
Pu air concentration levels around a plutonium fuel
fabrication plant are in the background range since con-
centrations at the stack release point are of the order
-3 3 • .
of 3 x 10 pCi/m and an additional dilution factor of
100 to 10,000 is imposed by atmospheric dispersion to
the site boundary. To obtain data at the 95 percent
3
confidence level, 10,000 m of air should be sampled.
If a weekly sample is desired, this would require an
air sampler flow rate of about 35 CFM, a relatively high
volume air flow rate. Thus, Pu isotopic analyses are
performed quarterly or monthly with resulting air sam-
ple flow rates of 2 or 9 CFM, respectively.
Although this collection technique results in a rela-
tively long term averaging process, the Pu containing
particulates are relatively uniformly dispersed on the
filter paper collector and an annual average dose is
required. Thus, sample collection is not the major
cause of uncertainty in the analysis of dose to humans.
The major cause of variability for this analysis is pro-
duced by utilization of different radiochemistry tech-
niques. There are several basic techniques noted in
A 5
the literature. * Examples of the basic techniques
A
are: the HC1 acid leach method of Talvitie, the LASL
3
method proposed by the AEC in regulatory guide 4.5 and
5
the complete fusion method proposed by Sill at NRTS.
Modifications or "improvements" to these techniques
have developed such as the "modified HASL-LASL method.6"
In all of the above techniques a tracer (usually Pu-236)
of known quantity is added and the dissolved sample is
plated out for alpha spectroscopy counting. The tracer
itself may introduce contamination. Thus, it is best
to add the tracer in about the same concentration as
the sample being analyzed. Then the counting statis-
tics on the tracer alpha energy and the Pu-238 and
Pu-239 alpha energies are comparable and no "interfer-
ence" from trace induced contamination should result.
However, this requires that one know the concentration
of the unknown beforehand. Another problem with the
tracer techniques is that the tracer may not be pure,
e.g., it may contain trace impurities of Pu-238 and
Pu-239, such that 1t produces an undesired "background"
which may be above the level being detected. Most of
the above problems can be avoided by utilizing Pu-242
as a tracer since Pu-242 has a lower alpha energy than
the Pu Isotopes being sought and the Pu-242 tracer sam-
ples usually contain a very low content of other Pu
Isotopes. Thus, little interference should result.
Known samples must be Included with the unknown samples
to assure that the chemistry technique 1s being util-
ized consistently since a laboratory may add its own
modifications to any of the basic techniques. The pro-
ponents of each method argue that their method 1s "best"
but the results are often different. Consistency of
radiochemical anaylsis can be improved by contractually
requiring that the same chemical analysis techniques be
utilized for all samples.
Another problem with adding the tracer is that the
tracer may not be in the same chemical form as the un-
known arr* thus a high fraction of recovery of the trac-
er may not ^-"vintee the same high fraction of recovery
for the unknown.
The nuclear analysis procedure also has inherent prob-
lems. For example, the alpha spectrometry system re-
guires long counting times with low background and good
(50-100 kv) a energy discrimination.
In spite of the many limitations of the current state-
of-the-art monitoring of Pu in air, this parameter is
believed to be one of the more accurate of the environ-
mental type measurements. These data have been more
self consistent than some of the other types of Pu en-
vironmental monitoring measurements. However, this does
not guarantee that these measurements are more accurate
in any absolute sense, because they still suffer the
uncertainties related to the accuracy of the radiochem-
ical techniques utilized.
Environmental Monitoring of Pu in Soil
Monitoring for plutonium in soil suffers all of the
problems noted above for air particulates plus a few
additional ones. Since soil is of varying composition
and the particles, in general, are not dispersed uni-
formly, it is quite likely that one analysis will vary
o
from another. A regulatory guide has been prepared by
the AEC for monitoring this medium and its common use
should help to eliminate some of the variables. How-
ever, there are still some difficulties with soil moni-
toring for Pu. The guide proposes either of two meth-
ods for obtaining soil samples: the "core method" and
the "trench method". Both of these techniques might
have to be modified for use at sites in the northeast-
ern part of the U.S. because the required soil samples
are not readily available. For example, the guide sug-
gests sampling at a minimum of 13 locations with the
closest sampling site being located about one mile from
the plant. Furthermore, the guide recommends that 10
cores, 3-1/2 inches in diameter and 12 inches deep from
an area of about one square foot be sampled which would
result in collection of a total of 40 to 80 lbs of soil
from each sampling location. All of this soil is taken
to obtain a 10 gram sample for analysis. Even for rela-
tively uninhabited areas, such as surrounding the RFP
site in Anderson, South Carolina it may be very diffi-
cult to remove 40 to 80 lbs of soil for each sampling
site on each collection period at 13 locations entirely
beyond the site property boundary. This amounts to a
total of about 1,000 lbs of soil removed for each sam-
pling period.
The radiochemistry techniques available for soil analy-
sis are the same as for air particulates. The method
selected in regulatory guide 4.5 is similar to the one
proposed by Talvitie4 and 1s the one currently in use
at the Los Alamos Scientific Laboratory (LASL).
Although the development of procedures like Regulatory
Guide 4.5 is certainly a step 1n the right direction,
such standards will not solve all of the problems atten-
dant with obtaining consistent, reproducible and accur-
ate Pu concentration data in soil. They may, in fact,
introduce new problems of a practical nature (such as
how to obtain 1,000 lbs of soil at each sampling period
without significantly altering the environment being
sampled). Although Regulatory Guide 4.5 provides a
standard procedure for radiochemical analysis, use of a
standard control sample is also recommended. The re-
cently Issued NBS Sediment Sample #4350 could be used
for this purpose.
Trace concentrations of Pu can also be determined by
mass spectrometry. In the past this method has suffer-
ed from loss of detection efficiency of at least one or
two orders of magnitude for environmental plutonium sam-
2
19-6
-------
g
pies as compared to pure plutonium mixtures. However,
in a recent paper by Storms et al^ mass spectrometric
method for determination of plutonium in soils is des-
cribed which is inherently capable of greater sensitiv-
ity than alpha spectrometry. It is reported that using
this technique, Pu isotopic compositions are success-
fully determined for concentrations as low as 2 x lO"1^
g Pu per gram soil based on 10 grains of soil or other
environmental sample.
Better instrumentation for in-place survey of the soil
O
would also be useful (such as an improved Fidler
Q
probe). Until such time as improved soil monitoring
techniques are developed, it is recommended that soil
sampling be considered a secondary monitoring method.
The primary methods for monitoring airborne deposition
should be precipitation and vegetation monitoring.
Thus, soil sampling should be performed to obtain semi-
quantitative data on contamination of soil to verify
that buildup of Pu remains small.
Environmental Monitoring of Pu in Fallout
(Precipitation) ~~~
Collection and isotopic analysis of rainwater provides
the primary method for monitoring plutonium deposition
from airborne releases. The principal problems associ-
ated with collecting the samples result from the sample
freezing (in a cold climate) overflowing (in a wet cli-
mate), or evaporating (in a dry climate). Once the
sample is obtained there does not appear to be any dif-
ficulty with this monitoring technique except for the
problems of obtaining consistency in radiochemical
analysis results for water samples. The difficulties
associated with radiochemical analysis of these liquid
samples is the same as that for ground and surface
water which will be discussed next.
Environmental Monitoring of Pu in Ground and Surface
Waters
Sampling of plutonium in either ground or surface water
has some of the same inherent problems as in soil.
First, there is no standard procedure for procuring the
water sample from the environment. The usual procedure
is to take a "grab" sample and then transfer the con-
tents to a plastic container. Secondly, there are no
uniform radiochemical analysis procedures for analyzing
for plutonium once the sample is taken. An example of
the type of variability that can occur as the result of
the nonuniform sampling and analysis procedures is
given by the data obtained in the Cheswick, Pa. Pluto-
nium Fuel Development Laboratory Monitoring Program10
with samples taken from the Allegheny River upstream
of plant discharges. The results of an analysis on a
split sample taken on March 2, 1974 by two different
laboratories as given in Table I shows a difference by
about a factor of 10 in reported Pu-239 concentration
and undetectable Pu-238. Analysis of a second grab
sample taken on April 27, 1974 indicated more Pu-238
than Pu-239. Normally, background fallout activity
has a ratio for Pu-238/Pu-239 of 0.1. Such variations
in concentrations cannot be explained by normal varia-
tions in "background" Pu concentrations or laboratory
techniques.
To better understand the above water data, several
check analyses were performed with the assistance of
the Westinghouse R&D Laboratory. A sample of known
Pu-239 concentration was prepared by the R&D Lab and a
duplicate amount of this sample was sent to each of
two laboratories. The first lab indicated the control
sample contained 2.07 + 0.11 pCi/1 of Pu-239 and <0.02
pCi/1 of Pu-238. The second laboratory reported the
activity in the control sample to contain 1.05 + 0.05
pCi/1 of Pu-239 and <0.05 pCi/1 of Pu-238. The control
sample activity as prepared by the R&D laboratory was
calculated to contain 1.23 pCi/1 of Pu-239 and no de-
tectable Pu-238. Although the first laboratory showed
some discrepancy from the calculated value for the con-
trol sample, this discrepancy is not nearly large
enough to explain the variation in the river water sam-
ples as noted above.
Some further investigation of the above discrepancies
was made. In early 1974 ASTM subcommittee D 19.04 be-
gan interlaboratory analyses of a method of measuring
Pu in water. From these analyses it was determined that
Pu was being lost by plate out on the walls of the sam-
ple container or by some other mechanism. By mid-1974
everything but plate out was eliminated and a definitive
test to measure sample container plate out was devel-
oped.11
In this analysis, PWR coolant water samples were "spiked"
with 10 dpm and 50 dpm samples of Pu-239 and the spiked
samples and a blank sample were analyzed by two labora-
tories. The results of these studies11 showed that (1)
between 13 and 50 percent of the Pu-239 remained on the
container walls after a single water rinse and (2) a
single nitric-hydrofluoric acid rinse removed over 90
percent of the activity in the container wall.
In another case where variability in water sample re-
sults was observed, differences in counting techniques
were determined to be the cause. For two samples, a
recount of the activity level in the water was reported
to be some 50 to 100 percent different from the original
data in spite of the fact that the original data were
reported with a 95 percent confidence level and a stan-
dard deviation of less than 10 percent. It was deter-
mined that two different sets of energy channels on the
multi-channel analyzer were used for the two different
analyses.
To summarize the data on Pu in water at environmental
levels, it is noted that Westinghouse experience has
been that variability of up to a factor of 10 has been
observed between samples which would be expected to be
similar. A major part of this variability is attributed
to the nonsystematic nature of the environmental sam-
pling methods used, but other sources of these differ-
ences are attributed to plate out on the container walls,
nonsystematic counting methods, differences in radio-
chemical analysis techniques and possible laboratory
contamination of samples.
Environmental Monitoring of Pu in Fish
12 13
Fish in both fresh water and salt water have been
found to concentrate Pu by factors ranging to about 4
over that of their native water environment. Thus,
fish would be expected to be a more sensitive indicator
of Pu than water. This hypothesis appears to be corrob-
orated by evidence of Pu found in fish analyzed for the
RFP preoperational program.1 However, although it ap-
pears that Pu has been found in some samples of fish
analyzed on this program, the consistency of the data is
so poor that no definitive background levels can be es-
tablished. Also, although very little data is reported
on Pu in fresh water fish, the data obtained by the
Woods Hole Oceanographic Institution in the Great
12
Lakes showed that Pu specific activity levels were
3
19-6
-------
approximately 1,000 times lower than those obtained in
the Anderson, South Carolina area Preoperational Sur-
veys.^ Thus, the Anderson, South Carolina data analyzed
by commercial laboratories appear to be inconsistent
with the expected background.
Some effort has also been expended in attempting to un-
derstand the Anderson, South Carolina area fish data.
After an initial preoperational survey taken in February
1973 identified positive indication of Pu-239 in all
four fish samples taken (see Table II),1 with specific
activities ranging from 0.04 up to 0.12 pCi/g (dry),
further surveys were made to attempt to establish exist-
ing background levels. A second sampling performed in
May of 1973 (Table III) showed positive indications in
only 2 out of 7 analyses and one of these was on only
the bone component of catfish.14 Furthermore a "split"
sample of these fish was sent to the EPA Easter/i Envi-
ronmental Radiation Laboratory and the EPA analysis re-
sults showed the fish to contain no measurable Pu
activity in spite of the fact that their minimum detect-
able levels were 6 to 30 times lower than commercial
laboratory results. The minimum detectable levels for
the EPA laboratory were 6-7 x 10"5 pCi/g (wet) for fish
flesh and 3.5 x 10"4 pCi/g for fish bone. These data
are shown in Table IV. To better understand this appar-
ent discrepancy, a second set of fish samples were ob-
tained from Lake Hartwell near Anderson, South Carolina
and these three samples were ashed and the ashed samples
were divided into three equal parts and were sent to the
EPA Eastern Lab and to two other laboratories. This
comparison, given in Table V, showed all Pu analyses to
be below detectable levels except for one laboratory
analysis which showed 0.22 pCi/g ash. This analysis
was repeated on a second 10 g aliquot of the 33 grams
of ash sample and was also found to be below detectable
levels of 0.002 pCi/g ash for Pu-238 and 0.003 pCi/g
ash for Pu-239.
Conclusions
pies and thus to reduce the causes of extreme variabil-
ity previously observed.
References
1. Recycle Fuels Plant Environmental Report, Nuclear
Fuel Division Westinghouse Electric Corporation,
July 1973, Docket No. 70-1432.
2. "Preparation of Environmental Reports for Nuclear
Fuel Fabrication Plants", Draft, December 31, 1973.
USAEC Regulatory Division.
3. Regulatory Guide 4.5, Directorate of Regulatory
Standards, "Measurements of Radionuclides in the
Environment, Sampling and Analysis of Plutonium in
Soil," May 1974.
4. Talvitie, N.A., "Radiochemical Determination of
Plutonium in Environmental and Biological Samples
by Ion Exchange," Analytical Chemistry, Vol. 43,
November 1971, p. 1827.
5. Sill, Claude W., Puphal, K. W,, and Hindman, F. D.,
Health Services Laboratory, USAEC Idaho Falls,
Idaho 83401, "Simultaneous Determination of Alpha-
Emitting Nuclides of Radium through Californium in
Soil," prepublication copy transmitted by letter
January 7, 1974.
6. Rourke, F. M., and Mewherter, J. L., "Effects of
Chemical Processing on the Thermal Ionization
Efficiency of Plutonium," Annual Conference on Mass
Spectrometry and Allied Topics, pp. 192-193, June
4-9, 1974.
7. Storms, H. A., Hunter, F. F, and Carlson, D. C.,
"Determination of Plutonium in Soils by Mass Spec-
trometry," April 2, 1976, presented at Second An-
nual Conference Nuclear Methods In Environmental
Research, July 29, 30 and 31, 1974, University of
Missouri, Columbia.
The latter sample results noted above were Intended to
help resolve differences noted in previous results.
However, the opposite effect has resulted since these
data further cloud the Issue and at this time it does
not appear possible to say unequivically whether or not
there is plutonium in fish located In the near vicinity g.
of the proposed RFP site. It Is possible that part of
the difficulty In establishing quantitative background
levels of Pu 1n fish could be due to sporadic and selec-
tive uptake and/or nonuniform distribution of Pu in the
fish. However, 1t could also be due to the other d1f- 10.
flcultles which were previously noted for other type
media monitoring.
11.
Although present effluent monitoring technology is as-
sessing the releases of Pu to the environment, and in
cases WESD has evaluated to date, the control of re-
leases resulting from normal Pu fuel operations is well 12
below the established guidelines, the plutonium environ-
mental monitoring technology needs significant Improve-
ment to assure that background levels 1n the environment
can be accurately determined. This Improvement In tech-
nology will require establishment of better sampling and
radiochemical analysis technique standards and quality 13,
control guidelines for Pu environmental monitoring and
some development of the Instrumentation technology.
Such improvements will require research and development.
Two factors are Involved 1n the accurate determination
of Isotoplc Pu concentration in the environment: 1)
the performance of the analytical laboratory (precision, 14.
accuracy, etc.). and 2) the Inherent variability 1n sam-
ples. By careful attention to technique and proper
quality control, 1t should be possible to measure and
control laboratory analyses of Pu in environmental sam-
LRL-2039, "Plutonium: A Review of Measurement Tech-
niques for Environmental Monitoring," Budnltz, Robert
J., November 1973, presented at the IEEE Nuclear
Science Symposium, San Francisco, California, Novem-
ber 14-16, 1973.
Tinney, J. F., Koch, J. J., and Schmidt, C. T.,
"Plutonium Survey with an X-Ray Sensitive Detector,"
Report UCRL 071362, Lawrence Llvermore Laboratory,
1969.
Westinghouse Cheswlck Site Fuel Development Labora-
tories, Environmental Report, September 1974, Revi-
sion 1, February 1975, Revision 2, June 1975).
Nuclear Fuel Division, Advanced Reactors Division,
Private communication, J. A. Corbett, Manager Analyt-
ical Laboratories, W Advanced Reactors Division, to
Wayne Bickerstaff, Industrial Hygiene and Safety
Department, W Cheswick Site, May 7, 1975.
C00-3568-3, General Summary of Progress Report, 1972-
1973, "Plutonium Concentration Along Fresh Water Food
Chains of the Great Lakes, U.S.A.," Bowen, V. T.,
Noskin, V. E,, Woods Hole Oceanographlc Institution,
Woods Hole, Massachusetts, 02543.
Noskin, V. E., and Bowen, V. E., "Concentrations and
Distributions of Long-L1ved Fallout Radionuclides in
Open Ocean Sediments," Reprinted from "Radioactive
Contamination of the Marine Environment," pp. 671-686,
IAEA, Vienna 1973.
WESD-NP-443 (Internal Memo H. C. Woodsum to W. S.
Gelger), "Summary of Data on Pu Concentration 1n
F1sh Samples," March 8, 1974.
19-6
-------
TABLE I
COMPARISON OF WATER SAMPLE ANALYSES FOR ISOTOPIC PLUTONIUM
PLUTOSIU* ISOTOPIC AM Ciiii c. kKt 3 ACTIVITY LEVELS
FOR SAMPLES OBTAItiEO FROM ftfi» C.Ii S7*.iA«S Of! KAY 13, 1973
Laboratory 1
Laboratory 2
Laboratory 2
Control Sample
Laboratory 1
Laboratory 2
Oate of Sample
Procurement
3/2/74
3/2/74
4/27/74
3/2/74
3/2/74
3/2/74
Date of Sample
Processing
Analysis Results, dCi/1
Pu-233 ZSZ22T
RIVER WATER SAMPLE*
4/9/74
3/25/74
6/27/74
<0.02 <0.02
<0.06 0.22+0.05
0.25 + 0.15** <0.14**
"STANDARD" MATER SAMPLE
3/2/74 0 1.23 + 0.06+
3/25/74 <0.02 2.07+0.11
3/5/74 <0.05 1.05+0.05
* Samples obtained upstream of PFDL Cheswlck site discharges, near Mew
Kensington Bridge.
** Minimum detectable levels for Pu-228 and Pu-239 1n this case were increased
to 0,14 pC1/l since less than the desired amount (one liter) was available.
t Calculates value.
Location
Wet/Ory Vrt. (q!
Grot-. :
; Gr&si
i Beta
Pu-S33
Mouth of Weens
Creek
Minnows
84.7/19
(175.8/37.6)*
<0.5
10.8
+ 0.98
<0.005
0.022 + 0.004
Gentrostee Creek
Jus*, below en-
trance >'aeas
Creek
Winnows
25.7/ 4.5
( 53.5/ 9.6)*
<0.5
6.6
+ 0.64
<0.005
<0.005
Weens Creek
near East Creek
Minnows
K.0/21.0
(188.0/41.0)*
<0.5
12.2
~ 1.1
<0.035
<0.G05
Gencrcstae Creek
Just balo*
Waens Creek
Catfish
(flesh)
76.5/16.0
(196.5/44.2)*
<0.6
11.2
~ 1.0
<0.005
<0.005
Ganerostee Creek
Juit bfelsw
Wee.ti Creek
Catfish
(bone)
25.7/ 4.8
( 26.6/ 8.2)*
<2.8
27.1
~ 3.1
<0.005
0,007 + 0.004"
Ser.erostee Creak
near bridge
Catfish
(flesh)
246,6/ 46.1
(CS7.5/122.2)*
<0.5
8.8
~ 0.93
<0.005
<0.005
Sencrostee Creek
near bridge
Catfish
(bone)
82.4/15.7
(164.8/31.4)*
-<2.9
35.0
i 3.2
<0.005
0,012 ~ 0.003
' Values 1n pireruhosas are wit/dry weUjht of original simple. Values indicated above are those
analyzed by Corr.erciul U& 1. A najor pjrt cf the ror.a1n.der (difference betwaan values 1n paren-
theses and values above) were sen: to EPA, Montgomery, Alabjnia for 4 split analysis. (Sea Table IV.)
' The minimum decectable level (HDL) is 0.005 pC1/g (dry). The statistical uneertalr.ty u such that
In this case within the 96 percent confidence Units 1t cannot be deterolned that the KOI is
exceeded. Thus, It cannot be determined that a measurable quantity was observed.
PLUTONIUM AIIO CROSS o AND 8 ACTIVITY LEVELS FOR PISH
SAMPLES OBTAINED FROM RFP SITE STREAMS ON FEBRUARY 23, 1973
Sarele
5a3te
Weeeis Creek
below East Creek
Veens Creek
above East Creek
Qenerostee Cn«k
below Weens
fiererostee Creak
below Weens
HoT Wet/Dry Wt.
75 23.0/ 4.9 g
76 40.5/11.0 9
94 14.0/ 3.5 g
95 9.0/ 2.0 9
Gross Alpha
<0.50
0.54 + 0.40
<0.50
<0.50
Radioactivity (pC1/q dry)*
feross Beta Pu-Z38
1.01+ 0.58** <0.01
1.57 + 0.61
0.94 + 0.57
<0.50
<0.01
<0.01
<0.01
Eu-233
0.04 + 0.02
0.06 ~ 0.02
0.10 + 0.04
0.12 +0.03
t humans
TABLE IV
COMPARISON OF EPA AND COMMERCIAL LABORATORY FISH ANALYSIS RESULTS^14'
ON SPLIT SAMPLES OBTAINED ON MAY 13. 1973
Sample Type
Catfish (flesh)
Catfish (bone)
Small Fry
Original
Weight*
wet/dry (q)
854/166.0
201/ 39.6
417/ 88.2
* Weight before splitting.
* Whole body samples (including bone) of species of fish not generally considered edible by f
were analyzed. Sample sizes were less than recommended minimum mass Units. Thus, these pre-
liminary results may not be typical of background concentrations 1r. edible portions of fish proposed
for t*e preoperational and operation** monitoring programs. (See Table 6.1-3 in Section 6.1-5 of
Reference 1).
** Uncertainties are for two t1i*c> t^ .~jndsrd deviation based on counting statistics.
Radioactivity in 10"^ pCi/g (wetl
Commercial Lab 1
Pu-238 Pu-239
<1.0 <1.0
<1.0 2.1+0.8
<1.1 2.1+0.4
EPA
Pu-238
Pu-239
<0.06
<0.35
COMPARISON Or PREOPERATIONAL RADIOCHEMICAL
AIMiWK «:Sv';T£ FOR LAKE HARTWCLL FISH ASH SAHPLfS
Cars
Laboratory
LAB 1
LAI 2
EPA
Ash weight.
grams
23.7
10
10
Gross a
<1
<6
<2
Cross 8
46 + 4
145 + 7
57 + 2
Pu-238
<0.005
<0.01 *•
<0.002
Pu-239
<0.005
<0.02**
<0.002
Radioactivity (oCi/o ash)*
LAB ^ mnr
33.28
<1
<0.005
<0.006
10
<5
125 *6
0.01 ~ 0.004**
0.22 + 0.02**
10
<2
40+2
<0.002
<0.002
White bass
Carp
Short-nosed Gar
e29
1,160
886
66.3
71.2
33.3
HI I 'Hi 8 —TOT
22.1
<1
70 + 4
<0.005
<0.006
* Total initial Mights on these three fish species and ashed reulnt were:
S?fc1?s Wet Weloht (ms) Ash Weloht font)
10
<4
M + 6
<0.02**
<0.03**
10
<2
37 + t
<0.002
<0.002
The ashed remains were distributed equally between the three laboratories.
~ These activity levels were carefully recfteckad on a duplicate 10 ( of ashed sample and were found
to be below 0.002 pCl/g (ash) for Pu-238 In Gar and &asi. below 0.003 and 0.002 for Pu-239 for tar
and Best, respectively, and below 0.016 and 0.014 pCi/g for Pu-238 and Pu-239, respectively, In Carp.
5
19-6
------- |