RESEARCH TRIANGLE INSTITUTE
SEMINAR/WORKSHOP PROCEEDINGS:
PERSISTENT ELEVATED POLLUTION EPISODES
(March 1979, Durham, North Carolina)
Compilers
Harry L. Hamilton, Jr.
N. Stuart Jones
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
Contract No. 68-02-3000
Project Officer
William E. Wilson
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27709
-------
SEMINAR/WORKSHOP PROCEEDINGS:
PERSISTENT ELEVATED POLLUTION EPISODES
(March 1979, Durham, North Carolina)
Compilers
Harry L. Hamilton, Jr.
N. Stuart Jones
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
Contract No. 68-02-3000
Project Officer
William E. Wilson
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, North Carolina 27711
Prepared for
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
-------
PREFACE
The Seminar/Workshop on Persistent Elevated Pollution Episodes (PEPE's)
was sponsored by the Regional Field Studies Office, Environmental Sciences
Research Laboratories (ESRL), Environmental Protection Agency (EPA) to provide
for the exchange of information among scientists and to generate, from the
Workshop, a set of recommendations for the conduct of future regional scale
field investigations. These investigations will be directed toward obtaining
the data base required to elucidate the atmospheric chemical and physical pro-
cesses involved in the development of persistent regional scale areas of high
concentrations of pollutants, particularly aerosols that can cause degradation
of visibility.
This PEPE seminar/workshop is the second to be sponsored by EPA/ESRL to
provide guidance for field studies. The first, a Workshop on Regional Air
Pollution Studies, was held at Boone, North Carolina, in June 1976. The re-
sults of that workshop led to the development of a program to investigate
Sulfur Transport and Transformation in the Environment (STATE).
The PEPE seminar/workshop is directed toward the continuation of the
STATE Program through the exchange of information acquired from recently
completed observation and measurement programs, reviews of the present state-
of-the-art measurement techniques applicable to the detection and delineation
of PEPE's, and discussions of methods for identifying and tracking air parcels
in dynamic flow fields.
The papers presented at the seminar appear in this volume in the order in
which they were presented. All of the speakers were invited and asked to
speak on subjects germane to the objectives of the subsequent workshops.
After 2h days of formal presentations, four concurrent workshops were
held to provide definitive recommendations on experiment design, measurement
strategy and tactics, and communications requirements; meteorology, transport,
and multiday parcel tracking; model development and validation—measurement
needs; and instrumentation. The instrumentation workshop was divided into two
ii
-------
sessions: instrumentation for study transformation and instrumentation for
remote sensing.
The reports of the workshop chairmen were made to a final plenary ses-
sion. These reports were also distributed at the close of that plenary ses-
sion to all recipients of the Environmental Protection Agency Request for
Proposal DU 78-B246, "Conduct Research Program Designed to Study Transport of
Pollutants in Power Plant Plumes, Urban and Industrial Plumes, and Persistent
Elevated Pollution Episodes." Representatives of each recipient firm had been
invited to attend and participate in all sessions of the seminar/workshop.
Dr. William E. Wilson, Director, Regional Field Studies Office, EPA/ESRL,
was the Project Officer for contract 68-02-3000 under which the seminar/work-
shop was conducted, and served as General Chairman of the sessions; Mr. James
J. B. Worth of the Research Triangle Institute was the Technical Chairman.
Mr. Harry L. Hamilton, Jr., also of RTI, planned and coordinated the seminar/
workshop.
iti
-------
19 March 1979
Contents
Page
Identification of PEPE's By Visibility Isopleths 1
R. B. Husar and D. E. Patterson
Evidence of Transport of Hazy Air Masses
From Satellite Imagery 19
Walter A. Lyons
Remote Sensing of Regional Air Pollution
From Satellites 46
E. J. Friedman and E. L. Kertz
Characterization of Regional Sulfate/Oxidant Episodes
in the Eastern United States 95
E. Y. long, M. T. Mills, B. L. Niemann, and L. F. Smith
Statistics of Elevated Pollution Episodes 123
Gerald F. Watson and Walter J. Saucier
Some Dynamic Aspects of Extended Pollution Episodes 152
William J. King and Fred M. Vukovich
Application of Numerical Models to Prolonged Elevated
Pollution Episodes 178
C. Shepherd Burton and Mei-Kao Liu
EPRI's Air Quality Studies 195
Glenn R. Hilst
Ensemble Trajectory Analysis of Summertime Surfate
Concentrations in the Eastern United States 202
Perry J. Samson
20 March 1979
Some Results of the Surfate Regional Experiment and
the PEPE Experiment Design 214
G. M. Hidy, P. K. Mueller, and T. F. Lavery
Fluorocarbon Tracer System for Atmospheric
Transport and Dispersion Studies 228
Gilbert J. Ferber
TV
-------
20 March 1979 (continued)
Contents (continued)
Page
MAP3S Programs Applicable to PEPE's 236
Michael C. MacCracken
Regional Scale Aircraft Mapping of Hazy Air Masses
During Project MISTT, 1975 and 1976 243
D. E. Patterson, J. M. Holloway, and R. B. Husar
Ground Measurements of Regional Haze 341
Kenneth T. Whitby and Peter H. McMurray
Comparison of Air Quality Within the Mixed Layer
With That at the Surface 367
James J. B. Worth
Passive Remote Sensing of SO2 406
Milla'n M. Milla'n
Summary of the NASA Remote Sensing Program 425
S. H. Melfi
Validity of Balloons for Air Parcel Tracking 450
Frank B. Tatom and George H. Fichtl
Navigational Aids for Air Parcel Tracking 469
James G. Haidt
Satellite Positioning of Tracers. . . 480
James J. B. Worth
Radar Tracking 489
C. Raymond Dickson
Light-Weight, Low-Cost Radar, Transponder Systems 500
Earl F. Pound
Miniaturized Laser Doppler Velocimeter 504
David H. Dickson
21 and 22 March 1979-Workshops
I. Experiment Design, Measurement Strategy and Tactics,
Communications Requirements 529
Roy Evans, Chairman
II. Meteorology, Transport, and Multiday Tracking 545
Harry L. Hamilton, Jr., Chairman
-------
Contents (continued)
Page
21 and 22 March 1979 (continued)
III. Model Development and Validation—Measurement Needs 551
George M. Hidy, Chairman
IV. Part 1. Instrumentation for Study of Transformation 560
Robert K. Stevens, Chairman
IV. Part 2. Instrumentation for Remote Sensing 567
Frank Allario, Chairman -
VI
-------
PARTICIPANTS
Abdul J. Alkezweeny
Battelle Pacific Northwest Laboratories—Area Office
Muskegon MI
Frank Allario
Langley Research Center
National Aeronautics and Space Administration
Hampton VA
A. Paul Altshuller
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Kurt Anlauf
Atmospheric Environment Service
Downsville, Ontario
Wendell Ayers
Langley Research Center
National Aeronautics and Space Administration
Hampton VA
John D. Bachman
Strategies and Air Standards Division
U.S. Environmental Protection Agency
Research Triangle Park NC
W. D. Balfour
Radian Corporation
Austin TX
Francis Binkowski
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Vtt
-------
Donald L. Blumenthal
Meteorology Research, Inc.
Santa Rosa CA
C. Shepherd Burton
Systems Applications, Inc.
San Rafael CA
Michael Chan
Aerovironment, Inc.
Pasadena CA
Jason Ching
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Robert Chunn
Contracts Management Division
U.S. Environmental Protection Agency
Research Triangle Park NC
John F. Clarke
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Samual C. Coroniti
Dynatrend, Inc.
Burlington MA
Richard Coulter
Argonne National Laboratory
Argonne IL
Dagmar Cronn
Washington State University
Pullman WA
Walter F. Dabberdt
SRI International
Menlo Park CA
Vi ii
-------
Kenneth L. Demerjian
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Rosa de Pena
Department of Meteorology
Pennsylvania State University
State College PA
C. Raymond Dickson
Idaho National Engineering Research Laboratory
National Oceanic and Atmospheric Administration
Idaho Falls ID
Allen C. Dittenhoefer
Department of Meteorology
Pennsylvania State University
State College PA
Ron Drake
Battelle Pacific Northwest Laboratories
Richland WA
Jack L. Durham
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research triangle Park NC
Thomas G. Dzubay
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Gary Eaton
Research Triangle Institute
Research Triangle Park NC
John Eckert
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Las Vegas NV
ix
-------
James G. Edinger
Atmospheric Sciences Department
University of California at Los Angeles
Los Angeles CA
Roy Evans
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Gilbert J. Ferber
Air Resources Laboratory
National Oceanic and Atmospheric Administration
Silver Spring MD
Edward Friedman
The MITRE Corporation
McLean VA
G. T. Gay
Sandia Laboratories
Albuquerque NM
Noor Gillani
Department of Mechanical Engineering
Washington University
St. Louis MO
Dan Golomb
Office of Energy, Minerals and Industry
U.S. Environmental Protection Agency
Washington DC
James G. Haidt
Research Triangle Institute
Research Triangle Park NC
Harry L. Hamilton, Jr.
Research Triangle Institute
Research Triangle Park NC
George M. Hidy
Environmental Research and Technology, Inc.
West Lake Village CA
-------
Glenn R. Hilst
Electric Power Research Institute
Palo Alto CA
George C. Holzworth
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Rudolf B. Husar
Department of Mechanical Engineering
Washington University
St. Louis MO
N. Stuart Jones
Research Triangle Institute
Research Triangle Park NC
Edwin Keitz
The MITRE Corporation
McLean VA
William King
Research Triangle Institute
Research Triangle Park NC
Robert G. Lamb
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Bryan W. Lambeth
Radian Corporation
Austin TX
Lee Langan
Environmental Measurements, Inc.
San Francisco CA
Paul Lioy
Institute for Environmental Medicine
New York University Medical Center
New York NY
xi
-------
Mei-Kao Liu
Systems Applications, Inc.
San Rafael CA
Man's Lusis
Air Resources Branch
Toronto, Ontario
Walter A. Lyons
MESOMET, Inc.
Chicago IL
Michael MacCracken
Lawrence Livermore Laboratory
Department of Energy
Livermore CA
David Mage
Environmental Monitoring and Support Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Ronald J. Massa
Dynatrend, Inc.
Burlington MA
Peter H. McMurray
Department of Mechanical Engineering
University of Minnesota
Minneapolis MN
S. Harvey Melfi
Office of Space and Terrestrial Applications
National Aeronautics and Space Administration
Washington DC
Edwin L. Meyer
Monitoring and Data Analysis Division
U.S. Environmental Protection Agency
Research Triangle Park NC
Ron Meyers
Brookhaven National Laboratory
Upton NY
xii
-------
Millan M. Millan
Atmospheric Environment Service
Downsview, Ontario
Dave Miller
Battelle Columbus Laboratories
Columbus OH
R. L. Myers
Rockwell International
Creve Coeur MO
Larry E. Neimeyer
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Leonard Newman
Brookhaven National Laboratory
Upton NY
Brand L. Niemann
Teknekron, Inc.
Berkeley CA
A. F. Oberta
Environmental Resources Group, Inc.
Danville VA
Brynjulf Ottar
Norwegian Institute for Air Research
Royal Norwegian Council for
Scientific & Industrial Research
Lillestrom, Norway
T, G. Pace
Monitoring and Data Analysis Division
U.S. Environmental Protection Agency
Research Triangle Park NC
Robert A. Papetti
Office of Air, Land and Water Use
U.S. Environmental Protection Agency
Washington DC
Xfti
-------
David Patterson
Department of Mechanical Engineering
Washington University
St. Louis MO
Francis Pooler
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Earl Pound
Utah State University
Logan UT
Rudolf Pueschel
Environmental Research Laboratory
National Oceanic and Atmospheric Administration
Boulder CO
Madhav Ranade
Research Triangle Institute
Research Triangle Park NC
J. C. Romanovsky
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Perry Samson
N.Y. State Department of Environmental Conservation
Albany NY
Dave Sanchez
Control Programs Development Division
U.S. Environmental Protection Agency
Research Triangle Park NC
Walter J. Saucier
Department of Geosciences
North Carolina State University
Raleigh NC
xtv
-------
Frank A. Schiermeier
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
St. Louis MO
Joseph Sickles, II
Research Triangle Institute
Research Triangle Park NC
Lowell Smith
Office of Energy, Minerals and Industry
U.S. Environmental Protection Agency
Washington DC
Chester W. Spicer
Battelle Columbus Laboratories
Columbus OH
Lester L. Spiller
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
Gene E. Start
Idaho National Engineering Research Laboratory
National Oceanic and Atmospheric Administration
Idaho Falls ID
Gary Stensland
Illinois State Water Survey
Urbana IL
Robert K. Stevens
Environmental Sciences Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park NC
George Sverdrup
Battelle Columbus Laboratories
Columbus OH
Gary Tannahill
Radian Corporation
Austin TX
xv
-------
Frank B. Tatom
Engineering Analysis, Inc.
Huntsville AL
Prekimi U. Tawath
The MITRE Corporation
McLean VA
James Tommerdahl
Research Triangle Institute
Research Triangle Park NC
Eugene Tong
Teknekron, Inc.
Berkeley CA
Edward Uthe
SRI International
Menlo Park CA
William M. Vaughan
EMI, Inc.
University City MO
Fred Vukovich
Research Triangle Institute
Research Triangle Park NC
Scott Wagner
Office of Space and Terrestrial Applications
National Aeronautics and Space Administration
Washington DC
T. L. Waldron
Rockwell International
Creve Coeur MO
Gerald Watson
Department of Geosciences
North Carolina State University
Raleigh NC
XVt
-------
Marvin L. Wesely
Argonne National Laboratory
Argonne IL
Kenneth T. Whitby
Department of Mechanical Engineering
University of Minnesota
Minneapolis MN
William E. Wilson
Regional Field Studies Office
Environmental Sciences Research Laboratory
Research Triangle Park NC
George Wolff
General Motors Research Laboratories
Warren MI
James J. B. Worth
Research Triangle Institute
Research Triangle Park NC
Bernard Zak
Sandia Laboratories
Albuquerque NM
xvit
-------
ABBREVIATIONS
AES Atmospheric Environment Services
APT automatic picture transmission
ATS Applications Technology Satellite
AVHRR advanced very high resolution radiometer
BNL Brookhaven National Laboratory
BOR Biggs Optical Range
b . light scattering due to aerosols
CBM carbon-bond mechanism
CIMATS correlation interferometer for measuring atmospheric trace species
CNC condensation nuclei count
COG center of gravity
COPE Carbon Monoxide Pollution Experiment
COSPEC correlation spectrometer
CVB constant-volume balloon
DARS differential absorption remote tensing
DIAL differential absorption lidar
EAA electrical aerosol analyzer
EML Environmental Measurements Laboratory (DOE)
EPE elevated pollution episode
EPRI Electric Power Research Institute
ERAQS Eastern Regional Air Quality Studies
ESRL Environmental Systems Research Laboratory
xviii
-------
FAA Federal Aviation Administration
FPD flame photometric detection
GCMS gas chromatography-mass spectrometry
GFC gas filter correlation
HK haze and/or smoke
HRWS helicopter remote wind sensor
HSI high-speed interferometer
HSRL high spectral resolution lidar
IHR infrared heterodyne radiometer
IHS infrared heterodyne spectrometer
IPM inhalable participate matter
IPMN Inhaled Participate Monitor Network
LAS laser absorption spectrometer
LCS laser crosswind sensor
LDV laser doppler velocimeter
LHS laser heterodyne spectrometer
LPI low-pressure impactor
LRC Langley Research Center
MAPS measuring air pollution from satellites
MAP3S Multistate Atmospheric Power Production Pollution Study
McIDAS Man/Computer Interactive Data Access System
MISTT Midwest Interstate Sulfur Transformation and Transport
MOPS multispectral observation of pollutants
MSS multispectral scanner
NASN National Air Sampling Network
NBS National Bureau of Standards
xix
-------
NESS National Environmental Satellite Service
NMHC nonmethane hydrocarbon
NOAA National Oceanic and Atmospheric Administration
NWS National Weather Service
OFT MAPS oribital flight test MAPS
OPC optical particle counter
ORBS Ohio River Basin Study
PEPE persistent elevated pollution episode
PMR pressure-modulated radiometer
PMV Plume Model Validation
PSD prevention of significant deterioration
RAMS Regional Air Monitoring System
Random Access Measurement System
RAPS Regional Air Pollution Study
RAQS Regional Air Quality Study
RBV return beam vidicon
RF radio frequency
RH relative humidity
SAGE Stratospheric Aerosol and Gas Experiment
SAM Stratospheric Aerosol Measurement
SAROAD Storage and Retrieval of Aerometric Data
SBUV-TOMS solar backscatter ultraviolet-total ozone measurement system
SMS/GOES Synchronous Meteorological Satellite/Geostationary Operational
Environmental Satellite
STATE Sulfur Transport and Transformation Experiment
SURE Sulfate Regional Experiment
TDMA tandem differential mobility analyzer
xx
-------
TSP total suspended particulate
TVA Tennessee Valley Authority
UMML University of Minnesota Mobile Laboratory
UV DIAL ultraviolet differential absorption lidar
VAS VISSR atmospheric sounder
VCO voltage-controlled oscillator
VISSR visual infrared spin-scan radiometer
XRF x-ray fluorescence
xxi
-------
IDENTIFICATION OF PEPE'S BY VISIBILITY ISOPLETHS
R. B. Husar* and D. E. Patterson
Washington University, St. Louis, Missouri
Abstract
Persistent elevated pollution episodes (PEPE's) may be identified by
regional scale high concentrations of secondary sulfate and ozone. Sulfate
aerosol causes visibility reduction, for which fine spatial and temporal
resolution is available. Visibility appears to be a suitable surrogate for
sulfate concentrations, so that the spatial distribution and trends of visibi-
lity offer information about the frequency of occurrence of PEPE's.
INTRODUCTION
The past several decades have seen a shift in the spatial scale of
visibility-reducing particulates, expanding from areas immediately surrounding
each factory to entire metropolitan areas to multistate regions covered by
haze. The control of smokestack particulate concentrations in the atmosphere
has reduced the occurrence of locally visible "hot spots". However, the
emissions of precursor gases such as S02, hydrocarbons, and NO have increased
A
greatly over this period. As will be discussed below, the average haziness
has increased over most of the United States in the past three decades, due to
the production in the atmosphere of secondary aerosols such as sulfate.
The atmospheric transmission of tall stack point source and urban area
source effluents has been studied extensively during the past decade, most
recently in the Midwest Interstate Sulfur Transformation and Transport (MISTT)
project and in the Sulfur Transport and Transformation in the Environment
(STATE) project, where the transport, chemical transformations, removal, and
process interactions in determining the sulfur budget of large plumes were
assessed.
*Speaker.
-------
URBAN PLUMES
Urban plumes constitute an aggregate plume originating from various
sources within a metropolitan area. One thoroughly studied urban plume is
that of the metropolitan St. Louis area, a major urban industrial center
including coal-fired power plants with a combined capacity of 4,600 MW, oil
refineries with a combined capacity of 4.4 x io5 bbl/d (1 bbl = 160 L),
various other industries, and a population of about two million (White et al.,
1976). Because St. Louis is remote from other major metropolitan areas, its
impact on the surrounding ambient air quality is relatively easy to identify;
air that has been modified by the aggregate emissions of the metropolitan area
form an "urban plume" downwind.
As part of project MISTT, the three-dimensional flow of aerosols and
trace gases in the St. Louis urban plume was studied (Wilson, 1978). The
plume was successfully tracked up to 240 km, and it was mapped quantitatively
up to 160 km (Figure 1). At these distances, the plume was still defined and
on the order of 50 km wide.
An increased concentration of light-scattering aerosols is a key character-
istic of the St. Louis urban plume. The primary contribution of project MISTT
was to quantify the flow of material at increasing downwind distances so as to
study-the transformations that pollutants undergo in the atmosphere.
The flow rate of ozone, b ., and particulate sulfur, S , all increased
SC3 L P
with distance downwind of St. Louis on July 18, 1975, reflecting the secondary
origin of ozone and most of the light-scattering aerosols (White et al.,
1976). Most of the increase in the b .flow rate was observed downwind of
the major increase in ozone flow rate, consistent with the finding of the
laboratory studies that aerosol production lags behind ozone production in a
photochemical system (Wilson et al., 1973). The ratio of the flow rate of
b . to the flow rate of S indicates that sulfate compounds accounted for
SCclT> P
most of the newly formed aerosol in the urban plume. This case study illus-
trates that emissions from a metropolitan area such as St. Louis cause reduced
visibility and elevated ozone concentrations in urban plumes, long after their
primary precursors have been diluted to low concentrations.
From extrapolation of measured plume transport, it is clear that the
urban plume of one metropolitan area may become a "background haziness" for
another. It is therefore reasonable to assume that the pollutant back-
-------
ground in nonurban regions of the eastern United States is likely due to the
combined emissions of metropolitan and industrial areas, as well as to emissions
from power plants far upwind. Regrettably, direct observations of plume
impacts have only been carried out up to a few hundred kilometers from their
sources. Beyond those distances, single or even urban plumes tend to elude
convenient aircraft mapping; also, they are likely to be superimposed on a
plume of another major source. It has been proposed that the observed large-
scale haziness of a large part of the eastern United States during stagnation
is in fact the result of the superposition of numerous urban and industrial
power plant plumes, as discussed in the next section.
Regional Scale PEPE Episodes
Early works on air pollution episode potential focused on local inversion
situations rather than stagnation of a regional scale air mass. Hosier (1961)
found that the southeast coast of the United States and the Smoky Mountain
region had the highest incidence of surface inversions nationally. Holzworth
(1967) evaluated the pollution potential of several U.S. urban areas based on
mean mixing heights and wind speeds. Korshover (1967) constructed a climatology
of stagnating anticyclones in the Eastern United States from 1936-1965 and
found the highest potential for large-scale stagnation east of the Rocky
Mountains to be in the Southern Appalachian and Smoky Mountain regions.
One of the earliest examinations of regional scale air pollution detailed
the evolution and impact of the Thanksgiving 1966 episode over the eastern
U.S. seaboard, in which a stagnating high-pressure system caused prolonged
elevated regional concentrations of primary pollutants and total suspended
particulates.
Formation of secondary pollutants (such as sulfate) in the atmosphere is
believed to be the major problem of regional air pollution today. The residence
time of secondary particles is the sum of the time required to create them and
the subsequent time for removal from the atmosphere. The flow diagram of
sulfur transmission through the atmosphere (Figure 3) illustrates, for example,
that most of the S02 emission is removed or transformed to sulfate within the
first day of atmospheric residence, but the atmospheric burden of the secondary
sulfate may remain relatively high for several days of residence time.
-------
Mean winds of 500 km per day are not uncommon in the eastern United
States (Blumenthal et al., 1978), so a reasonable scale of sulfate impact from
an individual S02 source is on the order of 1,000 to 2,000 km. The implication,
as seen in Figure 4, is that ambient light-scattering sulfate aerosol can
result from the additive effects of tens or even hundreds of S02 plumes of
different atmospheric ages. Long-range transport of fine aerosol sulfate
within a stagnating regional scale air mass can create a PEPE.
In the past decade, several researchers have presented evidence of long-
range transport of visibility-reducing fine aerosol. Rodhe (1972), Brosset et
al. (1975), Eliassen and Saltbones (1975), Smith and Jeffrey (1975), and
Szepesi (1978) are among those who have established the reality of long-range
transport of sulfates in Europe. Chung (1978) found evidence of transport of
sulfate from the northeastern Midwest of the United States into southeast
Canada.
In the United States, Altshuller (1976) noted the anomaly of decreasing
urban S02 concentrations with increasing rural sulfate trends and proposed
long-range transport as the cause of regional sulfate levels. Wilson et al.
(1976) also proposed long-range transport as the explanation for regional
sulfate concentrations, based on field studies of plume transmission. Gage et
al. (1977) demonstrated the sulfate episode potential when multiple S02
sources are aligned along the wind direction.
One of the earliest case studies of transport of large-scale hazy air
masses was that of Hall et al. (1973). Since about 1975, the evolution and
transport of regional scale hazy air masses have been receiving increasing
attention from numerous research groups. Detailed case studies of such
episodes have been reported by long et al. (1976), Husar et al. (1976), Lyons
and Husar (1976), Wolff et al. (1977), Samson and Ragland (1977), Vukovich et
al. (1977), Calvin et al. (1978), Hidy et al. (1978), and Chung (1978), among
others.
A common finding in recent studies is that formation of regional scale
haziness is usually associated with the presence of slow-moving high-pressure
systems. Because precipitation is relatively infrequent in anticyclonic
systems, the residence time of sulfate aerosol may be as long as a week or
more.
-------
An example of one such episode over a 2-week period in June-July 1975 is
presented in Figure 5 as visibility isopleths. Inspection of the sequence of
contour maps reveals that multistate regions are covered by a haze layer
-4 -1
in which noon visibility is less than 10 km (b .>4 x 10 m , outer contours).
The apparent movement of the areas of haze was confirmed by geostationary
satellite photographs. The region of haziness on July 30 may be clearly seen
in Lyons and Husar (1976), where the turbid air mass is visible over the
States of Arkansas, Missouri, Kansas, Iowa, and Minnesota.
The air mass of June 25, 1975 (defined by the shaded region of visibility
6.5-10 km) was of maritime origin in the Gulf of Mexico and had followed a
trajectory through Louisiana northward into Illinois and Indiana. Between
June 25 and 27, relative stagnation prevailed in the Great Lakes region,
leading to increasing haziness in this high emission density area. The tropi-
cal storm "Amy" moving northward off the northeast coast blocked the eastern
motion of the hazy air mass as a front approached the Great Lakes from the
northwest. , Between June 28 and 30 an easterly flow developed, causing the
hazy air mass to drift slowly westward, passing over St. Louis, MO, on June
28-29 and continuing across Missouri and Kansas by June 30. From June 30 to
July 2 the haze moved in a clockwise pattern to the southern Great Lakes. By
July 3, the advancing Canadian front formed the northern border of an exten-
sive hazy air mass occupying most of the midwestern and northeastern United
States. The hazy air mass again passed over St. Louis at this time. During
July 3-5 the cold front advanced rapidly to the south, pushing the heavily
polluted air mass across the southeastern States and out to sea.
The two passages of the hazy air mass over St. Louis, MO, resulted in
sharp increases of b . over the entire metropolitan region (Figure 6a).
Sulfate concentration also increased during the haze episode, from about 9 to
33 ug/m3. The spatial coherency of haziness is seen in the correspondence of
the extinction coefficient at St. Louis, MO, and Springfield, IL (Figure 6b),
again confirming that the observed haziness was primarily due to inflow of the
polluted "background" material of the hazy air mass rather than local contribu-
tions.
Sufficient sulfate data were available from the National Aerometric Data
Bank for comparison with visibility on two days during the episode period.
-------
Figure 7 indicates the substantial correspondence of sulfate and visibility:
sulfate concentrations in excess of 30 ug/m3 are seen to coincide with regions
of lowest visibility.
In order to examine the correlation between ozone and visibility reduc-
tion on a synoptic scale, contour plots of SAROAD daily maximum ozone were
also prepared (by hand) for the period June 25 through July 5, 1975 (Figure
8).
The plots show that the former U.S. standard for ozone (0.08 ppm) was
exceeded over large areas of the eastern United States on all days of the air
pollution episode. Comparison of the maps for visibility and ozone (Figures 5
and 8, respectively) reveals that the geographical location of high ozone
concentrations roughly corresponds to the areas of low visibility (and high
sulfate). As may be anticipated, however, the correlation with haziness (low
visibility) is much better for sulfates than for ozone.
During this period, the visual air quality was beyond the control of any
local jurisdiction. The Alabama Air Pollution Control Commission (AAPCC)
reported the following (Bulletin of the AAPCC, 1975):
During the weekend of July 5, 1975, a heavy haze layer enveloped the
state of Alabama and much of the Southeastern U.S. At that time, the
AAPCC technical staff received many comments from the public concerning
the origin and composition of the haze.
The National Weather Service in Birmingham did issue an air stagnation
advisory (ASA) for Alabama for this same time period; however, the
traditional pollutant measurements made by the AAPCC and local programs
did not show excessive levels. In fact, the measured local levels were
lower than had been measured under previous ASA's, making the dramatic
decrease in visibility more intriguing.
Husar et al. (1976) reported that in June-August 1975 there were at least
six episodes similar to the above. The work of other investigators confirms
that episodes of regional scale hazy air masses are not rare in the eastern
United States. Yet at present only the qualitative features of such episodes
are understood, i.e., the observed effect on visibility, the composition in
terms of secondary sulfate and ozone, and the apparent motion of the haze.
Important questions remain to be answered about regional scale episodes
of haziness, including the following:
-------
To what degree do the hazy air mass and the meteorologically defined
anticyclone coincide?
Are the effects of superimposing multiple S02 plumes and urban
reactive plumes simply additive?
What are the roles and possible feedback mechanisms of high pollutant
concentrations on dry and wet deposition, cloudiness, etc?
What is the actual residence time of the particulates in the atmos-
phere during such episodes?
TRENDS AND SPATIAL DISTRIBUTION OF HAZINESS
The upcoming Prolonged Elevated Pollution Episode (PEPE) project of STATE
will be specifically designed to sample such regional scale episodes of haziness
from their inception throughout their residence over the eastern United States.
PEPEs are important factors in determining the mean extinction coefficient (or
haziness). Therefore, historical information on visibility trends and spatial
distribution may be useful. Figures 9 and 10, with their associated captions,
summarize the seasonal and spatial trends of haziness over the eastern United
States since 1948.
SUMMARY
This report and the companion paper, "Regional Scale Aircraft Mapping of
Hazy Air Masses During Project MISTT, 1975 and 1976," present direct evidence
of long-range (200-300 km) transport of urban and power plant plumes. Regional
scale polluted air masses, defined by elevated concentrations of sulfate and
ozone, may be followed by the examination of visibility isopleths. The region
of most severe average summertime haziness includes the Ohio River Valley and
the Smoky Mountain States; therefore, the upcoming PEPE field study probably
should be located within this area.
ACKNOWLEDGMENT
This research has been supported by the Federal Interagency Energy/
Environment Research and Development Program through U.S. Environmental
Protection Agency Grant No. R803896. The work was conducted as part of
projects MISTT (Midwest Interstate Sulfur Transformation and Transport) and
STATE (Sulfur Transformation and Transport in the Environment) with the close
cooperation of Dr. William E. Wilson, Jr., Project Officer.
-------
REFERENCES
Altshuller, A. P. 1976. Regional Transport and Transformation of Sulfur
Dioxide to Sulfates in the U.S. J. Air Pollut. Control Ass.. 26:318-
324.
Blumenthal, D. L., J. A. Ogren, and J. A. Anderson. 1978. Airborne Sampling
System for Plume Monitoring. Atmospheric Environment, 12:613-620.
Brosset, C., K. Andreasson, and M. Perm. 1975. The Nature and Possible
Origin of Acid Particles Observed at the Swedish West Coast. Atmospheric
Environment, 9:631-642.
Chung, Y. S. 1978. The Distribution of Atmospheric Sulfates in Canada and
its Relationships to Longe-Range Transport of Air Pollutants. Atmos-
pheric Environment, 12:1471-1480.
Eliassen, A., and J. Saltbones. 1975. Decay and Transformation Rates SO ,
as Estimated From Emission Data, Trajectories, and Measured Air Concen-
trations. Atmospheric Environment, 9:425-429.
Gage, S. J., L. F. Smith, P. M. Cukor, and B. L. Nieman. 1977. Long-Range
Transport of SO /MS04 From the U.S. EPA/Teknekron Integrated Technology
Assessment of Electric Utility Energy Systems. Proceedings International
Symposium on Sulfur in the Atmosphere (ISSA), Dubrovnik, Yugoslavia.
Galvin, P. J., P. J. Samson, E. C. Peter, and D. Romano. 1978. Transport of
Sulfate to New York State. Atmospheric Environment, 9:643-659.
Hall, F. P., Jr., C. E. Duchon, L. G. Lee, and R. R. Hagan. 1973. Long Range
Transport of Air Pollution: A Case Study, August 1970. Mon. Weather
Rev.. 101:404-411.
Hidy, G. M., P. K. Mueller, and E. Y. Tong. 1978. Spatial and Temporal Dis-
tributions of Airborne Sulfate in Parts of the United States. Atmospheric
Environment. 12:735-752.
Hosier, C. R. 1961. Low Level Inversion Frequency in the Contiguous United
States. Monthly Weather Rev.. 89:319-339.
Holzworth, G. C. 1967. Mixing Depths, Wind Speeds and Air Pollution Potential
for Selected Locations in the United States. JAM. 6:1039-1044.
Husar, R. B., D. E. Patterson, C. C. Paley, and N. V. Gillani. 1976. Ozone
in Hazy Air Masses. Paper presented at the International Conference on
Photochemical Oxidant and its Control, Raleigh, NC, September 12-17, 10
pp.
Korshover, J. 1967. Climatology of Stagnating Anticyclones East of the Rocky
Mountains, 1936-1965. NAPCA.
Lyons, W. A., and R. B. Husar. 1976. SMS/GOES Visible Images Detect a Synoptic-
Scale Air Pollution Episode. Mon. Weather Rev.. 104:1623-1626.
8
-------
Rodhe, H., C. Persson, and 0. Akesson. 1972. An Investigation Into Regional
Transport of Soot and Sulfate Aerosols. Atmospheric Environment. 6:675-
693.
Samson, P. J., and K. W. Ragland. 1977. Ozone and Visibility Reduction in
the Midwest: Evidence for Large-Scale Transport. J. Appl. Meteorol.,
16:1101-1106.
Smith, F. B., and G. H. Jeffrey. 1975. Airborne Transport of Sulfur Dioxide
From the United Kingdom. Atmospheric Environment. 9:643-659.
Szepesi, D. J. 1978. Transmission of Sulfur Dioxide on Local, Regional and
Continental Scale. Atmospheric Environment, 12:529-535.
Thanksgiving 1966 Air Pollution Episode in the Eastern United States. NAPCA.
AP-45, 1968.
Tong, E. Y., G. M. Hidy, T. F. Lavery, and F. Berlandi. 1976. Regional and
Local Aspects of Atmospheric Sulfates in the Northeast Quadrant of the
U.S. Proceedings Third Symposium of Turbulence Diffusion and Air Quality,
AMS.
Vukovich, F. M., W. D. Bach, Jr., B. W. Crissman, and W. J. King. 1977. On
the Relationship Between High Ozone in the Rural Surface Layer and High
Pressure Systems. Atmospheric Environment. 11:967-983.
White, W. H., J. A. Anderson, 0. L. Blumenthal, R. B. Husar, N. V. Gillani,
J. D. Husar, and W. E. Wilson. 1976. Formation and Transport of
Secondary Air Pollutants: Ozone One Aerosols in the St. Louis Urban
Plume. Science. 194:187-189.
Wilson, W. E., Jr., D. F. Miller, A. Levy, and R. K. Stone. 1973. J. Air
Pollut. Control Assoc.. 23:949.
Wilson, W. E., R. J. Charlson, R. B. Husar, K. T. Whitby, and D. L. Blumenthal.
1976. 69th Annual Meeting of Air Pollution Control Assoc., Portland, OR,
June 27-July 1, 76-30-06.
Wilson, W. E., Jr. 1978. Sulfates in the Atmosphere: A Progress Report on
Project MISTT. Atmospheric Environment. 12:537-547.
Wolff, G. T., P. J. Lioy, R. E. Meyers, and R. T. Cederwall. 1977. An In-
vestigation of Long-Range Transport of Ozone Across the Midwestern and
Eastern United States. Atmospheric Environment, 11:797-802.
-------
Champaign^/
d.
Urbana
. Missouri f Illinais
A Pow«r plant
I Rafintrits
Figure 1. Concentrations of O3 and bscat values downwind of St. Louis on 18 July 1975.
Data are taken from horizontal traverses by instrumented aircraft, at altitudes
indicated in Figure 2A. Graph baselines show sampling paths; baseline concen-
trations are not zero (White et al.).
© 1976 by the American Association for the Advancement of Science. This figure originally appeared in an
article by W. H. White et al. in volume 194 of Science, pp. 187-189.
-------
•r I A
. mo neo, mi .nni
10 100 150
Dittanci downwind (km)
- 3
If
J
Figure 2. Traverse altitudes and pollutant flow rates in the St.
Louis urban plume on 18 July 1975. Data are plotted
against distance downwind of the St. Louis Gateway
Arch. (A) Location of horizontal traverses; closed
circles correspond to traverses shown in Figure 1.
Mixing heights were determined from aircraft sound-
ings. Approximate time (COT) of sampling is shown at
the bottom. (B) Flow rates (in excess of background)
of 03, bscat, and Sp (White et al.).
©1976 by the American Association for the Advance-
ment of Science. This figure originally appeared in an
article by W. H. White et al. in volume 194 of
Science, pp. 187-189.
SOT" AEROSOL
z-LJTl
ORT
wez
sy'/////////////,y'/'/'///////////'A/'/////////////<
I RRST DAY IScCOND DAYl THIRD DAY 1
Figure 3. Flow diagram of sulfur transmission through the atmosphere.
11
-------
Figure 4. Schematic illustration for the range of transport of fine particles. A reasonable
mean transport distance is considered to be 1,500 km. The actual transport
depends on the synoptic-scale wind field.
12
-------
Figure 5. Sequential contour maps of noon visibility for June 25-July 5, 1975, illustrates
the evolution and transport of a large-scale hazy air mass. Contours correspond
to visual range 6.5-10 km (light shade), 5-6.5 km (medium shade), and
<5 km (black).
13
-------
_ w
01
STATION 4
STATION 6
STATION 9
ui
u.
_J
91
. 6/23 6/24 6/25 6/2S 6/27 6/23 6/S 6/33 7/1 7/2 7/3 7/4 7/3
Figure 6. Local monitoring data in the St. Louis, MO, area during the June-July 1975
haziness episode, (a) Light-scattering coefficient (bscat) recorded at three widely
spaced locations in the St. Louis metropolitan area and daily average sulfate
concentrations; (b) extinction coefficients (bext) obtained from visibility observa-
tions at St. Louis, MO, and Springfield, IL, 150 km apart.
14
-------
aumnr oe
C9 20BJO
23
JUT S.3J3
ea •»
Figure 7. Comparison of noon extinction coefficient and daily mean sulfate concentration
on June 23 and July 25, 1975. The regions of highest sulfate concentrations
coincide with areas of lowest visibility.
15
-------
JJ£ 29
JX
JU£ 27
/
JULT I
•*•• ••• OZONE CCNC foam?
J2 to J6
EZ2 £8 to .12
Figure 8. Sequential contour plots of daily peak ozone concentration.
16
-------
OHIO RIVER VflLLEY
X
0)
..i i < i i
1950 I960 1970 1980 199
3UPRTER 1
19»0 1950 I960 1970 1980 1990 19*0 1950 I960 1970 I960 199
N.E. MEGflLOPOLIS
QUaRT^R ?
NEW ENGLRNO
QUPRTER t 3 QljflRr^R 2
t.
3.
1910 1950 1960 1970 1990 1990 1940 1950 I960 1970 I960 1990
SMOKY MOUNTfilNS
0 .-.u..!........,!.........!.........l..^i^ Q ..-. — ...... .............
1940 1950 I960 1970 1980 1990 1310 1950 1960 L97Q I960
19
-------
oc
GC
OJ
OC
cc
CE
CD
QC
OC
(X
ZD
O
OC
UJ
»—
OC
3.6
<6.6
24-3.0
8-10
>13.3
Figure 10. The spetial distribution of 5-year average extinction coefficients show the drastic increases of quarter 3
extinction coefficients In the Carolinas, Ohio River Valley, and Tennessee-Kentucky area. In the summers of
1948-1952. a 1,000-km size multi-State region centered around Atlanta, Georgia, had visibility greater than
15 mi, which has declined to less than 8 mi by the 1970's. The spatial trend of winter (quarter 1) visibility
shows Improvements in the N.E. Megalopolis region and some worsening in the Sunbelt region. Both spring
and fall quarters exhibit moderate but detectable Increase over the entire eastern United States.
18
-------
EVIDENCE OF TRANSPORT OF
HAZY AIR MASSES FROM SATELLITE IMAGERY
Walter A. Lyons
Mesomet, Inc., Chicago, Illinois
INTRODUCTION
Often, a new way of visualizing natural phenomena helps us crystallize
our understanding of basic physical processes. The ability of conventional
meteorological satellites to routinely observe massive areas of turbidity in
the atmosphere produced by both natural and anthropogenic sources has shed
new light-on long-range pollution transport. This paper briefly summarizes
some of the observations of major aerosol events in the atmosphere and
suggests that routine detection and tracking of synoptic scale pollution
episodes, along with quantitative measurements of their intensity, is entirely
feasible with existing spacecraft and data analysis systems.
Long-range transport of certain long-lived atmospheric pollutants was
graphically demonstrated three decades ago by the tracking of nuclear bomb
debris, in some cases, for several circuits of the globe. The Clean Air Act
of 1967 and 1970, however, had as an underlying assumption the notion that
the concentrations of various primary pollutants at some unspecified, but not
too great, distance downwind from the source would become indistinguishable
from the "natural background" due to dilution, wet and dry deposition, and
perhaps chemical transformation. During the 1970's, we have gradually come
to realize that these chemical transformations produce secondary pollutants
of significant concentrations, which become of greater concern at distances
typically greater than 100 km.
Blumenthal et al.1 measured photochemical oxidants being exported out of
the Los Angeles Basin at the rate of 100 tons/hr. Lyons and Cole2 noted that
high ozone values recorded in both urban and rural sections of Wisconsin were
the combined results of local emissions, mesoscale transport, and synoptic
transport, all comingled in a manner making it most difficult to separate the
19
-------
fraction resulting from each mechanism. Coffey and Stasiuk,3 Wolff et al.,4 and
Ott and Lyons5 were among many suggesting that ozone-laden air masses could
easily travel hundreds or even thousands of kilometers over several days.
Periods of elevated ozone in air masses undergoing long-range transport were
also found to be associated with reduced visibility by Samson and Ragland.6 As
research continued (e.g., project MISTT), it became clear that important
secondary aerosol-producing mechanisms were being found. The conversion of
S02 gas into aerosols within individual power plant or urban plumes was docu-
mented by many, including White et al.7 The progressive deterioration of
regional summer visibility, correlated with increased S02 emissions into the
upper portion of the boundary layer by the greater use of coal and higher
stacks, is reported by Husar.8
Reduced visibility, or visual range, is the result of decreasing contrast
of a distant object against a background by the scattering of light into the
intervening volume of air. It is reasonable, then, to assume that measure-
ments of visible wavelength radiation from an orbital platform would reveal
both increased atmospheric brightness and reduced contrast of ground targets,
if the aerosol were of appropriate size and concentration, and the remote
sensing device were of required sensitivity, resolution, and spectral response.
As early as 1971, Mohr9 noted anomalous areas of turbid air in ESSA
satellite APT images over Europe and suggested the possibility of using con-
ventional weather satellites to monitor regional air pollution events. In the
same year, McLellan10 had some success in using the digital brightness readings
from the geosynchronous ATS satellite to detect the smog buildup in the Los
Angeles Basin by its enhanced upward light scatter.
The prime contributor to widespread areas of turbidity appears to be the
conversion of S02 gas to sulfate aerosols, with transformation rates somewhere
around one to several percent per hour, according to Wilson et al.11 Figure 1
maps the region with the highest S02 emission densities in the United States.
The contributions of the coal-burning power plants, especially in the Ohio
River Valley and surrounding areas, are significant. The isopleths of maximum
recorded sulfate levels (Figure 2) show the pattern centered near the major
source regions, but with values exceeding 20 ug/m3 possible almost anywhere in
the eastern United States. It is in this area that our search for satellite
images revealing large-scale pollution episodes will be focused.
20
-------
SATELLITE DETECTION OF AEROSOLS
Various meteorological satellites have shown themselves capable of
detecting a wide variety of aerosol episodes in the atmosphere. Landsat
images were analyzed by Lyons,12 revealing point source smoke plumes of 100
km plus length over Lake Michigan. Also, Lyons et al.13 noted apparent
indications of high turbidity within entire air masses. Unfortunately, the
small areal and infrequent temporal coverage of Landsat makes synoptic moni-
toring impractical. The same could be said of 35-mm photography from manned
spacecraft missions, although some individual scenes of note were described
by Randerson.14
The operational meteorological satellite, while designed to detect
bright, high-contrast clouds, has proven itself suitable for mapping large-
scale areas of atmospheric dust, haze, and smoke. The Synchronous Meteoro-
logical Satellite/Geostationary Operational Environmental Satellite (SMS/GOES)
series has been used for this research.15 Two systems observe the United
States from 36,000 km above the equator, one centered at 75° W, the second at
135° W longitude. Every 30 min, 4-km resolution (subpoint) infrared (10.5-
12.5 urn) scans of the earth are made from pole to pole. During daylight
hours, visible images at 1.0-km subpoint resolution are obtained at the same
interval. The visual infrared spin-scan radiometer (VISSR) responds to
visible light in the 0.54- to 0.70-um band. Data are transmitted from the
satellite at 8-bit resolution. The enormous data volume is managed in many
ways, but primarily by conversion to photographic form. However, special
computer-based systems do allow investigations to access and manipulate the
data quantitatively.
Figure 3 is an example of a major aerosol event in the atmosphere—a
massive plume of smoke from the Hawaiian volcanic eruption of September 1977.
On subsequent days, the plume drifted in a cohesive manner in the trade winds
for many hundreds of miles. Shenk et al.16 were among the first to note the
cross-Atlantic transport of large clouds of Saharan dust. In Figure 4, a
dust cloud reached Cuba after having left the African coast less than 5 days
before, all the while visible in the SMS/GOES visible imagery. During the
U.S. drought of 1977, several major dust storms occurred in the central
United States. One such storm raged from 23 to 25 February 1977. On 24 Feb-
ruary the satellite detected a "plume" of dust stretching from the Oklahoma
21
-------
Panhandle to Georgia (Figure 5). It was subsequently seen to spread over the
Atlantic as far east as Bermuda. During the passage, local visibilities
dropped as low as 1 to 3 miles. Hi-vol aerosol measurements taken by the
Texas Air Resources Board found 24-hr averages ranging from around 600 ug/m3
near Houston to as high as 2,450 ug/m3 in northeast Texas.
An often overlooked source of manmade pollution is smoke from agricul-
tural burning. As noted by Parmenter,17 massive areas of slash burning in
southern Mexico, Yucatan, and Guatemala during the late winter dry season
generate massive smoke palls over much of the Gulf of Mexico. Figure 6 shows
one of the many such incidents noted in recent years. These smoke-filled air
masses frequently reach the southern United States.
Smoke from giant forest fires in the western United States can routinely
be monitored on SMS/GOES imagery. However, the plume of smoke from a small
northern Minnesota forest fire is shown in its seventh hour (Figure 7) after
having been tracked at 30-min intervals from its inception that morning.
DETECTION OF SULFATE AEROSOL HAZES
Lyons and Husar18 presented an SMS/GOES picture showing a large area of
"haze" covering the midwest. Comparisons with other data showed the area to
be associated with reduced visibility (6 nmi or less), high sulfate concen-
trations, and generally elevated ozone levels. The photograph showed a
"typical" episode: on the scale of 1,000 km, lasting for several days to
nearly 2 weeks, associated with a warm moist air mass that had significant
injections of S02 from major source regions. That particular image first
needed a certain degree of photographic darkroom enhancement to make the
"smog blob" or "hazy blob" or "elevated pollution episode" clearly recog-
nizable.
SMS/GOES image processing is generally geared to highlight high-contrast
cloud systems. Haze is not always easily apparent. A technique has been
developed by NOAA's National Environmental Satellite Service (NESS), as
reported by Parmenter and Anderson,19 to better delineate the subtly differ-
ent radiances of polluted versus "clean" air masses. A digital enhancement
curve, similar to that applied to infrared imagery, was found very useful.
Figure 8 shows an example of the NESS technique. The top panel is a segment
of a 2-km resolution image over Florida, as would be seen on a standard
22
-------
weather office satellite receiver. Application of the digital enhancement
makes the band of turbid air across the central part of the peninsula and
surrounding waters highly visible (middle panel). Reported surface visibil-
ities were under 7 miles within the haze band. Sufficient solar insolation
was reflected to result in significant differential ground heating, as evi-
denced by the supression of afternoon convection beneath the haze band (lower
panel). Another example of the technique was produced on 14 May 1977, when a
"clean" polar air mass descended into the northeast United States, displacing
a stagnant, polluted anticyclone southward (Figure 9). The enhanced image at
1230 GMT (Figure 10) shows a clear demarcation of the two air masses, espe-
cially over the ocean areas. This effective digital technique, however, is
not routinely available from NESS. Thus, researchers studying such phenomena
must rely on darkroom enhancement of prints made from negatives, or special
processing of "live" or recorded digital satellite data on special image
analysis systems. The problem in sulfate haze detection is therefore not one
of satellite sensor capabilities, but one of ground image processing techniques.
CASE STUDIES
The image described by Lyons and Husar18 was taken on 30 June 1975, the
midpoint of an episode lasting from before 25 June until after 5 July. It
was clearly visible on the prints made from SMS/GOES negatives. Figure 11 is
a photographically enhanced picture, 2 km resolution, 29 June 1975. A persist-
ent high-pressure cell over the northeast had caused an accumulation of
secondary pollutants over the lower Ohio River Valley and from Louisiana to
Minnesota. Trajectories made from 600-m winds confirm this overall flow
pattern.20 Of note in Figure 11, in addition to widespread obscuration of
ground features in cloud-free areas, are pockets of "reduced turbidity," as
suggested by lowered image brightness. Film animation shows that these
pockets travel with the general low-level winds, and retain continuity for a
day or more. There is considerable circumstantial evidence relating the
pockets to areas of "wash out" by prior thunderstorms. Knowledge of such
regions from satellite images would greatly aid in analyzing surface and
aircraft aerosol measurements. The volume of air undergoing significant wet
removal can also be calculated from such "convective footprints" if corre-
lated to radar data.
23
-------
Figure 12 summarizes some of the measurements made on 29-30 June 1975.
The visibility contours show that noontime readings (6, 4, 3, and 2.7 nmi)
closely correlate to the haze area seen on the satellite image. Those ozone
monitoring sites reporting 1-hr averages of 160 ppb or higher were located in
or near the "blob." Scattered sulfate measurements showed values as high as
80 ug/m3 within the haze boundary, and generally 15 ug/m3 or less outside.
A major episode occurred from 16 to 28 August 1976. This was discussed
in detail by Lyons et al.21 Figure 13 shows the "blob" on 26 August 1976. It
was particularly "bright" over the Chesapeake Bay area, and extended far
eastward into the Atlantic.
Figure 14 is a map of the 18 GMT, 22 August 1976 weather conditions, with
contours of the 5- and 3-nmi visibilities (corrected for precipitation), and
the boundary of the hazy area seen on the satellite picture. This analysis
was performed for each 13 days of the episode. In virtually all cases, the 5-
nmi visibility contour lay inside the visible boundary of the "haze blob." It
appears that as regional visual ranges drop below 6 or 7 nmi, the haze becomes
"visible" in properly printed SMS/GOES images, especially over water and other
low-albedo targets.
Figure 15 shows the various parameters for 28 August 1976, overlain with
isopleths of National Air Sampling Network (NASN) sulfate data. The areas of
Pennsylvania and New York State, where the visibilities were generally lowest
and the image appeared haziest, were associated with 24-hr sulfate levels
between 30 and 50 ug/m3. All sulfate readings made within the hazy area seen
on the 28 August, 2230 GMT image averaged 22.7 ug/m3, but outside the haze
area readings averaged only 5.7 ug/m3. Similarly, a series of sulfate measure-
ments were made in New York City by Brookhaven National Lab during this
period. When the blob, as seen on the satellite prints, covered the area,
sulfates averaged 23.7 ug/m3, but were only 7.5 ug/m3 when clean air covered
the area. This suggests overall enhancements of sulfate within satellite-
detected hazy blobs of 3.98 and 3.16 times for these two cases. Similarly,
for each of the 13 episode days, the highest hourly ozone measured at all
stations reported in the National Aerometric Data Bank were plotted. The
average for those sites outside the visible blob was 69.4 ppb; inside the
average rose to 90.8 ppb, or 30 percent higher. This suggests that the corre-
24
-------
lation of hazy, low visibility air to sulfates is stronger than to ozone
levels. This is partly due to the apparently greater spatial variability of
ozone, as a result of local NO sources, etc.
/\
As discussed in an earlier paper,21 the SMS/GOES data can be analyzed
quantitatively using such systems as the Man-Computer Interactive Data Access
System (McIDAS) developed by the University of Wisconsin's Space Science
Engineering Center. Digital data tapes from the August 1976 episode were
archived. McIDAS allows total pixel-by-pixel manipulations of the imagery.
All scenes are navigated and gridded, with animation of sequential images
easily accomplished via rapid access from video analog disk. A cursor allows
the operator to select any number of image pixels and measure the digital
brightness counts (DBC, Range 1-256), which can, by application of calibration
procedures, be converted into radiances. Over water, with apparently clean
air masses present (and confirmed by nearby land visibility of 7 mi or more),
DBC's typically ranged from 40 to 45 units. In hazy but otherwise cloud-free
areas, values would usually range from 70 to 85 DBC's. Thus the "signal" of a
hazy air mass in the SMS/GOES data was clearly well out of the noise level.
Over land, typical values in high-visibility polar air masses during August
would be around 60 to 80 DBC's, but could be as much as 50 percent higher when
apparent haze covered the target. By careful selection and calibration of
ground targets, it appears totally plausible to use digital satellite data to
routinely obtain a measure of "atmospheric brightness," and thus of turbidity,
visual range, and to a lesser degree, sulfates. What the technique lacks in
precision it would appear to compensate for in terms of spatial and temporal
coverage—particularly over water where virtually no routine air quality
measurements exist.
CONCLUSION
The current generation of meteorological satellite data in pictorial, and
especially digital, form, is capable of synoptic monitoring of the formation,
growth, and movement of large-scale hazy blobs known to be correlated with
elevated sulfate episodes. Details of the long-range transport and removal
process can also be studied. Figure 17 shows a "front" of haze moving south-
westward along the Atlantic coast on 25 August 1976. This feature was tracked
for 3 days; it finally disappeared on 26 August, after being advected into a
25
-------
line of thunderstorms along the Florida peninsula where its aerosols presumably
underwent washout.
The satellite data allow for rational extrapolation of surface visibil-
ity and sulfate measurements and provide a context for the analysis of air-
craft data. Tracking well-defined structural haze features provides a
possible means for verifying trajectory and transport models. The export of
polluted air masses into, and possibly across, the Atlantic can be studied.
(The!European METEOSAT satellite provides a similar capability in the eastern
Atlantic.) Actual digital processing of the SMS/GOES data allows for the
testing of statistical and physical models relating atmospheric turbidity,
visual range, and sulfate aerosol content.
ACKNOWLEDGMENTS
This research was primarily funded by the Regional Field Studies Office,
U.S. Environmental Protection Agency (Dr. William E. Wilson, Director), under
subcontract No. 68-02-3000, "Conduct Seminar/Workshop on Elevated Pollution
Episodes and Furnish Background Reports to Assist EPA Management in Planning
Future STATE Intensive Field Studies and Related Activities."
REFERENCES
1. Blumenthal, D. L., W. H. White, R. L. Peace, and T. B. Smith. Determi-
nation of the feasibility of the long range transport of ozone or ozone
precursors. Meteorology Research Inc. Report, EPA Contract 68-02-1462,
1974. 92 pp. [NTIS No. EPA-450/3-74-061].
2. Lyons, W. A., and H. S. Cole. Photochemical oxidant transport: Meso-
scale lake breeze and synoptic-scale aspects. Journal of Applied Meteor-
ology. 15:733-743, 1976.
3. Coffey, P. E., and W. N. Stasiuk. Evidence of atmospheric transport of
ozone into urban areas. Environmental Science and Technology, 9:59-62,
1975.
4. Wolff, G. T., P. J. Lioy, G. D. Wright, R. E. Meyers, and R. Cederwall.
An investigation of long-range transport of ozone across the midwestern
and eastern United States. Atmospheric Environment, 11:797-802.
5. Ott, S., and W. A. Lyons. Further evidence of long-range photochemical
oxidant transport inferred from acoustic sounder data. AMS/APCA Joint
Conference on Applications of Air Pollution Meteorology, Boston, 1977.
pp. 33-38.
26
-------
6. Sampson, P. J., and K. W. Ragland. Ozone and visibility reduction in
the Midwest: Evidence of large-scale transport. Journal of Applied
Meteorology. 16:1101-1106, 1977.
7. White, W. H., J. A. Anderson, D. L. Blumenthal, R. B. Husar, N. V. Gillani,
J. D. Husar, and W. E. Wilson, Jr. Formation and transport of secondary
air pollutants: Ozone and aerosols in the St. Louis urban plume.
Science. 194:187-189, 1976.
8. Husar, R. B.. Man's impact on the troposphere: Lectures in tropospheric
chemistry. NASA Reference Publication, No. 1022, 1978. pp. 319-348.
9. Mohr, T. Picture of the month: Air pollution photographed by satellite.
Monthly Weather Review. 99 (8):653, 1971.
10. McLellan, A. Satellite remote sensing of large scale local atmospheric
pollution. Proceedings, Second International Clean Air Congress, H. M.
Englund and W. T. Beery, eds., Academic Press, New York, 1971. pp. 570-
574.
11. Wilson, W. E., R. J. Charlson, R. B. Husar, K. T. Whitby, and D. Blumen-
thal. Sulfates in the atmosphere. Preprints, 69th Annual Meeting of
Air Pollution Control Association, Paper 76-30-06, 1976. 20 pp.
12. Lyons, W. A., and S. R. Pease. Detection of particulate air pollution
plumes from major point sources using ERTS-1 imagery. Bulletin of the
American Meteorological Society, 54:1163-1170, 1976.
13. Lyons, W. A., C. S. Keen, and R. Northhouse. ERTS-1 satellite observa-
tions of mesoscale air pollution dispersion around the Great Lakes.
Proceeding, AMS/WMO Symposium on Atmospheric Diffusion and Turbulence,
Boston, 1974. pp. 273-280.
14. Randerson, D. Quantitative analysis of atmospheric pollution phenomena.
Skylab Explores the Earth. NASA, SP 380, Available Government Printing
Office, 1978. pp. 381-400.
15. Ensor, G. J. User's guide to the NOAA geostationary satellite system.
U.S. Department of Commerce, NOAA, NESS, Washington, D.C., 1978. 101 pp.
16. Shenk, W. E., and R. J. Curran. The detection of dust storms over land
and water with visible and infrared measurements. Monthly Weather Review,
102 (12):830-837, 1974.
17. Parmenter, F. C. Monitoring air quality from satellites. Proceedings,
4th Joint Conference on Sensing of Environmental Pollutants. American
Chemical Society, Washington, D.C., 1978. pp. 254-257.
27
-------
18. Lyons, W. A., and R. B. Husar. SMS/GOES visible images detect a synoptic-
scale air pollution episode. Monthly Weather Review, 104 (12):1623-1626,
1976.
19. Parmenter, F. C., and R. K. Anderson. A satellite overview of inadver-
tent weather modification. Preprints, Sixth Conference on Inadvertent
Weather Modification. American Meteorological Society, Boston, 1977.
pp. 83-86.
20. Teknekron, Brand Niemann, personal communications.
21. Lyons, W. A., J. C. Dooley, Jr., and K. T. Whitby. Satellite detection
of long range pollution transport and sulfate aerosol hazes. Atmospheric
Environment. 12:621-631, 1978.
28
-------
SO2 EMISSIONS * 4
25 TONS/KM2/YEAR
Figure 1. Map of regions where SC>2 emission density exceeds 25 tons/km^/yr.
-------
w
o
(MAXIMUM OBSERVED
SULFATE VALUES
MAY-OCT.,1976
i
Figure 2. Plot of observed maximum sulfate values (jug/m^) at NASN sites during
the period May through October 1976.
-------
Figure 3. SMS/GOES 1.0-km resolution visible image, showing plume of smoke from
Hawaiian volcanic eruption drifting several hundred kilometers south westward,
at 0300 GMT, 16 September 1977.
-------
u>
ro
Figure 4. Full disk SMS/GOES 4.0-km resolution visible image, 1500 GMT, 18 June 1977.
Saharan dust cloud had crossed the Atlantic on 4 days. The position of the leading
edge of the dust on 14 June is indicated by an arrow.
-------
00
Figure 5. SMS/GOES 2.0-km resolution visible image, 1530 GMT, 24 February 1977,
shows cloud of dust stretching from Oklahoma Panhandle to Georgia. Dust
cloud followed cold front many hundreds of miles into the Atlantic.
-------
CO
/
r
Figure 6. SMS/GOES 2.0-km resolution visible image, 1700 GMT, 30 April 1975,
showing large area of smoke over Yucatan and the Western Gulf of Mexico,
the result of widespread agricultural slash burning.
-------
co
en
2200 Z
Figure 7. Biowup of SMS/GOES 1.0-km resoluteon visebSe Jmags, 2200 GMT,
29 September 1976, showing plume of smoke from a smalB forest fire
in northern Minnesota, having drifted about 100 km southeast after
about 7 hr.
-------
Figure 8(a). Portion of SMS/GOES 2.0-km resolution visible image over Florida,
1230 GMT, 15 June 1977.
Figure 8(b). Same scene, but with imagery digitally enhanced using experimental
technique developed by National Environmental Satellite Service.
Band of hazy air now distinct
Figure 8(c). Same region, but at 2030 GMT, 15 June 1977. Haze was sufficiently
dense to apparently suppress afternoon convection, present both north
and south of haze band.
36
-------
OJ
-vl
Figure 9. Surface weather map, 1200 GMT, 14 May 1977.
-------
co
00
Figure 10. SME/GOES 2.0-km resolution visible image, 1230 GMT, 14 May 1977,
specially enhanced by computer processing at NESS. A sharp discon-
tinuity in atmospheric turbidity is clearly found at the front separat-
ing the two anticyclones.
-------
co
Figure 11. SMS/GOES 2.0-km resolution visible image, 2200 GMT, 29 June 1975,
showing massive "smog blob" west of Appalachians to Kansas and from
the Gulf Coast to the upper Great Lakes.
-------
Figure 12. Synoptic map, 1800 GMT, 30 June 1975. Shaded areas represent visi-
bility contours of 6, 4, 3, and 2.7 nmi. NASN sulfate values (jug/m3)
recorded during June 29 and 30 are shown, and triangles mark oxidant
monitors that recorded maximum hourly values in excess of 160 ppb.
-------
Figure 13. SMS/GOES 2.0-km resolution visible image, 1430 GMT, 26 August
1976, showing widespread "smog blob" stretching from the Mississippi
River eastward into the Atlantic Ocean.
-------
Figure 14. Synoptic map, 1800 GMT, 22 August 1976. Stipple represents hazy
area as observed on satellite image. Contours of 5- and 3-nmi visibility
are also plotted.
-------
Figure 15. Synoptic map, 1800 GMT, 28 August 1976. Overlain on the hazy area
outline and 3- and 5-nmi visibility contours are the observed IMASIM sul-
fate readings (pg/m3). Readings in excess of 40 ;ug/m3 were found in
low visibility air in Pennsylvania and New York.
-------
30
20
10
SULFATE VALUES
NEW YORK CITY
AVERAGES
INSIDE BLOB 23.7
OUTSIDE BLOB 7.5
N
N
N
N
N
N
M
i-
o
E
Q
_l
O
O
M
16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
AUGUST 1976
Figure 16. Sulfate values, 24-hr averages, recorded in New York City by Brookhaven
National Laboratory from 16 to 28 August 1976. N indicates smog blob,
as seen on the satellite, was not over the area. Y or E indicates the moni-
tor was in or on the edge of the smog blob.
-------
Figure 17. Portion of 2.0-km resolution visible image, 2230 GMT, 25 August
1976, showing a front of turbid air moving south and west off the
U.S. Atlantic Coast.
-------
REMOTE SENSING OF REGIONAL AIR POLLUTION
FROM SATELLITES
E. J. Friedman* and E. L. Keitz
The Mitre Corporation, McLean, Virginia
Abstract
The potential for regional measurement of gaseous pollutants by use of
remote sensing from orbiting satellites is analyzed. Based on a study of
the capabilities of a wide variety of remote sensors to measure and monitor
air pollution over regional areas, two sensors were selected for possible
satellite application. These sensors are assumed to be orbiting in a typical
Space Shuttle orbit. A number of aspects of the problem are considered
including the data requirements of users, a model for statistical interpreta-
tion of the observations, the influence of orbit parameters on the spatial
and temporal sampling and an example of application of the model over the
eastern United States.
INTRODUCTION
This paper is based primarily on the results of two studies completed
recently. In the first of these an analysis was made of the potential for
remote sensing techniques to measure and monitor air pollution over regional
areas.1 The performance of such instruments was contrasted with that
expected from both ground-based monitors and airborne contact sensors.
Included in the study were evaluations of measurement strategies for the
long-range transport of both an urban oxidant plume and a fossil-fueled
power plant sulfur dioxide/sulfate plume. The study also addressed regional
measurement requirements and the current status of specific remote sensors,
which have a theoretical capability for regional measurement or monitoring.
the second study investigated the role that might be played by the
Space Shuttle in obtaining data that describe the air quality of the eastern
^Speaker.
46
-------
United States.2 A number of aspects of the problem were considered including
a model for statistical interpretation of the observations, the influence of
orbit parameters on the spatial and temporal sampling and an example of
application of the model. This study was limited to the measurement of
gaseous species.
The present paper discusses appropriate parts of these two studies and
focuses on the potential for regional measurement of gaseous pollutants from
orbiting satellites. The overall conclusion is that although the potential
for regional measurement of pollutant gases from satellites is clearly
evident, this potential is unlikely to be realized in sufficient time to be
of assistance in the experiments being discussed at this workshop. Although
one remote sensor (TRW/MAPS) is scheduled to fly on the second Space Shuttle
mission, the mission is experimental only and the characteristics of the
pollution monitoring will be constrained by the overall Shuttle flight
schedule.
MEASUREMENT REQUIREMENTS
A wide variety of scientific and regulatory groups have needs for
measurement of air quality on a regional scale. Not all of these groups,
however, have requirements for measurement of the same species or over the
same four-dimensional network. An extensive study of the users and needs
for local and regional air quality monitoring in the troposphere has been
completed by the National Academy of Sciences.3 This study provides a
comprehensive treatment of the topic. Table 1 presents a summary of the
results showing which of the various users have interests in four basic
categories of needs.
A summary of the air pollutants either recommended for measurement or
actually measured in various regional studies is presented in Table 2.
Care must be exercised not to precisely compare the various species lists
shown in the tables since each study focuses on either slightly or markedly
different objectives. These various objectives are summarized below.
1. National Academy of Sciences3--generalized tropospheric moni-
toring.
2. Sulfate Regional Experiments-investigation of sulfate formulation
from power plant emissions.
47
-------
3. Stanford Research Institute5--remote sensing of regional pollutants.
4. Regional Air Monitoring System/Regional Air Pollution Study6--
generalized urban and regional air quality.
5. Los Angeles Reactive Pollutant Programs-validation of mathemat-
'• ical models of air pollution photochemistry.
6. Environmental Protection Agency/Oxidant Study8--investigation of
rural oxidant levels as related to urban hydrocarbons.
7. Battelle Columbus Laboratories9--formation and transport of ozone
in nonurban areas.
Inspection of the table shows that the power-plant-oriented study is not
concerned with measurements of CO or hydrocarbons because these are not
generally emitted from efficient combustion processes. On the other hand,
studies concerned primarily with rural oxidants have not measured S02 or
related pollutants.
Based on the foregoing information and experience with similar programs,
a set of measurement requirements was developed for monitoring of both the
urban and power plant plumes. Table 3 presents a listing of the major
pollutants and meteorological parameters that should be measured in any
generalized monitoring program. Also shown are estimates of the accuracies
that should be required. Obviously, programs with specific scientific or
regulatory objectives may have requirements for additional or different
parameters.
Tracking of the urban plume oxidant chain primarily requires the measure-
ment of ozone, oxides of nitrogen, and hydrocarbons. Other gases such as
carbon monoxide and acetylene may be monitored as good tracers of the urban
plume's position and extent. Acetylene is particularly attractive since
there are no known natural sources of this substance. Tracking of the
single power plant plume principally requires measurement of those species
directly involved in the sulfur dioxide/sulfur chain. These are sulfur
dioxide, sulfates, and ammonia. Other species having a probably synergistic
effect in the sulfate chain are ozone and nitrogen oxides.
Table 4 presents recommended horizontal grid spacings for typical types
of regional monitoring. As in the case of the pollutants to be measured,
special regulatory or scientific objectives may indicate the need for finer
or coarser grids.
48
-------
Vertical grid spacing may vary from a single surface measurement for
some regulatory situations to as many as 10 or 20 levels for certain pollut-
ants or meteorological parameters in a scientific study. In general the
vertical spacing of measurements should be nonlinear because concentrations
generally change in shorter distances near the surface than at higher levels.
Table 5 presents recommended vertical grid spacings for various types of
regional monitoring.
The time scale required for regional pollutant measurements should
depend on the time scales of the major factors that influence regional
pollution. For urban plume studies, the principal factor influencing temporal
changes in concentrations is the variation in mobile source emissions through-
out the day. Thus changes occurring over a time span of 1 to 2 hr could be
important, particularly in scientific studies. Although the emission
patterns from power plants are not as readily identifiable, a temporal
measurement span of 1 to 2 hr also seems reasonable for regional scientific
studies. For regulatory use, the frequency of measurements should be much
less, particularly in those situations where scientific studies have developed
good correlations between concentrations measured at one specific time and
the overall diurnal pattern. In some cases, a measurement frequency as low
as one per day may suffice. Table 6 presents a summary of recommended time
intervals for typical regional situations.
REMOTE SENSING INSTRUMENT SURVEY
The performance characteristics of six categories of remote sensors
were reviewed to judge their ability to be used in a regional system whether
ground-based, airborne, or satellite-borne. These categories were:
1. Gas filter correlation,
2. Correlation interferometry,
3. LIDAR,
4. Laser heterodyne radiometry,
5. Laser absorption spectrometry,
6. Solar absorption spectrometry.
Specific instruments were analyzed and the outlook for the performance
capability of each over the next 5 years was estimated.
In this report, emphasis will be placed on the TRW/MAPS and the CIMATS
instruments because they were the instruments whose characteristics were
used in the satellite modeling study to be discussed later.
49
-------
Gas Filter Correlation (TRW/MAPS)
The MAPS instrument uses gas filter correlation in order to obtain high
radiation throughput and resolution. The present instrument is the third
gas filter correlator developed under the project (preceded by SR&T/MAPS and
SAI/AAFE/MAPS). A sample of the gas of interest, CO, is placed in a gas
cell located in an optical path carrying radiation from the scene to the
detector. Another, otherwise identical, path traverses an evacuated cell.
The difference provides an indication of the amount of gas of interest
between the infrared radiation source (normally the Earth and its atmosphere)
and the instrument. A further refinement in the MAPS instrument involves
the inclusion of a gas cell at a different partial pressure of the gas of
interest than is contained in the first gas cell. By comparing the different
signals derived from each gas cell and the evacuated cell, the measurement
is less subject to errors caused by the presence of interfering species in
the atmosphere.10 An added potential advantage in the use of two gas cells
of different pressures is the ability to obtain some coarse information on
several vertical layers of the atmospheric path.
Figure 1 shows general features of the MAPS instrument. The reference
blackbody is sampled alternately with the scene radiation to provide a
baseline reference level. The hot and cold blackbodies are adjusted to span
the anticipated temperature range of the scene. They are sampled at a
higher frequency than the scene energy and reference blackbody in order to
allow electronic separation of the signals for subsequent processing. Gain
controls on the electronics are driven by the balance sources so as to keep
the gain of the gas cell paths equal to that of the vacuum path for both
temperature extremes. This procedure effectively balances the system and
reduces the effects of variations of radiance reaching the two detectors.11
For the optical path through the evacuated cell, the modulated energy from
the balance sources is used for radiometric calibration, while that from the
scene and reference is used to determine scene brightness.
In the nadir-viewing (or vertically down-looking) mode, MAPS will
provide data on the total column burden of CO in cloud-free areas of the
Earth. Planned refinements in the choice of gas partial pressures and data
processing may result in data representative of two broad vertical bands in
the atmosphere.
50
-------
Extensive field tests of this instrument were conducted during the
summer of 1977 over Lake Michigan using a Cessna 402 as the platform.
Results of these tests have been given in detail elsewhere.12 13 These
tests demonstrated the ability of the instrument to measure the carbon
monoxide urban plume from Chicago as it progressed over the lake.
Since that time efforts have concentrated on developing an on-board
microprocessor capability for the instrument in an attempt to provide real-
time data. Although the work is not yet completed, progress has been signifi-
cant and this capability should be achieved shortly.
In April of this year the instrument will begin field tests on board
the NASA Convair 990 in the Indian Ocean. Testing will continue for a few
months. Future plans include a firm commitment to fly a satellite-hardened
version of the instrument on board the second Space Shuttle flight for the
measurement of global and regional CO patterns. The earliest launch date
for this mission is March 1980. The flight would be of 4 days' duration
with all MAPS data tape recorded on board rather than telemetered to ground
stations. This will be the first attempt to measure tropospheric gaseous
pollution from a spacecraft.
Correlation Interferometry (CIMATS)
The CIMATS instrument is a two-channel infrared (IR) interferometer,
which uses the same general principles of operation as its predecessor, the
Carbon Monoxide Pollution Experiment (COPE).14 1S 16 The instrument is a
variant of a normal Michel son interferometer. A time-varying path difference
is introduced into one of the optical paths by the rotation of a phase
plate. The other optical path is of constant length (phase). Recombination
of the two paths results in an alternating reinforcement and cancellation,
depending upon the phase difference of the two rays. The result is analogous
to a Fourier transform of the incident radiation. The effect of this design
is to transform the electromagnetic variation at optical frequencies to
electrical signals that can be conveniently processed. The basic components
of the instrument are illustrated in Figure 2. The range of scan (phase
delay) is chosen to optimize the signal-to-noise ratio for a particular
species of pollutant in the atmosphere.
51
-------
In the presence of interfering species, the interferogram does not
provide a unique indication of the species of interest. Table 7 lists the
major interferents and the expected signal-to-noise ratio for each measurable
species for both the reflected solar (Channel A) and the IR (Channel B)
ranges. To reduce interference effects, use must be made of weighting
functions that are chosen to minimize the contribution made by constituents
other than those of interest, while maximizing the signal-to-noise ratio of
the desired species. The weighting functions are determined both theoreti-
cally and experimentally. The best available atmospheric models and spectral
absorption properties are used to construct a "base interferogram." Variation
of the concentrations of various constituents results in a set of different
interferograms. Since each interferogram consists of a set of data points,
the set of interferograms may be considered a set of simultaneous linear
equations which, in turn, may be solved for the appropriate weighting func-
tions. Laboratory experiments are then performed to generate interferograms
based upon actual gas concentrations at a range of partial pressures.
In operation, interference filters limit the range of the radiation
incoming to the instrument to a very narrow band, selected to coincide with
the region where the effects of absorption by the species of interest are
greatest. Interferograms are generated while observing the area of interest
in a nadir-viewing (vertical down-looking) mode; one complete scan requiring
1 to 3 seconds. These interferograms are compared to the stored weighting
functions to determine the best fit of the data. Scale factors may be
applied to the weighting functions in order to facilitate this matching
process. When the correct weighting function is applied to the data, the
result gives the column burden of the specie of interest. In some cases,
CIMATS will require vertical profiles of temperature, pressure, and humidity
to perform calibrations.
The instrument was delivered to NASA/Langley in December 1978 by the
General Electric Company, the prime contractor. Channel A has been partially
tested against laboratory samples of pollutant gases. No testing of channel
B has been done to date. In the near future both channels will undergo
additional laboratory testing and channel A will be ground tested using
reflected sunlight as the source. Further plans are uncertain owing to
economic constraints.
52
-------
Fixed Wavelength LIDAR
Remote sensing instruments that use laser energy in some form generally
fit into three categories. The first of these, in which a fixed wavelength
pulse is transmitted through the atmosphere and the reflected or backscattered
signal is detected, is one type of LIDAR. The second, in which a laser(s)
is used as the local oscillator in a passive tunable remote system, falls
under the generic term of laser heterodyne radiometry. The third, in which
a tunable laser(s) is used in an active remote system, is called laser
absorption spectrometry.
Laser and lidar-type instruments have been used for many years in the
measurement of particulates, aerosols, and clouds. Since the first lidar
measurement of backscattering from tropospheric aerosols in 1963, similar
observations have been made by many investigators and by now are considered
to be routine.5 This discussion is limited to the ruby laser operated by
the LIDAR Applications Section of the Environmental and Space Sciences
Division at the NASA/Langley Research Center.
The NASA/LRC Ruby Laser is a ground-based semimobile instrument mounted
on a truck bed. The operating wavelength is 0.6943 urn in the red portion of
the visible spectrum. The laser operates at 1 pps with a field of view of 2
milliradians and an angular resolution of about 1 milliradian. The data are
digitized at 5 to 100 MHz, which provides a range resolution of about 1.5 m.
Elevation resolution is 1/15 degree. A PDP-11/10 minicomputer controls the
optical scanning and data collection. A "quick look" capability is provided
by a continually refreshed CRT, which provides 16 gray scales for the range-
height indicator display. Polaroid photography is used for hard copy of the
display.
The typical use of the instrument is in monitoring of power plant
plumes through aerosol backscatter. It is estimated that a typical power
plant plume would show the following particulate concentrations at a 10-km
range.
Wind speed Concentration
3 mps 35 M9/m3
10 mps 78 |jg/m3
53
-------
The 10-km range represents the probable greatest range for meaningful measure-
ments of cr or a with the instrument and 15 km is probably the maximum
range for detecting anything at all under typical conditions. The maximum
detectable concentration is not known because it is in part a function of
.the background concentration. In order to get any response, concentrations
on the order of 1-1/2 to 2 times the background are needed.
Laser Heterodyne Radiometry
This section discusses the present status and outlook for the future of
passive tunable remote sensing systems that use a laser(s) as the local
oscillator. Two versions of the instrument are under development by the
Laser and Molecular Physics Branch at the NASA/Langley Research Center. The
first version of the instrument is called the Infrared Heterodyne Spectrometer
(IMS). This instrument is completed and has been undergoing testing.17
Another version of the instrument is planned for Shuttle application. This
version is called the Laser Heterodyne Spectrometer (LHS). Since the LHS
program is similar to the IHS program, only the latter will be discussed
here.
The IHS is a 2-GHz bandwidth multichannel spectrometer being developed
for ground-based and airborne measurements of the concentration and altitude
distribution of tropospheric NH3 and 03 in both the atmospheric emission and
solar occultation modes. The instrument utilizes discretely tunable C130£6
laser local oscillators operating in the 9- to 11-um spectral region.
A number of gases of interest have spectral signature lines in the 9-
to 11-um region of the infrared. These include NH3) 03, D2H4, NH03, H20,
and several of the chlorofluoromethanes. However, after initial studies of
the calculated and measured spectra of the various gases, it was decided to
select ammonia and ozone for initial applications. NH3 is measured in the
10.78-um wavelength while 03 is measured at two wavelengths, 9.75 urn for
solar absorption and 9.66 urn for nadir radiance.
In a Shuttle application, this type of instrument may be capable of
measuring HN03, CIO, NH3, and CC12F2 (Freon 12) in a solar occultation mode.
Calculations based on stratospheric simulations have shown sensitivities as
low as 0.01 ppb for NH3 and 0.005 ppb for Freon 12.
54
-------
Laser Absorption Spectrometry
Three active tunable remote instruments were reviewed. They are:
1. Differential Absorption Remote Sensing (DARS),
2. UV Differential Absorption LIDAR (DIAL),
3. Westinghouse IR LIDAR.
The DARS instrument has been under development by the NASA/Langley
Research Center for a number of years and the basic theory and operating
principles are well known.18 19 The current version of the instrument is
under fabrication at the Jet Propulsion Laboratory. This version uses C02
lasers as both the energy source and the local oscillator. The operating
wavelengths will be in the 9- to 12-um band and should be runable over a 600-
to 800-MHz range. From an altitude of 10 km, surface horizontal resolution
would be a couple of kilometers. The data gathered should provide the total
burden of 03 (at 9.45 urn) and possibly any other species in the 9- to 12-um
band.
The DIAL system utilizes a commercial flashlamp-pumped frequency doubled
dye laser, a 0.25-m diameter receiver telescope, and a minicomputer for
controlling data acquisition and processing. The output of the laser is
typically 100 millijoules per pulse at 15 pps for wavelengths coincident
with the peaks and valleys in the sulfur dioxide absorption spectrum from
0.2960 to 0.3001 urn. The difference in absorption coefficient for each pair
of transmitted wavelengths is monitored using an auxiliary beam from the
laser. Results obtained from the calibration show the linearity of the
system with respect to the measurement of the total burden of sulfur dioxide.
The instrument should be capable of measuring the average concentration of
sulfur dioxide out to a range of 1 km with sensitivities less than 15 ppb.
Quantitative measurements of the average sulfur dioxide were obtained
around a NASA/Langley steam plant. The measurements were made at night at
ranges from 0.8 km to 1.9 km. Sensitivities varied from 10 ppb (at 0.8 km)
to 20 ppb (at 1.9 km), with the maximum observed concentration being 150 ppb
over a 1.2-km range. The maximum range achievable is estimated to be 7
km.20
Another LIDAR system with potential urban and regional application is
being built for NASA by Westinghouse Research Labs.21 Development is also
55
-------
being supported by the Environmental Protection Agency, U.S. Air Force, and
U.S. Army. The instrument utilizes an optical parametric oscillator.5 22
Development of this type of oscillator began in 196123 and the first tunable
system was demonstrated in 1965.24 At the present time it is possible to
obtain tunable.energy through the visible region, over a portion of the
near-UV and over most of the IR between 1 and 20 pm.
The instrument is expected to be delivered to NASA/Langley in the
Spring of 1979 and will undergo laboratory spectroscopy studies.21 The
ranging capability anticipated for the instrument during design and develop-
ment will not be fully achieved. This is due to the discovery that a signif-
icantly greater amount of energy per pulse than previously calculated is
required for an adequate return signal. The optical parametric oscillator
is limited in the amount of energy attainable. Thus, initial field measure-
ment will probably be limited to column content to a maximum altitude of 3
to 4 km. Pollutants currently expected to be measured are HC1, CH4, and a
broad spectrum measurement of hydrocarbons in the 3.5-|jm region. If any
ranging capability is eventually achieved it will probably be limited to a
few broad ranges over a distance of several kilometers.
Solar Absorption Spectrometry
The Radiometry Section of the Measurement Physics Branch at NASA/Langley
Research Center has developed two solar absorption type instruments for
close monitoring of pollution plumes.22 23 In the solar occultation mode
(Videcon), a conventional TV-type camera is used to monitor particulates,
N02, and total opacity. In the sky background mode, a UV videcon is used to
monitor $62- Each of these systems is described below.
e*
In the solar occultation mode, a conventional TV camera is used with
appropriate filters to monitor particulates, N02, and total opacity. The
operational wavelengths are as follows:
Wavelength (urn) Species measured
0.800 Opacity
0.600 Opacity
0.694 Particulates
0.400 Particulates plus N02
56
-------
Measurements made on clean stacks have shown opacities of around 3 to 4
percent with efficient precipitators in operation. Little or no N02 has
been found in these plumes.
The UV videcon is used to measure S02 in power plant plumes by monitor-
ing the sky background in the 0.310-um wavelength band. This measurement is
made on the wing of the 0.280-um St^ peak due to stratospheric ozone inter-
ference. Readings are made on the plume at 0.310 urn and off to the sides at
0.340 urn to establish zero background levels and eliminate interference from
background particulates. Data are recorded on tape and also displayed in
real time on a TV screen. S02 areas show on the screen in varying shades of
gray, depending on concentration.
Measurements to date have shown that S02 bubbles out of the stack
rather than emerging in a smooth steady flow. This allows tracking of
identifiable features in the plume yielding velocity measurements. In
actual practice, these features are tracked by cross correlation techniques
using two instruments with a small vertical separation. The practical lower
limit for S02 measurements is in the 50 to 100 ppm range with an estimated
accuracy of +50 ppm or approximately +10 percent. Velocity measurements are
estimated to be accurate to +20 percent. Measurements have been made at
distances up to 1 mile from the plume.
During the course of this study various other remote sensing instruments
were considered briefly and judged to be unsuitable for regional monitoring
either from aircraft or spacecraft.1 These will not be discussed here.
Instrument Performance Characteristics
Based on the information gathered during the instrument survey it is
apparent that not all of the remote sensors are will suited for regional
monitoring applications. Table 8 presents a summary of the basic informa-
tion in this regard. The basic suitability assumes that any instrument that
only measures CO and/or NO is not suited to power plant plume monitoring.
^
Also shown on the table is the expected operational mode of the instrument
(i.e., airborne or ground-based) and the limitation of the period of operation
(i.e., day and/or night). Obviously passive instruments operating in the
visible or UV are restricted to daylight operation and in some cases require
full sunlight.
57
-------
The various air pollution remote sensors encounter a series of limita-
tions on their operation. Some of these limitations apply to all instruments
while others apply to individual categories of IR, UV, visible, passive, or
active sensors. In addition, each of the instruments has certain limitations
on its capability to measure the surface value of the pollutant or provide
vertical profile data. These topics are discussed below and summarized in
Table 9.
All of the instruments under study require a cloudless line of sight
between the detector and the ground surface background (or sky background
for upward looking instruments). Thus any significant occurrence of clouds,
rain, or fog in the field of view prevents operation of all of the sensors
with the possible exception of the Videcon, which may have a limited capabil-
ity in light rain or light fog. In addition, the passive instruments operat-
ing in the visible or UV require strong sunlight for reflection or scattering
from the ground surface or sky background. Thus, high sun angles are gener-
ally required, which precludes operation during early morning or late evening
periods. Most of the instruments are affected to some degree by aerosols
such as smoke, dust, haze, and the like. The effect may range from negligible
to important as in the case of DARS and the ruby laser. For many of the
instruments the effect of aerosols has not been completely analyzed and
remains unknown.
The passive nadir-looking instruments in both IR and visible wavelength
bands are generally affected by rapid changes in the surface background
reflectivity or emission characteristics. This problem can be minimized by
use of techniques such as automatic gain control, signal integration over a
longer time or ground path, and/or auxiliary measurements. However, all of
the instruments involved perform better when operated with uniform background
reflectivity or background emissions.
In the thermal region, the signal received by a nadir-looking passive
instrument depends strongly on"the temperature difference between the pollut-
ant and the underlying surface.24 Thus little if any signal will be received
from the layers of the atmosphere nearest the surface where the surface and
air temperature are similar. In general, the difference in temperature must
be on the order of a few degrees Celsius in order for an instrument such as
58
-------
MAPS to receive an acceptable signal. This effect is sometimes referred to
as the vertical sensitivity of the instrument. A schematic representation
of this effect is shown in Figure 3. The lowest levels show little contri-
bution due to the small temperature difference while the upper levels have
only a small effect due to low pollutant density. Thus, large variations in
the near-surface profiles or pollutant concentrations have only a small
effect on the total signal received. In order to incorporate this effect
into the data processing, weighting functions must be produced for each
pollutant, preferably over a range of atmospheric conditions. The inability
to accurately compute this weighting function would cause serious errors in
pollutant concentration calculations particularly in the critical near
surface layers.
Many of the remote sensors, particularly the passive ones, require
auxiliary measurements in order to perform the data reduction or for proper
interpretation of the data. Usually it is some form of meteorological data
that is required. The data reduction programs generally require knowledge
of the temperature and/or water vapor profiles in the atmosphere. These
profiles are usually obtained from the nearest National Weather Service
station and this, of course, may introduce significant errors in the result-
ing data. In more precise usage, it may be necessary to require specialized
temperature or water vapor profile measurement at the exact time and location
of the remote measurement.
Other types of auxiliary data are sometimes also required. For example,
in the measurement of CO and CH4 with the COPE instrument, C02 is an interfer-
ing gas whose effect must be calculated and adjusted. This is accomplished
by means of a C02 detector built into the instrument.
All of the passive remote sensors receive incoming signals that have
interacted with the atmosphere throughout the entire path from the radiation
source to the sensor. Since there is no way to obtain ranging information
on these signals without the aid of auxiliary measurements, the received
signal provides only a column content measurements. Thus, it is not possible
to directly measure either the surface value of the pollutant or the value
for any specific altitude or vertical layer. There are two basic approaches
that can be used for obtaining vertical profiles or surface data with such
59
-------
instruments. The first involves the obvious technique of physically moving
the sensor through different altitudes to measure the differences in column
content corresponding to such movement. This technique will provide adequate
vertical information but does not justify the conclusion that the instrument
can provide vertical profiles of the data. Furthermore, its use is restricted
to airborne platforms.
The second method involves the postulation or assumption of certain
information about the vertical distribution of the pollutant such as an
assumed vertical mixing ratio profile. This profile can be used to differen-
tiate the column content data into individual vertical layers. The accuracy
of this method is highly dependent on the accuracy of the assumed auxiliary
information. Up to the present time, attempts to apply this method to
actual remote measurements have only been marginally successful. This
method is applicable to both airborne and satellite platforms.
Vertical or range information from active systems are much more practi-
cal. The accuracy and resolution of such systems depends to a great extent
on ability of the instrument to precisely control the pulse length and time
of the transmitted signal and to precisely measure the arrival time of the
return signal. In the case of pulsed laser systems, range resolution on the
order of a few meters is easily obtainable.
ORBITS
The characteristics of the orbit(s) chosen for the spacecraft observa-
tions of the polluted toposphere will exhibit a considerable influence on
data quality. The altitude is directly related to the spatial resolution of
the observations and determines, along with the orbit inclination, the
spatial and temporal coverage and sampling. Orbit parameters, along with
the launch date, control the solar elevation along the orbit track, which is
important for the proper operation of instruments working in the visible
spectrum. The orbit parameters also determine the repeat cycle of the
orbit, expressed as the number of orbits (or equivalently, the number of
days) between ground tracks that exactly coincide. The repeat cycle controls
the frequency with which observations are made of a particular site.
60
-------
Orbits of particular interest in the near future are those associated
with the Space Shuttle., The orbits that might be expected for Shuttle
missions will range from approximately 250 to 450 km in altitude and from
28.5° to 57° in inclination. Appropriate selection of values from each of
these ranges will result in orbits whose repeat cycles are an integral
number of days. Examples of such orbits are shown in Figures 4 and 5 and
Table 10. The figures illustrate the ground track that will occur for 1- and
7-day repeat cycles. Repeat cycles of 2, 3, 4, 5, and 6 days exhibit similar
coverage patterns that increase in density as the repeat cycle grows.
Conversely, the frequency of temporal sampling drops as the spatial sampling
increases.
One of the orbits (Figure 4) has an altitude higher than one might
expect for Shuttle but is included both because it shows the control of
orbit inclination on latitudinal coverage (no observations above 45° occur
in the example) and because it has a 1-day repeat cycle. In effect, this
example orbit illustrates one extreme in the trade-off between spatial
sampling density and frequency of sampling. The other extreme, for the
proposed 7-day mission duration, is shown in Figure 5. The orbit has a 7-
day repeat cycle, about four times as much spatial sampling density as the
1-day repeat cycle case but only one-seventh the frequency of sampling
(since it takes 7 days to complete the sampling pattern). There are two
factors that, along with the orbit parameters, determine the solar elevation
history along the ground track of the orbit: the local time of the various
observations and the declination of the sun as it varies with the seasons.
The solar elevation is primarily of importance in determining the occasions
in which the sensors using the visible spectrum can be used effectively.
Variation from measurement to measurement of the local sun time is of
significance because it indicates the amount of diurnal sampling that can be
achieved during the mission. If the mission is sufficiently long that each
hour of the day and night is sampled many times, the diurnal sampling is
better than in the case of a sun-synchronous or nearly sun-synchronous orbit
for which all observations at a particular point are made at the same local
time.
61
-------
Seasonal variations in the declination of the sun and the movement of
the orbit plane with respect to the Earth-Sun line determine the range of
latitudes for which observations can be made at various times of the year.
Figure 6 illustrates these limits for a 56° inclination orbit if 30° solar
elevation is to be achieved. The figure shows that missions as long as
approximately 20 days can enjoy the maximum possible latitude coverage in
one hemisphere or the other, as long as the launch date is chosen with care.
The time of day of the launch is also important (in this case it was assumed
that the first orbit crossed the equator at 12:00 noon). The figure also
shows that winter conditions limit the latitudes that can be covered. From
November to March only latitudes up to 30° can be covered in the northern
hemisphere.
MODEL FOR DATA INTERPRETATION
Characterization of the uncertainty in spacecraft observations of air
quality requires information on three factors: (1) the uncertainty in indi-
vidual observations as determined by the performance of the instrument and
data reduction methods; (2) the uncertainty associated with incomplete
sampling of the scene in space and time; and (3) uncertainty associated with
instrument response to spatial and temporal variability in the scene.
Identification of instrument uncertainty (both bias and random) is relatively
straightforward as a result of laboratory calibration and field testing.
Sampling uncertainties are somewhat more complex, however, since they involve
both the sampling characteristics provided by the orbit, as well as the
natural spatial and temporal variability of the pollutants and interfering
clouds. Models of the spatial and temporal response of the instrument are
presented first. From these models, a total response function is developed
and used to compute response errors. The next stage in the analysis is
computation of uncertainties associated with limited sampling. Finally, the
three contributors to the total uncertainty are used to calculate the results
that might be expected in observational programs.
Spatial Impulse Response Function
A number of parameters characterize the capability of a remote sensor
to respond to the variations in the pollution it is attempting to detect.
62
-------
These include:
velocity of the ground track - V (km/seconds)
time constant of the instrument* - ti(seconds)
full angle field-of-view - Q (radians)
altitude of the spacecraft - A (kilometers)
footprint of the field-of-view - D = QA
Each element that enters the field-of-view acts like a unit step
function whose amplitude indicates the absolute amount of pollution in that
element region. During its time in the field-of-view (D/V), that element
produces a response of the form
After the element has moved outside the field-of-view, the response approaches
zero according to
As the footprint progresses along the orbit path, each successive small
area interacts with the changed concentration and alters the output of the
sensor.
If the pollution level is described as g(x), then the contribution from
elements in the field-of-view at any place x is given by
1 I / I \ I 1 I " "• 1 I ^x'
(1)
Similarly, the contribution of the elements that have moved out of the
field-of-view is given by
x-D
1 I . / g(x' )exp (77I~~) dx'- (2)
0
*Assumes the instrument response to a step function grows a 1-e 1.
63
-------
The total response of the sensor thus becomes
X
f
x-D
1
x-D
."If
J J
/ i D
U(x-D)
(5)
Equation 4 is simplified further when expressed in terms of spatial
frequencies. The frequency spectrum Hx (ju^) is the Fourier Transform of
MX):
H
h(x)
dx
(6)
1 -
64
-------
Temporal Impulse Response Function
To determine the temporal impulse response function, the instruments
can be represented by the block diagram below.
IDEAL LOW
IrAuO C LIrfX£*£V
g' (t)/ g" (t)
N
r
SAMPLER
IDEAL HOLD
\ t J
g*(c)
FILTER
i " / \
tl \ C )
0
The ideal low pass filter in the front end of the sensor limits the
maximum frequency content in the pollution distribution to be observed by
the sensor and aids in determining the optimum sampling rate. The bandwidth
Af of this filter is approximately given by
Af * 7—
4T1
where
TI is the electronic time constant of the instrument.
The output signal g'(t), which is limited to the frequency band 0 to
Af, is sampled at a constant frequency of fo =f-» where f0>2Af. This
: ' 0. ~
sampling rate is required because to transmit a band-limited signal of
duration Tj and maximum frequency f , it suffices to send a finite set of
2f Tj independent amplitude samples obtained by sampling the instantaneous
amplitude of the signal at a regular rate of 2f samples/second. This is a
statement of the so-called Nyquist criterion.
g"(t) is a train of impulses given by
T / '
VT
0
g"(t) = /, g (nTQ) 6 (t-nTQ)
n=o
where 6 is the Dirac delta function.
TO is the sampling interval and as discussed above is related to Af and
the electronic time constant by
65
-------
T -
o
g*(t) is a sampled function consisting of short pulses, Flf F2, F , of
equal duration t3 at equal intervals TQ. Thus, the impulse response function
of the hold circuit is given by
h0(t) • U(t) - U(t-T ) for all TZ< T .
(9)
The last filter of the system gives a real time response and reflects
the time constant (TI) of the system. Its impulse response function is
given by
~t/Tl
h' (t) - e .
Finally, the convolution of h0(t) with h'0 (t) determines the temporal
impulse response function. The temporal impulse response function h2(t) is
given by
h (t) - h0(t) * hg(t)
U(t-T ) for all T
-------
Total System Impulse Response Function
The pollution distribution is a function of time as well as spatial
variations. In general, because of meteorology, the pollution distribution
at any point x2, at time t2, may be influenced by the pollution distribution
of point Xi at time t^ (tj < t2). The instrument responses in space and
time, as derived above, can be considered as independent of one another.
Thus, the total system impulse response function represented as h(x,t)
is given by
h(x,t) = h (x) • h (c)
-X/VT
1
-(x-D)/Vr
U(x) -
1
U(x-D)
(13)
U(t) - T
l
Taking the double Fourier Transform of Equation 13, we obtain the system bi-
frequency impulse response function as
33 SO
//• 1 Jww*% *•
J h(x,t) e • e ~ dx dt
-00 -00
(14)
1-e
• T,
1-e
"Jw2T2
This formulation is applied to instrument performance evaluation in the next
section.
Instrument Performance Evaluation
Let gi(x,t) be the input pollution distribution at any x, and t and
g2(x,t) be the instrument response to this input. To determine how accurately
the original distribution is reproduced, a criterion relating the average
power in the input and output pollution distributions may be used. The
average power of a random process x(t) is given by
67
-------
{"*<«}
where E is the expectation operator. The difference between the average
power of gi(x,t) and g2(x,t) will be given by
e =
The power spectrum or a spectral density Si (uji.u^) of the process
gi(x,t) is the Fourier Transform of its autocorrelation:
30 30 ;
~Ju
Rg<$l' V
, , w9
V-l-2 _
-00 -00
where
g
From Equation 16, using the Fourier inversion formula, it follows that
oo oo e • c
i r r ^wi i ^"22
\ (^1'V " ~"~2 / / Se (wi'U)'>) e e dwi dw?' (18)
1 ZL TT //I
-oo -oo
With |i = 0, and |2 = 0> tne above yields an expression for the average
power of the input distribution
00 00
du. dw .
Then the difference e becomes
03 00 00 O3
— r r )d» ~i c f-
4ir2 J J 8j 1' 2 1 2 4ff2 ^ J
s
4irfc 7 7 sl * * . * - 4TT* 7 J S2
— O3 —00 ^03 —00
68
-------
The spectral density S 2 (wi,u)2) of the output g2(x,t) is dependent upon the
y
system impulse response function and the spectral density of the input. It
is given by
So that
H( JW, > jun'
Sg1(wl'w2)
dw
— 7 I I H(jw . ,JUJ
(21)
(22)
and the percent difference is given by
co
% £ » 100
1-
L ~
J J |H(jw.,jwJ
-co —oa 1 1 J~2
"r "r s (w a.
/ / %S-2
CO -CO
1
2
21 I>tu2 Wl W2
> du du .
(23)
The difference of Equation 23 can be used as a part of the uncertainty
analysis.
Uncertainty Due to Sampling Strategies
One of the key issues in determining the performance of remote sensors
in space is the determination of the uncertainty in assessing pollution
distributions because of the sampling distribution in space and time.
The sampling problem arises because the selection of the orbit and the
inhibiting influence of clouds may make the spatial and temporal sampling
less than adequate for the purpose of characterizing the spatial and temporal
variability of the pollution.
For a scene with little spatial variability, a few observations at
different points may be adequate to obtain statistically meaningful averages.
The problem at hand, unfortunately, is characterized by a high spatial and
69
-------
temporal variability so that sampling limitations become important. It is
the purpose of this subsection to quantify the sampling by considering both
that which can be achieved as well as the spatial and temporal variability
in the pollution distribution.
The pollutants of interest may exhibit considerable spatial and temporal
variability. In fact, different pollutants may exhibit different patterns
of diurnal, seasonal, and spatial variability. The pollution can therefore
be regarded as a stochastic variable, with the pollution variations through
each day being a record from the ensemble of all possible records. Let u(t)
represent the ensemble mean of P.-(t), the pollution at time of the day, t on
+• K
the i day. Then, the average pollution for N days at time t is
N
Pi(t) (24)
Note that |j(t) is not independent of time because of the diurnal varia
tions and, therefore, the process is a nonstationary stochastic process.
Integrating Equation 24 over the times of the day that instruments are
capable of taking measurements, we obtain the time average pollution for N
days. This is given by
H» — / (i(t)dt . (25)
/
time of the day
Since Shuttle will be visiting one place only a finite number of times
in N days, Equation 25 is rewritten as a sum:
(26)
where
t. is the j hour
At
At - t..,-t.
70
-------
At will be determined by the orbit parameters since they control the
time between passes over the same place. Ideally, one would like to keep
At as close as possible to the sampling interval given by Nyquist criterion,
which may be impossible to achieve considering large variations in pollution
diurnal cycles. For example, if the temporal variability of the scene is on
the order of a day, which is not uncommon, the Nyquist criterion requires
that samples be made on a scale of fractions of a day. It is unlikely that
samples can be made with this temporal density. Therefore, the estimate of
the average pollution at time t., y(t.) may only be available from a single
th ** ^ th
measurement taken at the j hour of the i day, P-(t-). Let the variance
of this single estimate be a, defined as
(27)
Var
then the estimate of the ensemble average over N days will be given by
j-l J (28)
and the variance of this pollution estimate, j}, will be
Var [] - -2 T ^(t:
(29)
i-1, J-l
where p(i-j) is the correlation between the pollution estimates u(t.) and
(j(t.) made at different times.
This analysis can be carried further using the following assumptions:
1. The samples over the period are uniformly distributed over the
time of the day,
71
-------
2. The variance in any pollution estimate is a2 and,
3. The correlation between pollution estimates can be assumed to be
of exponential form, i.e.,
P(i-j) = .-tal
where
k is the index representing the k repeat cycle, and takes values
between 0 and n.
n is the number of repeat cycles in a time period of interest and
equals the number of samples of a place in that time period.
a1 is the correlation coefficient/repeat cycle.
If M represents the number of days/repeat cycle, and a is assumed to be
the correlation coefficient for one day, then
a1 = Ma.
Under these conditions, Equation 29 becomes:
Var ((1) - £- + *4
n 2
n2
(31)
Equation 31 can also be used to generate a set of curves for different
values of n, M, and a1, which illustrate the variance in estimates (in units
of a) for various mission durations, repeat cycles, and pollution correla-
tions. Curves developed from this equation are shown in Figures 7, 8, and
9. The mission durations considered in the examples are 7, 15, and 30 days.
The data pertaining to 15- and 30-day missions are included for comparison to
illustrate the value of missions longer than the expected 7-day Shuttle
missions. The correlation of pollution on 2 successive days is indicated by
p. The correlation coefficient, p, for 1 day is determined from using Equa-
tion 30, and is given by
72
-------
a = - ln(p)
(32)
Upon examination of the figures it is clear that for a very high correlation
of pollution of 2 successive days, the uncertainty in the estimate of pollu-
tion in units of a is not significantly reduced by making a large number of
samples. On the other hand, for a highly varying pollution scene, the
correlation of pollution on 2 successive days is low, and the potential
reduction in the uncertainty in the estimate of pollution provided by a
single observation is more pronounced.
Application of The Analytical Methods
The statistical tools developed in the previous sections can now be
used collectively to determine the data quality that might result from a
real mission. For example, a typical Shuttle orbit might have a 5-day
repeat cycle. This information, along with data on the correlation of
pollution and the probability of cloud cover, can be used to estimate the
sampling uncertainty. The inability of certain instruments to operate under
conditions other than relatively high solar elevation angles can also be
included.
Pollution Statistics--
An extensive listing of air quality statistics for New York City were
obtained from the Environmental Protection Agency. The data were provided
in the form of hourly and daily averages of pollution levels for July and
August 1974. From these data, weekly, biweekly and monthly means and standard
deviations were computed using conventional statistical methods. In addition,
the correlations were also computed. Data for one monitoring site appear in
Table 11. The correlation computation was carried out using the method
normally used in regression analysis. That is,
x.
n x. -
\2H
x. L
i J;
(33)
73
-------
where all sums range from 1 to n, the number of entries (for example, for a
30-day mission the number of entries will be 30). Thus, the coefficient p
represents the correlation between each measurement and the next successive
one.
Use of the Data—
The standard deviation and correlation data of Table 11 have been used
to compute the sampling uncertainty as a function of days per repeat cycle.
The uncertainty due to sampling is
x 100 (34)
Use of Equation 34 and Table 11 results in Table 12, which is an example
of the magnitude of sampling uncertainty that might occur in attempting to
make estimates of the pollution in the New York City area, assuming that an
IR sensor, unaffected by solar limitations, is used. A 50 percent chance of
cloudiness was assumed. The sampling uncertainty for a mission of the
expected duration (7 days) ranges from 20 to 36 percent, depending on the
number of days in the repeat cycle of the orbits. However, for a mission of
a month, the sampling uncertainty is reduced to the range of 10 to 22
percent, depending on the number of days per repeat cycle. As expected, the
1-day repeat cycle provides the smallest sampling error. However, the orbit
that produces a 1-day repeat cycle is also the one that results in the
poorest areal coverage (as illustrated in Figure 4).
Total System Performance —
Three major elements make up the total uncertainty in estimates of
pollution. They are the radiometric uncertainty imposed by the instrument
performance, uncertainties dictated by the spatial and temporal response
functions of the instrument, and the sampling uncertainty. Under the assump-
tion that these uncertainties are independent, and they appear to be, then
the data quality is defined by a root sum of squares calculation. The total
uncertainty is given by
1/2
uncertainty
cr
3
(35)
74
-------
where L is the number of observations during the mission
ov uncertainty in individual estimates.
a2: uncertainty associated with response.
a3: sampling uncertainty (derived from Table 12).
Table 13 illustrates the total system uncertainty for various mission
lengths and repeat cycles. As in the data of Table 12, 50 percent cloud
cover is assumed. Other assumptions used in the analysis are that each
pollution estimate has an uncertainty of 50 percent and that response uncer-
tainties may be as high as 10 percent, based on the use of Equation 23 and
typical pollution distributions.
An expected result appears immediately; mission duration has a signifi-
cant impact on the uncertainties, the shorter repeat cycle orbits exhibiting
the best data quality. Conversely, these short repeat cycle orbits have the
poorest density of sampling of the region of interest. The mission designer
.has the choice, then, of attempting to achieve the best possible data but
for limited locations. Alternatively, he may choose to gain the best possible
spatial sampling of the region at the expense of data quality at individual
locations.
Clearly, a considerable number of assumptions have been included in the
analysis above (frequency of cloudiness, uncertainty of individual pollution
estimates, etc.). It was the object of this analysis to illustrate the
methods to be used in analyzing sensor/orbit performance. Additional analysis
would be required if the assumptions were to be modified.
REFERENCES
1. E. L. Keitz et al. The Capability of Remote Sensing for Regional
Atmospheric Pollution Studies. The MITRE Corporation, MTR-7267, January
1977.
2. E. Friedman et al. Shuttle Applications in Tropospheric Air Quality
Observations. The MITRE Corporation, MTR-7628, August 1978.
3. Practical Applications of Space Systems, Supporting Paper 7,
Environmental Quality. The Report of the Panel on Environmental Quality
to the Space Application? Board of the Assembly of Engineering.National
Research Council, National Academy of Sciences, Washington, 1975.
4. G. M. Hidy et al. Design of the Sulfate Regional Experiment (SURE),
Volumes I-IV. Environmental Research and Technology, Westlake Village,
California, February 1976.
75
-------
5. M. L. Wright et al. A Preliminary Study of Air-Pollution Measurement by
Active Remote-Sensing Techniques. Stanford Research Institute, NASA
CR-132724, June 1975.
6. R. L. Myers and J. A. Reagen. The Regional Air Monitoring System, St.
Louis, Missouri, USA. International Conference on Environmental Sens-
ing and Assessment, September 1975, Las Vegas, Nevada, Institute of
Electrical and Electronics Engineers, New York, 1976.
7. R. B. Evans. Aerial Air Pollution Sensing Techniques. Proceedings,
Second Conference on Environmental Quality Sensors. National Environ-
mental Research Center, Las Vegas, Nevada, October 1973.
8. . Investigation of Rural Oxidant Levels as Related to Urban
Hydrocarbon Control Strategies. Prepared by the Research Triangle
Institute for the U.S. Environmental Protection Agency, March 1975.
9. C. W. Spicer et al. Final Data Report on the Transport of Oxidant
Beyond Urban Areas. EPA Contract No. 68-02-2241, January 16, 1976.
10. TRW, Inc. Monitoring Air Pollution from Satellites (MAPS). NASA CR-
145137, Final Report, 1 March 1977.
11. T. W. Ward and H. H. Zwick. Gas Cell Correlation Spectrometer:
GASPEC. Applied Optics. XIV:2896, December 1975.
12. S. M. Beck et al. Aircraft Instrumentation System for the Remote
Sensing of Carbon Monoxide. Proceedings, Fourth Joint Conference on
Sensing of Environmental Pollutants.November 1977, New Orleans, LA,
American Chemical Society, Washington, D.C.
13. J. C. Casas et al. Procedures Utilized for Obtaining Direct and
Remote Atmospheric Carbon Monoxide Measurements Over the Lower Lake
Michigan Basin in August of 1976. Proceedings, Fourth Joint Conference
on Sensing of Environmental Pollutants.November 1977, New Orleans, LA,
American Chemical Society, Washington, D.C.
14. M. H. Bortner et al. Analysis of the Feasibility of an Experiment to
Measure Carbon Dioxide in the Atmosphere. NASA CR-2303, October 1973.
15. R. Dick et al. Development of an Engineering Model Correlation Inter-
ferometer for the Carbon Monoxide Pollution Experiment. General
Electric Company, Philadelphia, Pennsylvania, November 1974.
16. M. H. Bortner et al. Carbon Monoxide Pollution Experiment: Final
Report. NASA CR-132717, January 1975.
17. R. K. Seals, Jr., and B. J. Peyton. Remote Sensing of Atmospheric
Pollutant Gases Using an Infrared Heterodyne Spectrometer. Presented
at the International Conference on Environmental Sensing and Assessment,
Las Vegas, Nevada, September 14-19, 1975.
76
-------
18. L. J. Duncan et al. Airborne Remote Sensing System for Urban Air
Quality. The MITRE Corporation, MTR-6601, February 1974.
19. R. K. Seals and C. Bair. Analysis of Laser Differential Absorption
Remote Sensing Using Diffuse Reflection from Earth. Presented at the
Second Joint Conference on Sensing of Environmental Pollutants, December
10-12, 1973, Washington, D.C.
20. J. M. Hoell, Jr., et al. Remote Sensing of Atmospheric S02 using the
Differential Absorption LIDAR Techniques. Presented at the International
Conference on Environmental Sensing and Assessment, September 14-19,
1975, Las Vegas, Nevada.
21. E. Browell, private communication, NASA/Langley Research Center, March
1979.
22. R. J. Exton, private communication, NASA/Langley Research Center, April
1976.
23. R. J. Exton and R. W. Gregory. A Four Channel Portable Solar Radiometer
for Measuring Particulates and/or Aerosol Opacity and Concentration of
N02 and S02 in Stack Plumes. NASA-TN-D 8182, June 1976.
24. H. G. Reichle. MAPS (Measurement of Air Pollution for Satellites).
NASA Langley Research Center, April 1973.
77
-------
RADIATION %~v«^
i
SAMPLE CELL
(Containing
Pollutant)
EMPTY
CELL
BLACK BODY
CALIBRATION
SOURCE
i
MIRROR
(Pivots for Calibration)
FILTER
1 1 DETECTOR
Figure 1. Gas cell correlation analyzer.
1
1
i \
1. A ;
1 v 5
1 - Entrance Apertur
2 - Lens
3 - Beam Splitter
4 - Adjustable Mirro
5 - Lens
6 - Detector
: — 1 , 7 - Moving Plate
' 8 - Mirror
— •
8
e
r
Figure 2. Correlation interferometer.
78
-------
Tropopause
UJ
o
Difference between
Surface and Air Temperature (AT)
Absolute Concentration
of Pollutant (C)
Signal Produced
f(CAT)
Figure 3. Schematic of effect of the difference between surface and air temperatures
on the instrument signal.
79
-------
00
o
44
43
42
/ -i
41
0)
T3
3
4->
•H
•U
39
38
37
36
35
I
85 84 83 82
81
80
79 78 77
West Longitude
'Indicate which orbit of the cycle is responsible for each ground track.
76
75
74 73
Figure 4. Ground track for orbit with 1-day repeat cycle
(altitude = 492 km, inclination = 45°).
72
71
70
-------
00
45
44
43
42
-------
00
RANGE OF LATITUDE AND TIME
OF YEAR FOR WHICH SOLAR
ELEVATION EXCEEDS 30°
200
DAYS FROM MARCH 21
I, , , \
Figure 6. Latitudes covered by a typical 56° orbit with the requirement that the,
solar elevation exceeds 30°. ; !
*. . i
j ' t i
/ : : / ; ;,
-------
1 2 3 4 5
No. of days in a repeat-cycle (M)
Figure 7. Errors due to sampling in a 7-day mission.
83
-------
'0.9.
07£T
0.7
o
CO
. J-l
•H
•s
R)
•H
0.3
0.2
0.1
0.85
0.05
2345
No. of days in a repeat cycle (M)
Figure 8. Errors due to sampling in a 15-day mission.
84
-------
1.0
0.9
^-r—tTT
0.7
O
01-
o..s~
(0
0.3
0.2
0.1
p = 0.95
0.85
2.3 4 5
No. of days in a repeat cycle (M)
Figure 9. Errors due to sampling in a 30-day mission.
85
-------
TABLE 1. USERS AND NEEDS FOR LOCAL AND REGIONAL AIR QUALITY
MONITORING IN TROPOSPHERE
00
(Tt
BASIC NEEDS
Near surface Regional air Predictive Major point source
Principal users sensing (local) quality data model inputs emissions
U.N. organizations
Federal agencies
EPA
AEC
NOAA
DOT
HEW
NSF
State and local agencies
Regulatory
Planning
Resource
Regional air pollution control boards
Environmental consultants
Engineers and architects
Research and scientific investigators
Medical and nursing professions
Public-interest groups
Industrials
News media
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
TABLE 2. SPECIES RECOMMENDED OR ACTUALLY MEASURED IN VARIOUS REGIONAL STUDIES
00
Species
Carbon monoxide (CO)
Formaldehyde (H2CO)
Halogens
Aanonla (NH3)
Sulfur dioxide (S02)
Hydrogen sulflde (H2S)
Nitrogen oxides (NOX)
Peroxyacetyl nitrates (PAN)
Hydrocarbons
Ozone (03)
Mercury (Hg)
Nitric oxide (NO)
Nitrogen dioxide (N02)
Aerosols or parti culates
Methane (CH4)
Hydrogen chloride (HC1)
Nonmethane hydrocarbon
(NMHC)
Total sulfur (S)
Acetylene (C2H2)
Ethylene (C2H4)
Ethane (C2H6)
Fluorocarbon-11 (CC3F)
Carbon tetrachloride
(CC14)
Recom-
mended
X
X
X
X
X
X
X
X
X
X
X
MAS SURE
Accu- Recom- Accu-
racy mended racy
10 ppb
1 ppb
1 ppb
10 ppb X NS*
10 ppb X NS
0.1 ppb X NS
10 ppb
10 ppb
1 ppb
10 ppb X NS
10 2 ppb
X NS
X NS
X NS
SRI RAMS/RAPS
Recom- Accu- Meas-
mended racy ured
X 10 ppb X
X 1 ppb
X 10 ppb
X 10 ppb X
X
X est 1 ppb X
X 10 ppb X
X
X
X 0.5 ppm X
X 0.5-1 ppm
X
Accu-
racy
NS
NS
NS
NS
NS
NS
NS
NS
NS
LARPP EPA/OXIDANT BATTELLE
Meas- Accu- Meas- Accu- Meas-
ured racy ured racy ured
X NS X
X NS
X NS X NS X
X NS X NS
X NS X NS X
X NS X NS X
X NS X
X NS
X NS X
X NS X NS X
X NS X
X
X
X
X
Accu-
racy
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
NS
*NS: Not stated.
-------
TABLE 3. REGIONAL MEASUREMENTS REQUIRED
Carbon monoxide
Ammonia
Sulfur dioxide
Hydrogen sulfide
Nitrogen oxides
Nitric oxide
Nitrogen dioxide
Total hydrocarbon
Methane
Acetylene
Ozone
Aerosols or parti culates
Sul fates
Surface and vertical temperature
Surface and vertical humidity
Surface wind velocity
Vertical profile of wind
Estimated
accuracy
10 ppb
10 ppb
10 ppb
0.1 ppb
10 ppb
10 ppb
10 ppb
1 ppb
0.5 ppm
0.05 ppb
10 ppb
25 ug/m3
0.5 ug/m3
2° C
10%
10°; 2 mps
10°; 2 mps
Urban
plume*
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Power
plant
plume
X
X
X
X
X
X
X
X
X
X
X
X
X
*It is assumed that the urban plume contains one or more power plant plumes.
88
-------
TABLE 4. RECOMMENDED HORIZONTAL GRID SPACING
Distance from
source
25 m
25-50 km
50 km
General
regulatory
25-100 km
25-100 km
50-100 km
Type of monitoring
Urban plume
scientific
5-10 km
10-25 km
25-50 km
Power plant
scientific
1-5 km
5-10 km
10-25 km
TABLE 5. RECOMMENDED VERTICAL GRID SPACING
Parameters
Pollutants
only
Nonsurface
weather
General
regulatory
1-4 levels*
Use NWS
data
Urban plume
scientific
5-10 levels*
15-25 levels,
surface to
1 km above
mixing layer
Power plant
scientific
5-10 level sf
15-25 levels,
surface to
2 km
*Decreasing logarithmically to 1 km above mixing layer.
tPoisson distribution starting at surface with maximum at plume level.
89
-------
TABLE 6. REGIONAL TEMPORAL MEASUREMENT REQUIREMENTS
Type of monitoring
Parameters
Urban plume or
General power plant
regulatory scientific
Pollutants and 1 to 4 per 6 to 24
surface day day
meteorology
Upper air Use NWS 4 to 12
meteorology data day
per
per
TABLE 7. INTERFERENTS AND ANTICIPATED SIGNAL-TO-NOISE
Channel
A
A
A
A
B*
B*
B*
RATIOS
Species
CO
CH4
N20
NH3
CO
N20
S02
[S/N] FOR
Assumed
background
0.1 ppm
1.5 ppm
0.2 ppm
1.0 ppb
0.1 ppm
0.2 ppm
2.0 ppb
CIMATS MEASUREMENTS
Center
wavelength
Cum)
2.33
2.33
2.27
2.23
4.62
4.62
8.70
[10]
Interferents
CH4
C02
CH4
C02
CH4
C02
C02
CH4
03,
, H20
, M
, NH3
, H20
, H20
, H20
, H20
, N20
H20
S/N
20:1
40:1
10:1
2:1
2:1
2:1
Not
avai 1
able
*S/N ratios of thermal channel are based on the assumption of no indepen-
dent measurement of temperature profile of the atmosphere. Availability
of such measurements will increase S/N ratios.
90
-------
TABLE 8. SUITABILITY OF REMOTE SENSORS FOR REGIONAL MONITORING
Instrument
SR&T/MAPS
TRW/HAPS
TRW/HAPS
COPE
CIMATS
(visible)
CIMATS
(IR)
CIMATS
(IR)
OARS
DIAL
Westlnghouse
IR LIDAR
IHS (solar)
IHS (nadir)
IHS (nadir)
Ruby Laser
UV Videcon
Videcon
Typical
Urban Power plant
plume plume
X
X
X
X
X
X
X
X X
X
X X
X
X
regional use
Power plant Species
close in Not suited monitored
CO
CO, CH«
NHa
CO, CH4
CO, N20, CH4
NH3
CO, N20
S02
03
X S02
HC1, CH4,
HC
X 03, NH3
03
NH3
X Particulate
X S02
X Particulate,
Operational mode
Ground-
Airborne based Satellite
X X
X X
X X
X
X X
X X
X X
X X
X
X
X X
X
X
X
X
X
N02 X
Period of
operation
Day
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Night
X
X
X
X
X
X
X
X
X
X
X
*The current version of the instrument is not engineered for this measurement.
-------
TABLE 9. SUMMARY OF..SPECIEIC INSTRUMENT LIMITATIONS AND CAPABILITIES
rs»
SR&T/MAPS
TRW/MAPS
COPE
CIMATS
(visible)
CIMATS
(IR)
DARS
DIAL
Westlnghouse
IR LIOAR
IHS (nadir)
Ruby Laser
UV Videcon
Videcon
Sensor
Cloudless
High sun line of
angle sight re-
required quired
X
X
X X
X X.
X
X
X
X
X
X
X X
X
limitations
Uniforn
Hinimun back-
of ground
aerosols desir-
required able
X
X
X
X
X
X
X
X
Vertical
sensi-
tivity
Inpor-
tant
X
X
X
X
X
X
Auxiliary data
Tenperature Water vapor
profile profile
required required
X X
X X
X X
X
X
*'
Vertical measure-
Bent capability
Limited Precise
Column ranging ranging
content or pro- or pro-
only file file
X
X
X
X
X
X
X
X
X
X
X
X
Surface measure-
ment capability
Requires
assumed
nixing
No practical ratio
capability profile
X
X
X
X
X
X
X
X
X
X
Measures
layer
near sur-
face
X
X
-------
TABLE 10. ORBIT DATA
Altitude
(km)
492
349
399
277
245
Inclination
(°)
45
56
56
57
57
Repetition
factor
15.00
15.50
15.33
15.75
15.86
Repeat
cycle
(days)
1
2
3
4
7
Nodal
movement
per
orbit
24
23.23
23.48
22.86
22.70
Orbits
per
repeat
cycle
15
31
46
63
111
TABLE 11. CARBON MONOXIDE POLLUTION DATA FOR A TYPICAL URBAN
LOCATION (FOR AUGUST 1974 AT
GREENPOINT AVENUE, NEW YORK, NEW YORK)
Averaging period
(days)
Mean*
(ppm)
Standard deviation
(ppm)
Correlation
7
15
30
3.36
3.34
3.35
1.20
1.19
1.20
0.46
0.45
0.45
93
-------
TABLE 12. PERCENT SAMPLING UNCERTAINTY IN NEW YORK AS A
FUNCTION OF DAYS PER REPEAT CYCLE
(INCLUDING 50 PERCENT CHANCE OF CLOUDINESS)
Percent error*
Mission duration (days)
Repeat cycle (days) 7 15 30
1
2
3
4
5
6
7
'Equal to .*r^R ** v 100
20 . 14
25.16
30.32
35.77
35.77
35.77
35.77
14.25
17.41
20.79
23.91
26.72
29.26
31.61
10.22
12.38
14.73
16.92
18.89
20.69
22.35
TABLE 13. TOTAL UNCERTAINTY (IN PERCENT) IN POLLUTION
ESTIMATES ASSUMING 50 PERCENT CHANCE OF CLOUD COVER
AND 50 PERCENT UNCERTAINTY IN INDIVIDUAL POLLUTION ESTIMATES
Mission duration (days)
Repeat cycle (days) 7 15 30
1
2
3
4
5
6
7
27.9
37.1
45.1
52.6
56.0
59.2
62.3
19.4
25.5
30.9
35.6
39.8
43.5
47.0
13.8
18.1
21.8
25.2
28.1
30.8
33.3
94
-------
CHARACTERIZATION OF REGIONAL SULFATE/OXIDANT EPISODES
IN THE EASTERN UNITED STATES AND CANADA
E. Y. Tong*, M. T. Mills, B. L. Niemann*
Teknekron, Inc., Waltham, Massachusetts
L. F. Smith
Office of Energy, Minerals and Industry, U.S. Environmental
Protection Agency, Washington, DC
Abstract
Current national interests in developing air quality modeling and pre-
dictive procedures have resulted in several analyses and measurement programs.
The test of these predictive techniques must ultimately be based on their
ability to simulate field measurements. This paper contains results of synop-
tic sulfate/oxidant analysis to provide a regional characterization of the air
quality and aerometric features over the eastern United States and Canada.
This regional analysis shows evidence to support the co-occurrence of elevated
concentrations of atmospheric particulate sulfates and ground-level ozone.
Furthermore, it was found that these "warm episodes" are not only related to
the trajectories of the polluted air parcels, but also to certain easily iden-
tifiable synoptic features at the surface and aloft. The most outstanding
episode is related to the eastward movement of occluded frontal systems and
the circulation around the Bermuda High. Within this warm and humid sector,
recirculation may occur, resulting in high sulfate/ozone concentrations within
a relatively small region. This circulation may be a subsynoptic system em-
bedded in a subcontinental polluted air mass.
INTRODUCTION
A meaningful assessment of the extent and magnitude of regionally ele-
vated concentrations of particulate sulfate and oxidant calls for the collec-
tion and analysis of large volumes of data from public and private sources.
Under the aegis of the EPA Office of Research and Development, Teknekron, Inc.,
^Speakers. c
-------
has compiled a relatively large aerometric data archive, on which this report
is based. The primary objective is to characterize the synoptic and aerometric
features of the interstate and regional occurrence of elevated ground-level
concentration of particulate sulfate (S(Q and ambient ozone (03). The re-
sults of the analysis should provide a basis for the development and evaluation
of regional air quality models and other predictive methods.
The fact that under certain meteorological conditions $04 and 03 are co-
occurring aerometric parameters in the industrial eastern United States and
Canada has been demonstrated by several investigators. Vukovich et al.1
showed the relationship between elevated 03 concentrations and the presence
and movement of anticyclones. They identified the back side of high-pressure
systems as a region most conducive to the occurrence of high 03 levels (i.e.,
contravention of the 0.08 ppm standard), particularly for those air masses
that have a relatively long residence time over the industrialized source
regions. These synoptic conditions are nearly identical to the intrusion of
warm and moist maritime tropical air from the Gulf of Mexico, conditions found
by Tong et al.2 to be most conducive for the occurrence of $04 levels ex-
ceeding, say, 20 ug/m3. Tong and Niemann,3 in their examination of regional
transport over the Ohio River Valley, have shown the close correspondence
between regional $04 and 03 episodes that were often related to an occluded
system moving from west to east over the Great Lakes. The presence of a more
or less stagnant west-east front located along the Canadian border area and
the westward intrusion of the Bermuda High is another configuration that gives
rise to regional excursions of SO^ levels in the Midwest and Northeast. Chung4
and Chung and Melo5 have largely confirmed these findings over Ontario, Canada.
In addition, Wolff et al.,6 Samson,7 Lioy and Samson,8 and others have demon-
strated, using the Moodus8 data, the eastward and north-eastward transport of
03 from the source or accumulation regions following the movement of a general-
ly eastward-moving anticyclone.
The analysis presented in this paper is directed at an examination of
data covering a period of 3 years—1975, 1976, and 1977—over the entire eastern
United States and the industrialized portions of eastern Canada. The approach
involves scanning both the S04 and 03 data by means of an episode retrieval
program and then selecting suitable (in terms of data coverage) cases for
96
-------
detailed aerometric analysis. Supplemental data such as the routine radiosonde
soundings and an enhanced satellite photograph were used whenever possible.
EPISODE TALLY AND CASE SELECTION
The episode retrieval program is designed to scan the air quality data
archive and tally the number of monitoring stations that reported a concentra-
tion greater than a designated threshold for each day of the month. The pro-
gram also gives the percentage of stations reporting on a given day to facili-
tate case selection. Tables I, II, and III present the 03 episode tallies
for each of the 3 years. In this analysis, the 1:00 p.m. EST 03 concentrations
were used. The entries in these tables show the number of stations that ex-
ceeded the indicated threshold values, and the percentage of stations report-
ing on that day. For example, on July 16, 1977, a total of 114 out of 238
stations reported 03 concentrations exceeding 0.08 ppm. A similar episode re-
trieval was performed for the $04 data from the National Aerometric Sampling
Network (NASN).
Examination of the 03 episode selection results for the 3 years shows
the well-known photochemical phenomenon of a summer 03 peak. During the
period from April through September of 1977 (Table III), area-wide contraven-
tion of the 0.08-ppm level was almost a daily occurrence. The July 16, 1977,
episode, for example, was quite extensive spatially with 30 stations exceeding
0.14 ppm. In fact, this episode lasted for 6 days from July 16 through July 21.
Regional sulfate data for case analysis are available for July 18, 1977. Un-
fortunately, the 6-day and 12-day $04 sampling frequency makes it impossible
to examine the regional distribution of S04 concentrations corresponding to
the peak 03 day (July 16).
The retrieval exercise for the $04 data used a threshold value of 25 M9/m3>
24-hr average. This screening resulted in the identification of nine regional
episodes that are candidates for case analysis. These cases are summarized in
Table IV with a brief description of the synoptic situation. An additional
03 case, which was analyzed by Wolff et al.6 and reanalyzed by Tong and Niemann,3
will also be examined to show the influence of the occluded system. Presented
below are three cases selected on the basis of the co-occurrence of S04/03
episodes: April 12-24, 1976; August 20-22, 1976; and July 16-21, 1977.
97
-------
TIME SERIES ANALYSIS
It is of interest to show the regional characteristics of 03 episodes and
their day-to-day development by means of a modified time-series plot. Figures
1, 2, and 3 are the time/regionality plots for April 1976, August 1976, and
July 1977, respectively. The number of stations exceeding the threshold value
has been normalized by the ratio of the percentage of the total reporting
stations to the average percentage of stations reporting during the period of
interest.
Examination of Figures 1, 2, and 3 and Tables I, II, and III shows that the
variation of regional 03 levels was more or less cyclic during the warm months
of the year. This cyclic characteristic is most likely a response to the
changes of synoptic systems as they migrate across the eastern United States
and Canada. This has been demonstrated to some extent by Lioy and Samson8
in their tracking of anticyclones in conjunction with the analysis of the
Moodus 03 data. Also given in Figures 1, 2, and 3 are the daily air mass
classifications for the Ohio Valley. These provide a rough indication of the
degree of moisture and heat intrusion. In addition, periods of stagnation
based on Korshover's classification scheme are also indicated in Figures 1, 2,
and 3 by horizontal arrows.
It is seen from the monthly plots in the figures that the synoptic meteor-
ological features exhibited certain identifiable macroscale characteristics.
First, days with regionally elevated 03 levels were, almost without exception,
associated with the back side of anticyclonic circulations over the study
region. These anticyclones originated in the interior of Canada and moved
eastward or southeastward into the eastern United States. An example of this
type of activity is the April 1976 case. During the summer months, the advance
and retreat of the semipermanent Bermuda High became the dominant factor. In
either case, the southerly and southwesterly flow on the western side of the
anticyclone was frequently the same in terms of the direction of pollutant or
precursor transport. In all cases, as shown in the figures, the peak 03 days
were largely associated with maritime tropical air masses originating over
tropical or subtropical waters.
Second, the correspondence of stagnation periods over the source region
(the Midwest) with the regional 03 peaks strongly supports the findings of
98
-------
Vukovich,1 Lioy and Samson,8 and others, that the onset of 03 episodes is
often preceded by a day or more of accumulation of 03 precursors. Presumably
the presence of the fair weather anticyclone would promote photochemical gener-
ation of 03 leading to peak 03 levels toward the end of the stagnation period.
The decline of the 03 episodes is often the result of frontal passage, which
is followed by the replacement of the polluted air mass by a fresh polar con-
tinental air mass.
CASE STUDIES
The analysis presented above leads naturally to a review of the regional
sulfate distributions analyzed by Hidy et al.10 and Tong and Batchelder11
during the past year. The cases selected are:
Case I—April 18, 1976, corresponding to the peak 03 day;
Case 2--August 22, 1976, one day after the peak 03 day; and
Case 3—July 18, 1977, two days after the peak 03 day.
What is of interest, in addition to the "indirect" analysis discussed above,
is to examine the spatial distribution of all relevant parameters in the tra-
ditional synoptic manner. This direct analysis for the actual episodes of
interest may be performed by mapping both the 03 and $04 data, along with other
relevant variables. The case analyses presented below provide this detailed
spatial examination.
Case 1—April 12-24. 1976
The 03 episode in Case 1 was originally presented by Wolff et al.12 who
showed the eastward movement of the 03 isopleth pattern. The tracking began
on April 12 when a small closed area of 03 >0.08 ppm was observed over Ohio.
This high 03 area expanded progressively as it moved eastward within the warm
sector of the occluded frontal system. Its size reached a peak on April 18,
covering a large region on the Atlantic seaboard. On that day (see Figure 4),
the area of 03 >0.08 ppm extended from southern New Hampshire and Vermont to
the southern Virginia border, then westward to cover West Virginia, Pennsylvania,
and part of New York State. Partial dissipation of this regional pattern
occurred when a cold front advanced southeastward, became occluded, and then
moved northward again. 03 levels in the warm sector remained high; e.g., 03 was
greater than 0.06 ppm in the warm sector, but was less than half that value
99
-------
north of the warm front. During the peak day on April 18, the backward tra-
jectories (see Wolff et al.12) showed directions generally from the west and
southwest, in accordance with the systematic macroscale translation of the
occluded system.
On April 18, the $04 pattern was almost identical to the 03 pattern in
that the area of maximum $64 coincided with that for 03 (see Figure 4). The
SO-ug/m3 $04 isopleth encompassed almost exactly the same region of 03 >0.08
ppm. The relative magnitudes of $64 concentration, with respect to the posi-
tion of the warm sector, showed the same variation as for 03, as shown in
Tong and Niemann3. Sulfate concentrations were available from five stations
around Boston on April 12, 18, and 24. The five-station averages for each of
the 3 days show that, before the onset and after the decline of the episode,
the $04 levels are generally around 10 ug/m3. During the peak day of this
episode, the $04 level exceeded 25 ug/m3 in Boston.
Case 2—August 18-23. 1976
The August 1976 case is an interesting episode in that it represents a
stagnation situation in which there is strong evidence of recirculation of
the pollutants within a relatively small region in eastern Ohio and western
Pennsylvania. In addition, it shows evidence of long-distance transport of
pollutants into western Tennessee and Arkansas from areas to the east. For
these reasons, this case is examined in considerable detail. The analysis
focuses on the day before, during, and after the date for which the $04 map
has been constructed on the basis of AQCR averages. Unfortunately, $04 data
for August 22 were available only for the midwestern States, southern Ontario
and some of the northeastern States.
The synoptic patterns during August 21, 22, and 23 are summarized in
Figure 5, showing positions of fronts and pressure centers, with the dates
indicated. During the several days preceding and during this episode, the
synoptic meteorology may be characterized by air stagnation with passage of
minor weak fronts. There was no significant movement of the high-pressure
center just south of Lake Erie during August 21 and 22. The entire eastern
United States was dominated by high pressure with a weak gradient. A weak
low-pressure center was present on the southern Atlantic Coast from August 19
through August 24.
100
-------
The 03 episode began on August 18 with 38 of the monitoring stations
exceeding the 0.08 ppm level. This number grew and reached a peak of 117 on
August 21 and then declined to 99 on the following day. The region (see
Figure 6) encompassing the western half of Lake Erie, northern Ohio, the
northeastern corner of Indiana, southeastern Michigan, and southwestern
Ontario reported 03 levels consistently greater than 0.14 ppm on August 21,
22, and 23. Most of the northeastern quadrant of the United States and
southern Ontario was within the region of contravention (>0.08 ppm) during
these 3 days. The decline of the regional peak came on August 23 and 24 with
the passage of a weak cold front that advanced from the northwest. The region
of high 03 began to diminish in size as the stagnation dissipated with the
frontal passage. The regional pattern did not disappear entirely because on
August 25 the eastern United States was again dominated by an anticyclone
centered over eastern Pennsylvania. Light but persistent moisture and heat
advection from the south and southwest continued through August 28.
As shown in Figure 6, the S04 distribution for August 22 is a typical
prefrental pattern with the cold front situated in the northeast-southwest
direction over the Great Lakes. Highest $04 was observed over the Detroit-
Windsor-Cleveland-Toledo region coinciding almost exactly with the persistent
regional 03 peak mentioned above but not shown in Figure 6. Again the area
within the 20-ug/m3 $64 isopleth is also coincident with the 03 pattern as
shown by the large shaded area in Figure 6. This area extended roughly from
the eastern half of Indiana to the New England coast.
The persistence of the 03 pattern and the stagnation of the high-pressure
center during this episode led to an examination of visibility patterns and
air parcel trajectories. These patterns are shown in Figure 7 for August 22.
Two visual range isopleths are plotted, one enclosing airport visibility values
between 3 and 4 miles; and another for the interval of 4 to 6 miles. The
general correspondence between the latter isopleth and the S04/03 pattern de-
scribed above is quite evident. Of interest is the small closed area of visi-
bility less than 4 miles over the regional S04/03 peak (Detroit-Windsor-
Cleveland-Toledo) on this day. Furthermore, an enhanced satellite photograph
was made available by Lyons13 who was able to delineate the boundary of the
haze layer adjacent to the ground surface. The location of this boundary is
roughly shown in Figure 5, showing the possible subcontinental extent of the
101
-------
polluted air mass on this day. In view of the persistence of the 03 pattern,
it may be speculated that this subcontinental haze layer may also have per-
sisted for at least 3 days.
In an attempt to infer the source-receptor relationships for this day,
selected air parcel trajectories were constructed and shown in Figure 7. Two
interesting results should be noted. First, around the anticyclone that per-
sisted for 2 days just south of Lake Erie, there was a recirculation of the
polluted air for 3 days beginning on August 20. Pollutants emitted in the
vicinity of the Ohio-Pennsylvania border moved slowly westward into Ohio,
curved northward into southeastern Michigan, and then returned to Pennsylvania
via southwestern parts of Ontario. This case is somewhat extraordinary due
to the long-distance circular transport that resulted in equally persistent
high $04/03 levels in the western Lake Erie multi-city complex.
The second point of interest is the trajectories originating in south-
central Tennessee. In their previous analysis, the authors had observed oc-
casional $04 "hot spots" over western Tennessee and Arkansas. Perhaps because
that region is only in the periphery of major source regions, namely the Ohio
Valley and the Tennessee Valley, no detailed analysis was made of these excur-
sions. On the basis of Figure 7, however, it appears that the westerly trans-
port of pollutants emitted in Tennessee and northern Alabama contributed to
the high 03 observed along the Tennessee-Arkansas border (see Figure 6).
Reisinger and Crawford,14 in their detailed analysis of three August 1976
episodes, reported a peak S04 concentration of 32 |jg/m3 at Blytherville,
Arkansas, confirming the consistency of these observations and the correspon-
dence between regionally elevated S04 and 03 concentrations.
Case 3-July 12-21. 1977
Case 3 is a regional 03 episode that was first analyzed by Wolff and
Lioy.12 This episode was associated with a prolonged stagnant air mass
covering the entire eastern half of the United States, with persistent hot
and humid conditions lasting for more than 9 days. This period was termed
by Wolff and Lioy12 a heat wave; and the transport was said to take place in
an "ozone river" that ran from the Texas Coast to the Washington-Boston corri-
dor. While the existence of this idealized "river" remains to be verified,
the transport of pollutant toward the northeast from the mid-South region
102
-------
appears to be a macroscale feature of the general flow. It is of interest at
this point to focus on the aerometric conditions during the 3-day period,
July 16, 17, and 18.
The synoptic conditions during this period, aside from the presence of
the generally stagnant air mass, are a broad pressure gradient throughout the
region, and the presence of a stationary front over the midwestern States as
shown in Figure 8. This stationary front delineated the two contrasting air
masses: maritime tropical to the south and polar air to the north. There was
a fairly broad transitional zone along the front that fluctuated somewhat in
the north-south direction. These conditions were not broken until 4 days
later on July 22.
The peak 03 day was July 16, on which a broad band extended from eastern
Arkansas eastward to the Carolinas, thence to southern Maine as shown in
Figure 9. Areas in which 03 exceeded 0.14 ppm were situated over the junction
of Tennessee, Alabama, and Georgia; over the northern Delmarva Pennisula; and
over southern New England. The fact that this broad band was deflected south-
ward was due primarily to the position of the stationary front. On the next
day, July 17, the pattern had diminished in size and 03 concentrations were
generally lower. The area of 03 greater than 0.14 ppm remained in only
southern Connecticut.
On July 18 (see Figure 10), a day when regional 504 data were available,
the stationary front became somewhat less defined. The 03 pattern moved
northward into the Midwest, with two areas exceeding 0.10 ppm, one located in
Indiana and Ohio, and another over central Tennessee and Kentucky. The $04
pattern over the Midwest is nearly identical to the 03 pattern. Sulfate con-
centrations of 40 and 50 ug/m3 coincided with the 03 peak in Indiana and Ohio.
Southern Ontario showed relatively low levels of $04 (and probably 03, also)
during these days due primarily to the frontal barrier along the border.
DISCUSSION
The compilation and analysis of S04/03 episodes are continuing to further
characterize the corresponding regional aerometric features. Examination of
these case studies and others not yet published (see Table IV) leads to the
following conclusions:
103
-------
1. The co-occurrence of S04/03 episodes is primarily a warm
and fair weather phenomenon, in which photochemical activ-
ities are thought to be much enhanced.
2. Pollutants such as $04 and 03 and/or their precursors are
transported by the eastward translation of an occluded sys-
tem, circulations around the backside of an anticyclone,
and circular advection around small and persistent high-
pressure centers.
3. Episodes in western Tennessee and Arkansas may occasionally
be due to advection of pollutants and/or precursors from
sources in the Tennessee Valley.
4. Frontal positions are good macroscale markers delineating
the boundaries of $04 and 03 episodes.
5. Since regional S0l/03 episodes are primarily a fair-weather
occurrence, the haze layer formed by these pollutants can
be detected by satellite imagery.
6. The spatial extent of $04/03 episodes may on occasion be
subcontinental in scale, which would require interstate and
international action for control.
ACKNOWLEDGMENT
The authors acknowledge with thanks the data processing and technical
assistance provided by Mr. David Hergert, Dr. Alan Hirata, and Ms. Anne
Mazzarelli of Teknekron. This work was sponsored by the U.S. Environmental
Protection Agency under Contract Number 68-01-1921.
REFERENCES
1. Vukovich, F. M., et al. On the Relationship Between High Ozone in the
Rural Surface Layer and High Pressure Systems. Atmos. Environ., 11:967,
1977.
2. Tong, E. Y., et al. Regional and Local Aspects of Atmospheric Sulfates
in the Northeast Quadrant of the United States. Proceedings, Third
Symposium on Atmospheric Turbulence, Diffusion and Air Quality, American
Meteorological Society, October 1976.
3. Tong, E. Y., and B. L. Niemann. Meteorological Similarities Between the
Occurrence of Ambient Ozone and Particulate Sulfate Episodes. Teknekron
Staff Research Memorandum, unpublished, April 1978.
104
-------
4. Chung, Y. S. Ground-Level Ozone, Sulphates and Total Suspended Partic-
ulates in Canada. Proceedings, International Clean Air Conference,
Brisbane, Australia, May 15-19, 1978.
5. Melo, 0. T., and Y. S. Chung. Regional Air Pollution-Sulphates.
Meeting of the Chemical Institute of Canada, Winnipeg, Canada, June 4-7,
1978.
6. Wolff, G. T., et al. Transport of Ozone Associated With an Air Mass.
Paper presented at 70th Annual Meeting, APCA, Toronto, Canada, June 20-24,
1977.
7. Samson, P. J., and K. W. Ragland. Ozone and Visibility Reduction in the
Midwest: Evidence for Large-Scale Transport. Jour. Appl. Met.,
1101-1106, 1977.
8. Lioy, P. J., and P. J. Samson. Ozone Concentration Patterns Observed
During the 1976-1977 Long-Range Transport Study. Paper submitted to
Environment International, 1979.
9. Korshpver, J. Climatology of Stagnating Anticyclones East of the Rocky
Mountains, 1936-1975. NOAA Tech. Memo. ERL-ARL-55, Air Resources
Laboratory, Silver Spring, Maryland, March 1976.
10. Hidy, G. M., P. K. Mueller, and E. Y. Tong. Spatial and Temporal Dis-
tributions of Airborne Sulfate in Parts of the United States. Atmos.
Environment, 12:735-752, 1978.
11. Tong, E. Y., and R. B. Batchelder. Aerometric Data Compilation and
Analysis for Regional Sulfate Modeling. Technical Report, Teknekron,
Inc., under EPA Contract Number 68-01-1921, Berkeley, California, 1978.
12. Wolff, G. T., and P. J. Lioy. Ozone Concentration Patterns Associated
With the July 1977 Eastern U.S. Heat Wave. Paper presented at 71st
Meeting of APCA, Houston, Texas, June 25-30, 1978.
13. W. Lyons, Mesomet, Inc., Chicago, Illinois, private communication, 1979.
14. Reisinger, L. M., and T. L. Crawford. August 1976 Sulfate Episodes in
the Tennessee Valley Region. TVA Technical Report, TVA/EP-79/04, 1979.
105
-------
0>
's
ll <°
(A "1
cO
l|
f! 00
oo 20
w °
April, 1976
— l-rf
M
'
TCT CC C JY
vi§ IfCIIC'lll
• lllllll tllll
(Part A)
M
M
M
15 10 15
.K
H
c
T
M C i
• Til T T T ^
I ' II^^T*' •
1 i 1 1 r u . i i I 1
20 25 30
Day of Month
NOTE: Letters indicate the predominant air mass over the Midwest
(M = maritime tropical; C = polar continental; T = transitional).
Arrows indicate the periods of air stagnation based on Korshover's
criteria.
Figure 1. Time-series regionality plots of ozone showing the normalized
number of stations that exceeded 0.08 ppm for April 1976.
106
-------
i*iu
0)
N 100
O
E
c.
a
00
9 80
o
a
c
'•5
OJ
Q>
x 60
UJ
M
c
JO
2 40
w
0
O
J3
| 20
Z
_
"
••
-
August, 1976
^ w
M
M
T
1
J
T
4 fe
M
M
M
I I
1 1
T
M
C
c
c
1
5 10 15
M (Part B)
-^
M
T
T
s c
|
w
M
M
C
' *
M
/I
M
M
T
C
_LL
20 25 30
Day of Month
NOTE: Letters indicate the predominant air mass over the Midwest
(M = maritime tropical; C = polar continental; T = transitional).
Arrows indicate the periods of air stagnation based on Korshover's
criteria.
Figure 2. Time-series regionality plots of ozone showing the normalized
number of stations that exceeded 0.08 ppm for August 1976.
107
-------
I^U
0
c
2 100
0
1
a
03
°. 80
O
ra
.£
0
S)
o 60
X
u
(A
C
_0
^
2 40
W
^
o
b
0)
A
§ 20
2
-
_
_
-
r
V
c
July, 1977
- .
M M
T
*
^
M
M
M
T
M
M
T
C
c
c
T
(Part C)
M
T
T
M
M
M
T
T
C
C
T
T
T
10 15 20
Day of Month
25
30
NOTE: Letters indicate the predominant air mass over the Midwest
(M = maritime tropical; C = polar continental; T = transitional).
Arrows indicate the periods of air stagnation based on Korshover's
criteria.
Figure 3. Time-series regionality plots of ozone showing the normalized
number of stations that exceeded 0.08 ppm for July 1977.
108
-------
Figure 4. Aerometric summary of Case 1, April 18, 1976, showing the
relationship of ozone and sulfate and pollutant confinement in an occlusion.
109
-------
8/23
8/21
Note the persistent anticyclone just south of Lake Erie, the presence and
advancement of a weak cold front, and the presence of a low-pressure
center on the southern Atlantic Coast.
Figure 5. Summary of synoptic events for Case 2,
August 21, 22, and 23, 1976.
110
-------
Note the correspondence of the two pollutant patterns and the area enclosed
by the 40-Mg/nrr SO^ isopleth in which O3 concentrations exceeded
0.14 ppm.
Figure 6. Spatial distribution of sulfate and ozone for August 22, 1976.
Ill
-------
August 22, 1976
—— Visibility 3-4 mi
_ _ Visibility 4-6 mi
Boundary of Haze Layer
Trajectory Starts at OOZ
Trajectory Starts at 12Z
ir Origin of Trajectory
Note the recirculation around the northeast comer of Ohio and the likely
regional transport of pollutants into western Tennessee and Arkansas.
Figure 7. Spatial distribution of visual range, the satellite-observed haze
layer, and the forward trajectories (beginning on August 20) of air parcels.
112
-------
Note the washing out and stationary nature of the frontal zone.
Rgure 8. Summary of synoptic events for Case 3,
July 16, 17 and 18, 1977.
113
-------
Note the reduction in size of the pattern and the persistent high O3 area in
Connecticut.
Figure 9. Spatial distribution of ozone for July 16 and 17, 1977, during the
peak of the ozone episode.
114
-------
Note the correspondence of O3 and SO^ over the region, particularly over
Indiana and Ohio.
Figure 10. Spatial distribution of sulfate and ozone concentrations
for July 18, 1977.
115
-------
TABLE I. RESULTS OF EPISODE RETRIEVAL FOR 1975*
Threshold Value 0.08 ppm
Day January February
1
2 1-31
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22 1-32
23
24
25
26
27
28
29
30
31
March
1-32
1-31
1-31
1-31
4-32
1-30
1-33
1-27
1-30
April
1-32
1-31
1-30
1-30
1-31
2-31
3-30
2-31
2-32
3-32
1-30
2-30
1-32
5-33
4-32
4-31
1-26
1-32
2-32
May
9-32
1-31
3-33
3-33
4-32
6-32
11-32
14-32
19-32
7-32
4-32
6-32
29-33
13-31
20-31
22-32
44-33
33-33
25-34
45-32
29-32
5-30
3-30
5-31
9-30
18-33
14-33
1-34
June
1-33
4-36
10-36
11-35
11-36
1-35
1-36
8-36
9-36
2-36
11-36
26-35
13-34
1-35
13-36
13-36
19-31
13-34
23-33
46-32
34-36
41-36
33-38
20-38
28-38
30-37
31-35
25-37
July
37-40
66-40
67-40
27-39
32-39
36-38
16-40
40-39
33-40
41-41
13-42
9-40
8-39
1-42
11-42
28-43
26-43
41-43
22-42
14-43
26-44
26-44
75-43
43-44
2-44
11-43
46-41
56-44
29-44
74-45
97-44
August
96-44
72-42
48-41
29-44
20-44
1-44
3-44
10-44
32-44
62-44
33-44
41-45
39-46
14-46
9-45
14-44
5-44
34-45
9-46
17-45
24-45
14-46
10-46
6-44
16-45
26-45
7-45
27-46
47-46
16-45
4-45
September
6-44
9-46
16-45
12-45
9-46
1-46
4-45
7-46
4-45
4-45
4-45
1-45
1-41
3-44
1-44
1-45
2-44
4-41
4-44
15-46
October November December
3-41
5-37
3-39
1-39
2-38
3-39 1-29
6-37
6-38
5-38
4-39
4-39
3-39
1-30
1-36
2-38 1-35
5-38
10-38
11-39
2-38
1-35
*See footnote at the end of table.
(continued)
-------
TABLE I (continued)
Threshold Value 0.14 ppm
Day January February
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
March Apri 1 May
1-32
1-32
1-32
1-32
1-33
1-32
3-33
5-34
6-32
5-32
1-33
June
3-33
1-32
3-36
7-35
1-38
1-38
2-38
3-37
1-35
July
2-40
9-40
10-40
3-38
1-39
1-40
2-43
3-43
3-43
2-42
2-44
11-43
1-44
2-43
3-44
9-45
27-44
August September October November December
25-44
16-42
5-41 2-45
10-44 1-45
3-44
1-44 1-37
1-44 1-38
1-45
3-46
1-46
1-44
2-45
3-46
1-46
1-45
*First number of each entry is the total number of stations exceeding the threshold value, and the second number is
the percentage of stations reporting for that day. The total number of reporting stations is 466.
-------
TABLE II. RESULTS OF EPISODE RETRIEVAL FOR 1976*
00
Threshold Value 0.08 ppm
Day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
January February March
2-39
1-41
1-40
1-41
1-36
1-40
1-39
1-41
1-42
1-42
7-41
2-41
2-39 1-41
2-41
2-40 3-41
3-39 1-40
6-32
April
1-42
2-41
7-40
7-40
4-41
3-41
6-41
6-42
3-42
7-42
3-43
9-43
9-43
20-43
18-42
48-41
55-40
47-41
23-42
9-44
2-44
9-43
12-43
2-44
3-45
3-45
5-45
May
10-44
6-45
2-44
8-42
19-42
1-42
1-42
7-42
24-42
12-43
9-42
7-43
3-43
2-44
1-44
1-45
1-44
20-44
30-44
14-44
5-44
3-45
6-44
5-44
10-44
13-44
11-43
3-42
7-41
June
1-45
7-46
20-45
30-45
52-44
68-45
116-45
100-45
107-46
108-46
71-46
40-45
22-45
49-45
20-45
12-46
39-46
1-44
4-44
6-44
10-45
20-45
17-46
9-47
27-45
57-43
48-46
29-46
5-45
July
19-44
25-45
35-45
21-45
29-44
59-45
20-47
55-47
41-47
41-47
41-44
3-47
18-45
28-47
33-47
27-45
12-45
30-43
59-47
67-47
27-47
11-47
27-48
38-47
18-47
26-47
59-48
39-46
15-46
26-46
26-45
August
3-43
14-45
50-46
71-46
35-45
11-45
13-45
29-45
37-46
70-47
58-48
44-46
11-45
6-47
9-47
38-47
45-48
66-47
117-47
99-46
74-47
60-48
56-49
61-48
31-48
25-47
6-47
5-48
18-47
September
41-49
2-50
9-51
6-51
4-51
5-50
11-51
52-52
47-51
4-51
7-51
27-51
46-50
61-49
27-50
5-50
16-51
29-50
34-48
18-50
1-50
9-51
8-52
7-52
4-51
1-50
2-50
1-50
October November
6-50
34-48
41-48
21-50
1-44
4-42
1-42
6-45
14-48
13-48
12-49
11-50
17-50
1-43
1-41
1-46
1-45
1-41
1-45 1-41
1-45
1-42
December
1-41
1-40
2-39
1-39
1-39
1-39
1-39
1-40
1-39
1-38
*See footnote at the end of table.
(continued)
-------
TABLE II (continued)
Threshold Value 0.14 ppm
Day January February March April
1
2
3
4
5 1-41
6
7
8
9
10
11
12
13
14
15
16
17 . 1-41
18 6-40
19 5-41
20 1-42
21
22
23
24
25
26
27
28
29
30
31
May June
3-44
1-45
1-45
2-45
8-45
13-46
7-46
9-45
2-45
3-45
4-46
1-44
3-46
1-45
3-43
7-46
3-46
July
1-45
3-45
2-44
9-45
2-47
1-47
1-45
1-47
5-44
1-45
1-47
1-47
2-45
7-47
3-47
3-47
3-48
8-46
1-46
August September October November December
2-50
3-48
1-48
6-46
1-45
2-52
4-51
2-46
7-47
12-48 1-50 1-49
6-46 7-49
1-47
5-47 1-50
3-45
3-47
10-47
2-46
2-47
1-48
2-49
7-48 1-45 1-41
*First number of each entry is the total number of stations exceeding the threshold value, and the second number is
the percentage of stations reporting for that day. The total number of reporting stations is 466.
-------
TABLE III. RESULTS OF EPISODE FOR 1977*
Threshold Value 0.08 ppm
Day
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
January February March
4-42
1-39 1-41
1-41
1-42
1-41
1-42
1-40 4-42
4-42
3-42
1-43
1-41 1-42 1-43
1-41
1-42
1-41 1-42
2-42
1-41
1-42 2-42
2-43
1-41 5-42
1-42
April
3-46
1-45
1-47
1-47
2-48
1-47
8-47
8-46
20-48
60-48
39-48
11-50
22-50
28-48
31-47
27-48
17-47
17-48
20-49
22-49
1-48
3-49
8-49
4-50
6-49
May
19-45
1-49
1-49
3-49
21-48
1-47
4-47
2-49
14-51
39-51
63-53
62-50
50-50
57-52
109-52
72-52
57-53
74-53
61-51
37-50
30-51
67-51
64-51
38-50
60-52
117-51
52-50
41-46
38-51
June
9-50
25-50
14-49
71-50
73-49
14-48
1-51
10-52
9-52
22-54
41-52
48-51
33-53
44-51
66-53
56-54
50-53
34-51
6-50
22-51
4-52
5-52
14-52
29-53
9-51
19-50
51-51
40-51
15-51
25-52
July
10-50
14-48
38-47
59-47
53-49
55-50
44-51
34-51
31-50
30-50
9-51
15-50
50-50
49-49
94-52
114-51
78-50
51-49
78-51
70-49
61-51
2-51
36-49
51-48
1-51
4-50
10-51
55-51
21-52
41-51
30-50
August
19-50
18-51
37-51
56-51
41-51
22-50
10-48
9-50
8-51
14-51
20-51
11-50
1-49
12-48
8-50
7-50
3-51
18-52
17-50
18-50
9-51
10-50
2-50
3-50
11-50
23-50
26-48
28-49
15-47
21-48
September
18-50
40-50
17-48
23-48
10-47
13-49
10-49
21-50
10-50
3-49
1-48
3-49
6-51
1-49
1-48
10-47
6-48
1-49
2-49
7-49
1-49
3-48
1-48
3-49
October November December
3-45
1-48
1-42
1-49
5-48
1-47 1-40
1-48
1-48
1-47
11-47
3-46
2-45
1-46
3-46
1-45
*See footnote at the end of table.
(continued)
-------
TABLE III (continued)
Threshold Value 0.14 ppm
Day January February March April
1
2
3
4
5
6
7
8
9
10
11
12
13 1-48
14
15
16
17
18
19
20
21 1-49
22
23
24
25
26
27
28
29
30
31
May
1-49
1-53
1-50
2-50
3-52
2-52
5-52
3-63
1-51
3-51
5-51
2-50
2-52
9-51
3-50
June
1-50
1-50
1-49
1-54
1-51
5-53
3-54
3-53
5-51
1-50
3-51
2-51
July
3-47
2-49
3-50
2-51
3-51
3-50
1-50
2-50
3-49
4-52
30-51
4-50
2-49
11-51
8-49
6-51
1-49
1-48
2-51
1-51
1-50
August September October November December
2-50
5-50
1-51
1-51
8-51
4-50
1-51
1-48
1-52
6-48
4-49
2-47
"First number of each entry is the total number of stations exceeding the threshold value, and the second number is
the percentage of stations reporting for that day. The total number of reporting stations is 466.
-------
TABLE IV. SUMMARY OF SULFATE EPISODES DURING 1975, 1976, and 1977
DATE
DESCRIPTION
1975
June 23 High-pressure center on Atlantic Coast dominating eastern half of
United States for several days. Sulfate advection from midwest and
southcentral States to as far as northern New York and New England.
July 17 Identical synoptic situation as June 23, except Bermuda High is far-
ther inland on the Atlantic Coast. This situation resulted in high
sulfate concentrations in Wisconsin, Illinois, Indiana, and Michigan,
as well as pollutant advection into Ontario, New York State, and
New England.
August 10 Bermuda High dominating eastern United States. High-pressure center
persisted for several days on the southern Atlantic Coast. Pollut-
ant transport is similar to that on July 17, with high sulfate
levels also being advected into the mid-Atlantic and northern
Atlantic States.
1976
June 11 Data available for only a few of the States in the northeast quadrant
of the United States. Sulfate levels exceeding 40 ug/m3 in Ohio
and West Virginia. Stationary front just north of Great Lakes.
Persistent high-pressure center situated in northern Georgia-western
Carolinas for several days.
August 22 A persistently stagnant situation with high-pressure center just
south of Lake Erie. Sulfate levels exceeded 50 ug/m3 in the northern
Ohio-southern Michigan region. Recirculation is evident over several
States. Westward advection from western Pennsylvania, northerly
and northeasterly transport into Michigan, Ontario, and recurvature
southward to New York State and Pennsylvania. Westerly transport is
also evident in the TVA region, resulting in episodic concentrations
in western Tennessee and Arkansas.
September 9 Typical case of pollutant confinement in the warm sector of an oc-
cluded system. High sulfate concentration in Ohio, Pennsylvania,
New York, and States abutting them to the south.
1977
May 31 Another example of the occluded confinement. The warm sector is
farther south than the previous case.
June 24 Nearly identical to May 31 case.
July 18 A persistent Bermuda High situation with a west-east stationary
front along the Canadian border. Eastern half of United States
experienced hot and humid conditions for many days. High sulfate
levels in a tier of States from Missouri to New Jersey, excluding
northern parts of New York and New England.
August 5 Poor data coverage but high sulfate levels in Pennsylvania and New
York. A typical Bermuda High situation.
August 11 Moderate sulfate levels in Ohio, West Virginia, and western
Pennsylvania. A peripheral occlusion situation.
August 23 Similar to August 11 case.
October 22 Similar to August 11 and 23 cases.
122
-------
STATISTICS OF ELEVATED POLLUTION EPISODES
Gerald F. Watson* and Walter J. Saucier*
North Carolina State University, Raleigh, North Carolina
INTRODUCTION
Climatologies! studies of pollution episodes are few and have mainly been
indirect in that the meteorological conditions often associated with pollution
are studied rather than the pollutants themselves. Thus, Holzworth (1972,
1974) compiled statistics on air pollution potential by means of a mixing
depth/boundary-layer mean-wind index. Korshover's (1975) 40-year climatology
of stagnating anticyclones is well known. One of the few studies directly
concerned with pollutants is Altshuller's (1978) 10-year climatology of high
oxidant levels at four observation sites in the eastern United States.
The present study will report on some of the statistical characteristics
of elevated pollution episodes (EPE's) based on the evolution of haze and/or
smoke (HK) areas east of the Rocky Mountains. This study, too, is an indirect
approach to the climatology of pollutants. However, many investigators (e.g.,
Altshuller, 1978; Wolff et al., 1977; Hall et al., 1973) have established
that persistent HK areas (blobs) are often, if not always, associated with in-
creased oxidant concentrations. Related light-scattering aerosols produce
the low visibilities accompanying the hazy conditions. In so far as this
relationship holds, the following results based on haze may be interpreted
as indicative of the statistical behavior of atmospheric pollutants—natural
and manmade.
PROJECT DESCRIPTION AND GOALS
A recent 5-year period, June 1973-May 1978, was selected, and daily NMC
Weather Depiction Charts were obtained from the National Climatic Center.
The primary chart time was 1900 GMT (2 p.m. EST), chosen to minimize the
interference of morning radiation fogs. Also, sufficient convective mixing
*Speakers.
123
-------
should have occurred by this time so that transient high concentrations of
early morning HK would not be mistaken for day-long persistence of interest
here.
The major goals of the project were to:
1. Identify and track prominent HK blobs.
2. Determine the geographical distribution of monthly HK occurrence
and persistence, and to relate these characteristics to the
synoptic-scale weather and flow patterns.
3. Compile statistics on the dimensions, trajectories, visibilities
and other associated weather parameters (e.g., humidity, cloud
cover, etc.).
4. Relate HK episodes to actual pollutant concentrations via EPA
SAROADS data.
In the 3 months since formal work on the project began, some of these goals
have been attained and other at least explored. The progress thus far
achieved is in large measure due to the enthusiastic help of seven of our
meteorology students--three graduate students and four seniors. I would like
to acknowledge their help at this time.
IDENTIFICATION OF HK EPISODES
The identification of HK episodes involved a day-by-day scan of the
Weather Depiction Charts looking for hazy areas that exhibited continuity over
a few days. Obvious mesoscale and/or transient situations were eliminated;
more substantial cases were outlined on tracing paper for their duration. These
daily isochrones of blob outline would include limited areas of fog and precip-
itation if associations were clear. Otherwise, only stations reporting haze
or smoke were included.
Incidentally, it is somewhat regrettable--from the point of view of this
project—that NMC changed the Weather Depiction Chart symbology for weather
and obscurations in about the middle of the 5-year period (July 1976). In the
earlier part of the period, aviation code symbology was used; synoptic code
symbology was adopted in the latter half. The aviation code permitted the
observer to report, for example, the occurrence of haze with other kinds of
weather as fog or rain. The computer-printed synoptic symbology displays only
one type, the dominant weather, thus making decisions about the incorporation
of weather areas in close proximity to a hazy area somewhat more subjective.
124
-------
In any case, the daily HK isochrones for each of the episodes were to
serve many purposes in this project, and much effort went into their construc-
tion. One of the uses of the HK isochrones was to determine the geographical
distribution of HK persistence. Persistence is certainly of importance in
the determination of the possible adverse health effects. Also, these charts
helped to formalize the definition of an HK episode for purposes of this
project.
EXAMPLES OF HK EPISODES
Figure 1 shows isopleths of persistence (i.e., the number of consecutive
days with haze or smoke) for two episodes. Persistence values were determined
by superimposing the daily isochrones on a 1° latitude/10 longitude grid and
marking all grid squares with at least 50 percent of the area included within
the isochrone.
The first example in Figure 1 is a mini-episode, June 22-25, 1976. The
dashed line is the episode "envelope," or the area within which haze/smoke was
reported on at least one day. During its brief 4-day lifetime the blob affected
some 350,000 km2 over the eastern Great Lakes.
The isopleths of 2, 3, and 4 consecutive days of haze/smoke are drawn.
Though the episode was of short duration, it was sufficiently stationary so
that an area of more than 50,000 km2 experienced the maximum HK persistence
of 4 days.
The other HK episode shown in Figure 1 lasted for 21 days (July 4-24, 1977),
and during this time affected an area of at least 2.8 million km2. How far the
HK area may extend into the Atlantic is unknown. Maximum persistence of 9
days occurred along the Ohio Valley and coastal regions of Virginia and North
o
Carolina.
Examination of many such cases from the 5 years of data led to what is
believed to be a reasonable working definition of a "haze/smoke episode." An
episode is one in which the region of 3 or more days' persistence covers an
area of at least 10s km2. Such an area is a box about 3° latitude square, or
about the area of North Carolina.
Application of HK episode criterion to those cases originally obtained
resulted in the neglect of some and in the addition of a few. The brief and
localized June 1976 case shown in Figure 1 would be readily accepted as an
episode under the criterion.
125
-------
TABULATION OF HK EPISODES
Table 1 lists dates of the 42 episodes identified during the 5-year period
June 1973-May 1978. Episodes that only marginally satisfied the above criterion
are indicated by asterisks. The arrows indicate episodes that are continuous
across months.
The remarkable 3-month eposide extending from July 1 to October 4, 1973,
is to be noted. This is one episode because a 3-month continuity of HK areas
can be traced, although the areal extent varied from occupying the eastern
one-third of the United States to near disappearance for a day or so. As a
matter of fact, except for what appear to be clean breaks on June 30 and
October 5, the episode may have been 4 months long—June 9 to October 13!
Nothing quite like this occurred in the other years, although there were a
few episodes approaching a month in length.
For purposes of counting episodes, July, August, and September (including
the first 4 days of October) 1973 are each assigned one episode of a full month's
duration. The number of episodes in each month, and their minimum and maximum
duration, are given at the bottom of the table. The total number of monthly
episodes (46) differs from that of the actual total (42) because episodes that
spanned 2 months were credited to both if at least one-third of the episode
duration occurred in one or the other month. Table 1 shows a broad maximum
in episode number and duration centered on July.
The table also gives (below the dates) the general geographical region(s)
of persistence >3 days. None of the episodes extended west of 100° longitude;
the six geographical regions in the eastern United States are defined in
Figure 2. The abbreviations are as follows: central region (C), north
central (NC), northeast (NE), east central (EC), southeast (SE), and south
central (SC). Episodes concentrated primarily along the boundaries between
regions are indicated by underscoring.
The percentages shown in Figure 2 represent the number of times each geo-
graphical region appears in Table 1 divided by the total number of episodes (42).
Thus, for example, >3 days' persistence occurred in the northeast region in 74
percent of the 42 episodes. This region was most frequently visited by persis-
tent HK episodes, the south central region least frequently.
126
-------
SPATIAL AND TEMPORAL VARIABILITY OF HK EPISODES
Tabulations of frequency of occurrence of HK in 1° latitude/10 longitude
grid squares for each episode were partitioned into the respective months in-
volved. Monthly statistics over the 5-year period of HK spatial and temporal
variability represent the major completed work of the project. Some examples
will now be presented.
The distribution of the frequency of occurrence of haze/smoke during
July (1973-1977) is shown in Figure 3. The outer long-dashed curve represents
the extremity of haziness composited over the 5 years; the western boundary
along the 100° meridian is to be noted.
The isopleths are lines of constant probability (percent) and were ob-
tained by dividing the number of days of HK by the total number of July days
in 5 years. For practical purposes the total days is taken to be 150 for all
months; then 3 days of HK out of 150 days is 2 percent. In general, the
actual number of days with HK can be obtained by multiplying the isopleth
values by 1.5.
The proper interpretation of Figure 3 is that it shows the probability
of the occurrence of HK on any individual day in July. This is HK associated
with episodes and is not necessarily the same as that obtained if all HK
events were included. One notes in the figure that HK is most likely along
the Virginia-North Carolina border where the probability exceeds 40 percent
(or more than 60 days of HK during July over the 5 years). Norfolk, VA, is
located in the region of maximum probability.
The chance of one day of haze in every three (32 percent) encompasses
most of North Carolina, the Great Smoky Mountains, and a secondary maximum
(36 percent) along the Ohio River in southern Ohio. This pattern, including
the prominent minimum (<24 percent) on the Virginia-West Virginia boundary,
is recognizable in other months, June-September.
Of greater concern from the point of view of the impact of episodes on
human health is the persistence of polluted conditions over a given locale.
Monthly statistics on the likelihood of >2, >3, >4, and >5 consecutive days
of HK are provided in the final report. An example is shown in Figure 4 in
which the isopleths represent the probability (percent) of >4 consecutive
days of HK occurring at least once during July. The 20 percent value corre-
127
-------
spends to such an occurrence observed once in the 5-year period, 40 percent
to twice in 5 years, and 100 percent to an occurrence in each of the 5 years.
Figure 4 suggests that EPA could maximize its chances (>80 percent) of
encountering an extended pollution episode by situating the STATE field
experiments in the upper Ohio River Valley; the Washington, DC-Baltimore, MD,
area; and/or along the Virginia-North Carolina border.
The charts provided in the final report can be combined in a variety of
ways to show other aspects of HK variability. For example, Figure 5 depicts
regions in which the probability is 60 percent that 2, 3, 4, 5, and 6 consecu-
tive days of HK will occur. The hatched areas indicate that persistence is
of shorter duration than the isopleth value. The figure shows a better than
even chance of >2 successive days of HK during July over the eastern one-
third of the country. Such a probability for at least 4 days in a row is
restricted to the mid-Atlantic States and the Ohio River Valley.
Figure 6 is similar to the previous figure except that the probability
is 100 percent. A nearly certain chance of 3 successive days of HK sometime
during July is limited to North Carolina and a few other locales. Regions in
which stagnant conditions lasted a week or more, and which occurred at least
once in the 5-year period, are shown in Figure 7.
Some idea of interannual variability is illustrated in Figure 8. Iso-
chrones of 4 or more consecutive days of haze in July 1973, 1975, and 1977
are shown, and differences are evident. The southern extent of the haze
areas along the east coast is found farther northward as time progresses, and
the east coast areal coverage in 1977 is much smaller than in the other 2
years. Another 1977 area is centered over the Ohio River; this region was
free of persistent haze in 1973 and 1975.
Monthly variability is demonstrated by Figure 9. The isopleths show re-
gions in which there is a 60 percent probability of >4 days of haze for all
months between May and October. However, only 3 of the months--June (JN),
July (JL), and August (AU)—exhibited such areas. Along the east coast the
areal coverage in July is much larger than 1 month earlier or later. August
appears to have the larger and better organized region over the Ohio Valley.
HK EPISODE DIMENSIONS AND VISIBILITIES
Figure 10 shows the variability in areal and linear extent of the HK
area during one of the major episodes (June 18-July 13, 1975), as well as
128
-------
associated visibilities. Area! extent is the total overland area encompassed
by isochrones, which sometimes imply continuation over adjacent water. This
episode commences along coastal Virginia with a typical 10s-km2 area, and
expands as it moves westward. The areal coverage increased by an order of
magnitude to 106 km2 during the first 10 days. At the time of maximum size
(July 3, 1.5 x io6 km2), the blob extended from the coast of the mid-Atlantic
States west-northwestward to the eastern Dakotas. The general pattern of
decline in areal extent was similar to that of the rise. One should note the
occasional rapid fluctuations in size that verify one's impressions of day-
to-day expansions-contractions, separations-recombinations of haze areas.
Maximum and minimum linear dimensions of overland HK areas have also
been tabulated. No matter what the actual blob shape, some measure of length
(major axis) and width (minor axis) was attempted for the largest blob. When
the blob extended out to sea, the major axis would often be the distance
along the coast and the minor axis the inland extent.
Figure 10 plots of major and minor axes show a more or less steady
increase from IO2 km to IO3 km during the first half of the period. Axis
ratios vary from about 1:1 (nearly circular) to 2:1, with 3:2 being most
frequent. The long and narrow shape at the time of maximum areal extent and
near the episode end (July 8-10) are to be noted. The latter result because
large haze areas extend into the Atlantic.
Areal mean and absolute minimum visibilities over the episode duration
are also shown in Figure 10. The mean visibility for each day is a subjec-
tively determined, weighted mean of visibilities inside isochrones. The
minimum visibility is the lowest reported value associated with haze/smoke.
One should keep in mind that haze/smoke cannot be reported unless the visibil-
ity is 6 miles or less.
The inverse relationship between visibility and areal extent is notable--
as the haze area grows it also becomes more "intense." Lowest mean visibili-
ties are 3 miles; the overall mean for this case is close to 4 miles. Minimum
visibilities are usually confined to small regions and the absolute minimum
most often reported by one station. During the "mature state" in episode
development, local minima are frequently 2 miles or less and occasionally as
low as 1 mile.
129
-------
Some statistical characteristics of HK blob dimensions and visibilities
are presented in Figures 11 and 12. Here September 1973-77 is used as an ex-
ample. The histogram in the former figure is percent frequency of haze area
in categories of 10s km2 centered on the value indicated. Total haze days in
September was 85. The 50 percentile value is around 4.5 x 105 km2 (about 6°
latitude square). There appears to be only a 1 in 10 chance of the HK area
equaling or exceeding 106 km2 (about 9° latitude square). The significance
of the trimodal character of the frequency distribution is not clear.
Major and minor axis measurements (° latitude) are represented as a
scatter diagram in Figure 11. The average major axis is 9.8° latitude (1,088
km) long, the average minor axis about 4.5° latitude (500 km) long. An eye-
estimated line of best fit passing through the mean (solid line) represents
the most probable minor axis length given the major axis. This would be a
crude estimate at best because of the large scatter about the line.
The relative frequency (percent) of lowest visibilities during September
is shown in Figure 12. Visibility categories are at 1-mile intervals centered
on the values shown. The mean, median, and mode are, respectively, 2.6, 2.5,
and 2.0 miles.
HK BLOB TRAJECTORIES
The evolution of HK blobs, as controlled by synoptic-scale flow patterns
and weather distributions, is recognized as an important problem whose solu-
tion could lead to better forecasts of the onset, migration, duration, and
intensity of EPE's. One aspect of the problem has received attention in
recent years, namely, the following of well-defined pollution and/or HK blobs
by means of trajectories—the "long-range transport" problem.
Case studies of long-range transport have, on occasion, provided insight
into the association of HK blobs and the synoptic situation. Hall et al.
(1973) and Wolff et al. (1977) have noted that pollution blobs, usually
formed in association with an anticyclone, will persist as the initiating
circulation system moves on. Wilson (1978) emphasizes the special challenge
of studying chemical transformations and pollution concentration variability
within migrating hazy air masses.
130
-------
It is hoped that a continuation of the present project will produce
statistical relationships among HK areas and synoptic-scale pressure and
weather distributions. At the moment, an interesting case study of a blob
trajectory and the interacting pressure systems that produced it, will be
presented. First, however, a comment about HK blob trajectories.
RTI studies, among others, point out that many factors contribute to the
formation, maintenance, and dissipation of large, hazy air masses. The
factors include available source regions, chemical transformations related to
sunshine and humidity (perhaps temperature also), the wind field with associ-
ated small-scale mixing and large-scale convergence, and the vertical mixing
depth, itself dependent on several processes.
Parts of a large HK blob may be translated by the local wind, but the
area as a whole is more likely a propagation phenomena resulting from the
interaction of many processes. Thus, blob trajectories and air trajectories
originating in the same general area may not be the same. What feature of a
large polluted air mass, then, can be identified and tracked? It was decided
to try to construct the trajectory of the "center of mass" or "centroid."
The centroid of an HK blob was determined subjectively, but carefully,
from its outline by attempting to balance areas within the blob. If multiple
blobs existed on a particular day, the centroid of each was determined, and
the area-weighted mean center of the entire system determined. It was satis-
fying to find that the individual daily centroids did fit together in a
reasonably continuous manner. The centroid trajectory for the June 18-July
13, 1975, case is shown in Figure 13. This is the same case described in the
previous section, some aspects of which were presented in Figure 10.
Daily centroid positions are plotted in Figure 13. There was a general
westward movement from June 18-July 1, complicated by a series of anticyclonic
loops. The last such loop lasted for about 6 days, June 28-July 4. The cen-
troid then executed a cyclonic U-turn across the southeastern States and
headed northward along the Atlantic coast. It dawdled for about a week (July
7-13) around the Virginia-Maryland-Delaware region before dissappearing at
almost the same location where it had begun nearly a month earlier!
What produced such an interesting trajectory? Indeed, does it make
sense? The episode began in the climatologically hazy area of coastal Virginia
131
-------
(see Figure 4). The initial HK area was associated with a weak anticyclone
(1,022 mb) located over eastern Tennessee. The relationship of the centroid
trajectory to this circulation system and others is shown in Figure 14.
The initiating or "parent" anticylcone (the abbreviation PA is appropri-
ate) moves westward over the next few days, essentially disappearing as a
1,019-mb high over western Kentucky on the 21st. The haze area during this
time performed a small anticyclonic loop so that it eventually headed in the
same direction as the PA.
Anticyclogenesis occurred over Lake Ontario on the 20th-21st and this
system became the second dominant anticylone (A2) in the life of the migrating
HK blob. In response to A2's southeastward movement and circulation pattern,
the blob centroid moved northward as far as Lake Erie. As A2 moved out to
sea, a new high formed in eastern Tennessee (A3), and the HK area responded
to this system's movement and circulation by heading southward and completing
a second anticyclonic loop.
As A3 weakened, a moderate polar high (A4) moved rapidly eastward north
of the Great Lakes and took control of the HK blob. In the easterly and
southeasterly flow south of A4, the blob moved westward. Around the 26th,
yet another anticyclone (A5) took over. Third and fourth anticyclonic loops
were completed before A5 relinquished control to a developing high on the
Texas-Oklahoma border (A6).
A6 dropped south and then eastward into the central Gulf. The HK centroid
dutifully followed until July 6. For the next week, until the episode ended,
a nearly stationary pressure pattern formed along the east coast—a north-
south trough along the Appalachians and a high pressure ridge extending from
the Gulf northeastward into the Atlantic. The general southerly flow along
the coast stretched the haze area northward beginning on the 6th. A combina-
tion of large-scale ascent in the vicinity of the trough and the convective
mixing and precipitation scavenging associated with rain showers eroded the
haze area until it finally disappeared.
Thus, a succession of six anticylones influenced the movement of the
haze area in this case. Perhaps the story should be recounted in terms of
cyclones and fronts in order to clarify their large-scale interaction with
132
-------
haze areas. Also, the role of other processes, such as those mentioned above,
should be examined to determine why the anticylonic environment seemed to be
favorable to the formation and long-term maintenance of hazy air masses.
ASSOCIATION OF HK EPISODES AND STAGNATING ANTICYLONES
The association of the HK episodes with the large-scale weather patterns-
lows, highs, and fronts—remains to be investigated. However, Korshover's
(1975) listing of stagnation periods (4 or more days' duration) accompanying
warm anticyclones was available, and included a Z^-year overlap (June 1973-
December 1975) with the present project. The HK episode decisions (Table 1)
were made independently of any considerations of weather pattern associations.
It thus appeared to be of interest to examine the association between the two
sets of data.
To do so, the 2^-year period was divided into weekly intervals starting
with June 3, 1973. Each week was assigned a "yes" (y) or "no" (n).depending
on the presence of a stagnating anticyclone, and HK episodes were treated
similarly. When either situation spanned 2 consecutive weeks, both weeks
were credited with a "yes" if the condition existed for at least 2 days of
that week.
The results for the warm seasons of each year (June-October, 1973; May-
October 1974 and 1975) are presented as two-way contingency tables in Figure
15, Case A. The numbers in the table are the numbers of weeks of the associ-
ation indicated for the box. The smaller printed numerals are row and column
totals and the grand total. Thus, in Case A there is a total of 76 weeks, 46
of which included HK episodes and 26, stagnating highs. Perfect correspond-
ence between the two phenomena would find numbers only in the boxes y-y and
n-n.
Following standard statistical procedures (Panofsky and Brier, 1958) a
"No Relation" table was constructed. In this case, the "No Relation" table
is the same as the "Observed".' If Chi-Squared is used to test the signifi-
cance in any difference between the two tables, x2 = 0. Thus the null hypoth-
esis of no relationship could not be rejected at any level of confidence.
This would imply that stagnating anticyclones are not sufficient—not even
necessary—for the occurrence of HK episodes.
133
-------
It was thought, however, that some of the lack of association may be due
to the inclusion of 1973, which, as mentioned previously, contained nearly 4
months of continuous haze. Thus, there are many consecutive weeks of "yes"
for HK occurrence, but few for stagnating anticyclones. Case B in Figure 15
is similar to Case A except that only May-October 1974 and 1975 are considered.
This time there is some difference between the "Observed" and "No Relation"
tables and x2 = 1-3. For 1 degree of freedom, this x2 value occurs at the 25
percent confidence level. Thus, one would take considerable risk in rejecting
the null hypothesis in this case also.
The problem in this surprising result may lie in incompatible definitions
of stagnating anticyclones and/or in HK episodes. Indeed, Altshuller (1978)
found that in associating elevated oxidant levels with Korshover's stagnating
highs over a 10-year period, there were many instances of the former occurring
at times when the stagnation criteria were not satisfied. Perhaps, then,
what is being indicated is that other meteorological factors must also be
considered.
CONCLUDING REMARKS
In the brief 3 months of work on this project, some headway has been
made in compiling 5-year statistics on the evolution of haze/smoke episodes
and their meteorological context. Identification of episodes and the compos-
iting of monthly statistics on the geographical distribution of HK occurrence
and persistence and of dimensions and associated visibilities has occupied
most of the time. These materials are available in the final report to RTI.
Studies of the interaction of hazy air masses with the synoptic-scale
flow and weather distributions have been initiated. The development of
statistical models would seem to require detailed case studies and thoughtful
consideration to means of compositing individual results. The statistical
association of haze episodes with elevated oxidant and other pollutant levels
should be determined. This may shed further light on the relative contribu-
tions of natural and manmade sources to haze production.
The speaker, for one, looks forward to pursuing these problems.
REFERENCES
Altshuller, A. P. 1978. Association of Oxidant Episodes with Warm Stagnating
Anticyclones. Journ. Air Pollu. Control Assoc., 28:152-155.
134
-------
Hall, F. P., C. E. Duchan, L. G. Lee, and R. R. Hagen. 1973. Long-Range
Transport of Air Pollution: A Case Study, August 1970. Mon. Wea. Review
101:404-411.
Holzworth, G. C. 1974. Meteorological Episodes of Slowest Dilution in Con-
tiguous United States. Environmental Monitoring Series (EPA-650/4-74-002),
NERC-Environmental Protection Agency, Research Triangle Park, N.C., p. 80.
. 1972. Mixing Heights, Wind Speeds, and Potential for Urban
Air Pollution Throughout the Contiguous United States. Office of Air
Programs Publication No. AP-101, Environmental Protection Agency, Research
Triangle Park, N.C., p. 118.
Korshover, J. 1975. Climatology of Stagnating Anticyclones East of the Rocky
Mountains, 1936-1975. NOAA Tech. Memorandum ERL ARL-55, Air Resources
Laboratory, Silver Spring, MD, p. 26.
Panofsky, H. A., and G. W. Brier. 1958. Some Applications of Statistics to
Meteorology. The Pennsylvania State University, University Park, PA,
p. 224.
Robinson, E., F. M. Vukovich, and D. H. Pack. 1978. The Issue of Oxidant
Transport. Journ. Air Pollu. Control Assoc., 28:209-210.
Wilson, W. E. 1978. Sulfates in the Atmosphere: A Progress Report on Project
MISTT. Atmospheric Environment. 12:537-547.
Wolff, G. T. 1977. An Investigation of Long-Range Transport of Ozone Across
the Midwestern and Eastern United States, Atmospheric Environment, 11:
p. 797-802.
135
-------
co
100
100
Figure 1. Examples of haze/smoke episodes. Chart A: minor episode June 22-25,1976. Chart B: major
episode July 4-24,1977. The outer dashed curve is the episode envelope. Number of succes-
sive days (persistence) of haze/smoke is shown.
-------
100
Figure 2. Geographical regions (defined in the text) with the percentage of all
episodes (total 42) in which the region experienced >3 successive days
of haze/smoke.
137
-------
100
Figure 3. Probability (%) of any one day in July experiencing haze/smoke asso-
ciated with an episode. Based on 5 years' data 1973—1977.
138
-------
100
Figure 4. Probability (%) of >4 consecutive days of haze/smoke occurring
at least once in July.
139
-------
100
Figure 5. Number of consecutive days of haze/smoke in July with a probability
of 60 percent
140 '
-------
100
Figure 6. Number of consecutive days of haze/smoke in July with a probability
of 100 percent.
141
-------
100
Figure 7. General areas in which persistent haze/smoke lasts for a week or more
at least once in 5 years.
142
-------
"-u
100
Figure 8. Illustrating interannual variability. Isopleths of >4 consecutive days
of haze/smoke for the years 1973, 1975, and 1977.
143
-------
100
Figure 9. Illustrating monthly variability. Isopleths of 60 percent probability
of >4 consecutive days of haze/smoke in the 3 months June (JN),
July (JL), and August (AU).
144
-------
=i
%
i
§
g
03
OO
2,0
1,5
1,0
,5
0
2,0
1,5
1,0
,5
0.
5
3
2
1
0
V ^ B__-
y
J rr •-,
-> ' a
/ Vy \
>'"'"" /' ^
15 20 25 30 5 ID 11
JUNE JULY
MAJOR AXIS ^
-> — O — *> --^ 0 / \
^ v\ ' ' 'i
/ * "
3^-o-- 'f/ * '^SX^6' '• /\ ' ^
o— Q /' ' MINOR AXIS -\ 0'-
,* "-o .'
^v *- -
| 1 1 1 1
L5 20 25 30 5 10 15
JUNE JIJLY
X • X X XX'
'•/' /^- /\ '
- rV— ox v y > *• * > i 'rf X * - * X • O ^^— .
*j v A. ^ , f^AN \ /
v / "*, •'
X— x • X-A-* •' 'X— X 'x X
3 6 p --O--0 0 O— 0 / 9--O
o' MIN o . 7
0
13 o
1 1 i 1 1
[5 20 25 30 5 10 15
Figure 10. Evolution of areal and linear dimensions of HK areas during the major
episode June 18—July 13, 1975. Associated mean and minimum visibil-
ities are also shown.
145 '
-------
20
10
o
o
10
12
AREA (105 KM2)
8
CO
0.
0
•3
O
8
12
16 .
20
' MAJOR AXIS (9 LAD
Figure 11. Upper chart: histogram of relative frequency of overland area! coverage
of haze/smoke blobs for all episodes, September 1973—1977. Lower
chart: scatter diagram of major versus minor axes of haze/smoke blobs
for the same 5-year period. The solid line is an eye-estimated line of
best fit.
146
-------
50 —
30-
FREQUB1CY
20-
10 —
0
46
27
19
2 3
VISIBILITY (MI)
5
Figure 12. A relative frequency histogram of minimum visibilities during haze/smoke
episodes, September 1973-1977.
147
-------
100
Figure 13. The centroid trajectory of the haze/smoke area during the June 18—
July 13, 1975, episode. One-day intervals are indicated and selected
dates shown. The outer dashed curve is the episode envelope.
148
-------
I
1
1
!•_--
\ '
\ A6 ^,
\ *
Y J
(
/ \
( !•
'\ i
•/
- -n ^
J^^
^
X
y^
100
Figure 14. Showing the association of the June 18—July 13, 1975, episode's centroid
trajectory with interacting anticyclones. PA represents the path of the
"parent" anticyclone; A2,..., A6 are other anticyclones successively
controlling the haze trajectory during periods of corresponding symbols.
Selected dates are indicated, and central pressures are underscored.
149
-------
OBSERVED
O R E L A T I 0 II
STAGNATING HIGH
y n
y is
H Y
n 10
30 n
STAGNATING HIGH
y , n
6 16
20 30 10
30
20
26 ' 50 76 26 ' 50
X2=o
y 12
U V ,, „ _.
n 7
16 2
A
3 10
19 26 9
18
17
19 35 54 19 ' 35 •
70=1.3
B
Figure 15. "Observed" relationship between the occurrence of warm stagnating high
pressure systems and haze/smoke (HK) episodes. The "No Relation" table
results from assumed random association. Case A: June-October 1973
plus May-October 1974 and 1975. Case B: same as A excluding 1973.
See the text for further discussion.
150
-------
TABLE 1. HAZE/SMOKE EPISODES EAST OF THE ROCKY MOUNTAINS IN THE 5-YEAR PERIOD
JUNE 1973-MAY 1978 (see the text for explanation)
^Ftonth
YfiaT~~~ _
1973
1974
1975
1976
1977
1978
i Episodes/
Duration
Range
(days)
J.in
/
0
Feb
/
11-15
C
1-5
C NC
~9^12 *
. C HE
15-22 *
HE EC
4/4-7
Mar
/
0
Apr
/
15-21
HE
1/7
Mny
/
4-9
SC
17-26
NE EC
16-24
EC
30-i
NE EC
22
NC NE
5/3-16
Jun
HF. EC
28
18 1
C NC NE
8-16
NE EC SE C
22-25
NE
1 *
11-18
NE EC C
>6 /
7/3-21
.l.il
NE KC SE
> 15 *
NE EC
14 - 31
NE EC C
-> 13
EC SE
16-25 *
HE
5-12
EC SE
20-31
NE EC
4-24
NE EC
27
NE EC
/
9/6-31
Aug
NC NK KC SE
7-12
NE
20-28
NE EC
X ift
NE EC
28 —
C.
10-16
NE
20-28
c HE
>i *
11-16 *
C NE
20-28
EC
31-
/
8/6-31
Sop
i >
NR F.C SE SC
'i-13
HE
14-22
SC
>11 *
J£
2-5
KC
15-21
EC SE
>14
C NE
/
7/4-30
Oct
> 4
C
6-13
NK
11-15 *
NF.
29 —
NE
24
NC C
/
4/5-9
Nov
-»3 *
i
/
1/6
Dec
/
0
-------
SOME DYNAMIC ASPECTS OF EXTENDED POLLUTION EPISODES
William J. King and Fred M. Vukovich*
Research Triangle Institute
Research Triangle Park, North Carolina
Abstract
A boundary layer trajectory model was used to compute the residence
time of air parcels within transient high-pressure systems of Canadian
origin traversing the eastern Ohio River Valley. A strong correlation was
noted among the presence of a transient anticyclone, elevated pollution
levels, deteriorating horizontal visibility, and the length of time an air
parcel had been entrained in the boundary layer of the high-pressure system.
INTRODUCTION
A large body of research over the past decade has been devoted to the
"elevated pollution episode" (EPE) and the "persistent elevated pollution
episode" (PEPE). One phase of this work, as performed by the Research
Triangle Institute, has revolved around the investigation of nonurban ozone
buildups, particularly as they relate to high-pressure systems and their
movement (Bach et al., 1976; Vukovich et al., 1977; Decker et al., 1976;
EPA Report No. 450/3-74-034; and EPA Report No. 450/3-75-036.) This extended
research effort has suggested the following major conclusions:
1. Atmospheric conditions in a high-pressure system lend themselves
. to the rapid synthesis of ozone, providing sufficient ozone precur-
sors have been injected into the system.
2. High ozone during the summer months is associated with high-pressure
systems.
3. The concentration of ozone within an air parcel in a high-pressure
system correlates with the residence time of the air parcel within
the system.
^Speaker.
152
-------
4. Air parcels on the "back side" of a moving high-pressure system
have had the longest residence time and contain the highest ozone
concentrations in the system.
5. Maximum residence times within a moving high-pressure system vary
inversely with the speed at which the system is moving.
6. Maximum ozone concentrations in transient high-pressure systems on
the continental United States are higher the farther east (where
the population density is greatest) the system moves, suggesting
the importance of anthropogenic sources of ozone precursor material
to the synthesis of ozone within the high.
There has been limited evidence that pollutants other than ozone reach
high levels in high-pressure systems (Husar et al., 1976, and Vukovich, 1979).
Restricted horizontal visibility due to lithometeor obstructions to vision
(smoke and/or haze) has been suggested as a possible indicator and method of
monitoring the progress of an EPE. This paper presents a preliminary evalua-
tion of the dynamic meteorological conditions in a restricted visibility
state utilizing the same basic research techniques that were developed
during the course of the RTI summer ozone field programs. Comparisons have
been made among the various data components to determine the relative coin-
cidence of peaks in air pollution levels and minimum visibility readings in
the area of the Ohio River Valley. Then the dynamics of several transient
anticyclones are examined to see just how well system residence times com-
pare with in situ visibility and pollution concentration data.
Several surface analyses were performed over a macroscale study area
centered along the Ohio River Valley and including most of the northeast and
north central United States from approximately 36° to 46° north and from 73°
to 92° west. This region was selected because, in the summer of 1974, a
number of transient anticyclones traversed the area during periods when two
major air pollution monitoring programs were operating. The Research Triangle
Institute's Summer Ozone Study field network stations were operating at
Wilmington, McConnellsville, and Wooster, Ohio; DuBois, Pennsylvania; and
McHenry, Maryland. At the same time, the Electric Power Research Institute's
SURE network was taking 24-hr average measurements of sulfate, sulfur dioxide,
and total suspended particulate. The SURE stations were located at Rockport,
153
-------
Madison, and Lawrenceburg, Indiana; Collins, Illinois; Huntington and Wheeling,
West Virginia; Scranton, Pennsylvania; and Albany, New York.
Weather data for the period including visibility, present weather,
station pressure, dry bulb and dew point temperatures, surface analyses, and
anticyclone movement were obtained from U.S. Department of Commerce publica-
tions such as Daily Weather Maps; Climatological Data, National Summary; and
Local Climatological Data summaries for 61 National Weather Service (NWS)
first-order stations over an 18-State area.
In addition to the primary study area, a 500 x 500 km square mesoscale
study area centered around Columbus, Ohio, was used for certain detailed
analyses including the effects of system residence time upon pollution
levels and horizontal visibility.
PEAKS IN STATION PRESSURE, AIR POLLUTION, AND RESTRICTED VISIBILITY
A simple test was devised for comparing the relative coincidence of
occurrence of maximum values of station pressure, various pollutant concen-
trations and restricted visibility. The data for this test were taken from
the mesoscale study area previously described and were compiled as follows:
1. Station pressure—the area-averaged, daily average station pres-
sure (inches of mercury) computed from the following cluster of
NWS stations in northern Ohio: Akron, Cleveland, Columbus, and
Youngstown;
2. Restricted visibility—the area-averaged, 1700 EOT visibility
reported at the following cluster of NWS stations: Akron, Cleve-
land, Mansfield, and Youngstown. The 1700 EOT visibility was used
in order to remove the effects of morning fog.
3. Ozone—the area-averaged, 8-hr average (1200-2000 EOT) ozone
~3
concentration (ug m ) computed from the following RTI ozone
monitoring sites in Ohio: Wilmington, McConnellsville, and Wooster.
4. Sulfate, sulfur dioxide, and total suspended particulate (TSP)--the
area-averaged, 24-hr average concentration (ug m ) computed from
the following EPRI SURE stations: Lawrenceburg, Indiana; Huntington
and Wheeling, West Virginia.
Numbering the days consecutively beginning with 16 June 1974 (day 1)
and running through 31 August 1974 (day 77), the day relative to day 1 when
154
-------
peaks were reached In each of the above listed parameters are given in
Table I.
From these data, station pressure seems to peak generally before the
other parameters, particularly restricted visibility. This is precisely the
predicted relationship: the high-pressure system moves through an area; the
pressure rises and peaks as the center passes the monitoring station; during
this time, pollution levels rise and visibility deteriorates with peaks
being reached in the air on the "back side" of the high. In order to deter-
mine which of these parameters most closely follows the trend in restricted
visibility, the data in Table 1 are used to compare the date of occurrence
of maximum restricted visibility relative to the date of occurrence of peaks
in the other parameters (Table 2).
These 14 cases suggest that maximum restricted visibility correlates
best with peaks in sulfate, ozone, and total suspended particulate. Also,
the primary pollutant, sulfur dioxide, appears to reach a maximum concentra-
tion in advance of the other pollutants and then tapers off, most probably
because of its reactive nature. This contrasts with the results of Vukovich
(1979), which suggest that the S02, S04, and TSP peaks occur simultaneously.
Noteworthy is the fact that, with only three exceptions, minimum visibility
was reached at the same time or after each of the selected pollutants peaked.
FOUR CASE STUDIES
In order to examine the relationships of visibility to air pollution in
more detail, the following case study periods were defined:
July 9-15, 1974 (days 24-30)
July 15-19, 1974 (days 30-34)
July 19-23, 1974 (days 34-38)
August 4-8, 1974 (days 50-54)
Each case study covers a period between major frontal passages when a transient
anticyclone of Canadian origin traversed the Ohio River Valley including the
500 x 500 km mesoscale study area. It has been found that west to east
moving Canadian high-pressure systems remain relatively free of pollution
buildups until they penetrate the industrialized north central United States
(Bach et al., 1976; Vukovich et al., 1977). In every case, the center of
155
-------
the anticyclone passed to the north of Columbus, Ohio; three times within
200 km and once within 400 km.
As suggested earlier, as long as an air parcel remains in the domain of
a moving high-pressure system, potential exists to develop higher concentra-
tions of certain pollutants. This is particularly true of secondary pollut-
ants such as ozone and sulfate. Figure 1 shows the residence time of air
parcels within a circular symmetric anticyclone relative to the speed at
which the system is moving. These residence time representations evolved
from a boundary layer trajectory model (Vukovich et al., 1977). This same
model was used to calculate the residence time isopleths within the 500 x
500 km study area for each day of each case study period. There follows a
discussion of the results of this modeling effort and the general meteoro-
logical conditions for each case study.
July 9-15 Case Study
A rather intense anticyclone developed on July 9 to the northwest of
Hudson Bay and began immediately moving to the south. An idealized repre-
sentation of the system's location and direction of movement each day at
1700 EOT is given in Figure 2. For simplification, the anticyclone is
represented by a circle with a 1,000-km radius. System speed and approxi-
mate central pressure are also indicated.
On day 29, the high-pressure system had stalled and was beginning to
exhibit the first signs of anticyclolysis. The approximate position of the
system on day 29 is indicated by the dashed circle centered in the extreme
southeast corner of the Columbus study area. The system had totally dissi-
pated by the following morning.
Figure 3 shows the 500 x 500 km study area centered on Columbus, Ohio.
The isopleths indicate the residence time of air in the high-pressure system.
Pollutant and meteorological milestones are noted. As the high-pressure
system advanced, it slowed steadily, the center traversing the northern
portion of the study area on day 27 (Figure 2). On days 28 and 29, the air
over the study area was on the "back side" of the system in the region of
longest residence times (Figure 3). The total suspended particulate concen-
tration peaked out on day 28, with visibility and each of the other pollu-
tants following suit on day 29 (Figure 4).
156
-------
On day 30, a weak cold front traversed the study area. The frontal
zone showed little temperature variation but could be identified by an
elevated dew point temperature and a 180° wind shift across the frontal
zone. However, surface winds remained relatively light both behind and
ahead of the front. As the front passed, the pollutant concentrations
decreased. The visibility increased, except in an isolated region around
Lake Erie extending south into West Virginia where smoke and haze continued
to restrict visibility after the front had passed and well into the evening
of day 30 before clearing. Why this anomaly occurred is not readily apparent
from the available data.
July 15-19 Case Study
On July 15 (day 30), a ridge of high pressure developed over the north
central United States and central Canada with its center northwest of Lake
Superior. Figure 5 documents the progress of this large anticyclone, the
center of which traversed the northern portion of the Columbus study area on
the evening of day 31, picking up speed as it progressed. By the afternoon
of day 33, the center was off the coast of Virginia. Although this idealized
representation indicates the system had passed out of the Columbus area, the
actual synoptic picture shows the system to be considerably larger with the
study area still on the "back side" of the high. Figure 6 shows the residence
time modeling analysis.
Of the four case studies, this July 15-19 anticyclone best exhibits the
relationships suggested by the research in the previous section about visibil-
ity deterioration, anticyclone movement and system residence time. The
reactive primary pollutant, sulfur dioxide, peaks early (day 31), followed
the next day by the pressure peak as the anticyclone center traverses the
study area. On day 33, the study area is on the "back side" of the high
where air has reached maximum residence time within the system. On this
day, visibility reached a minimum and the sulfates, total suspended particu-
late, and the secondary pollutant ozone, which is normally associated with
photochemical smog, reached a maximum. Figure 7 shows the pattern of re-
stricted visibility on day 33 on the "back side" of the high. By day 34,
the strong cold front shown in Figure 7 had moved through the study area.
The passage of this front, which was accompanied by rain showers and a
157
-------
pronounced wind shift, resulted in a dramatic reduction in air pollution
levels and a corresponding increase in visibility.
July 19-23 Case Study
An anticyclone formed in the Beaufort Sea off the north shore of Alaska
on July 13. It then took 6 days to make its way southeastward across Canada,
entering our study area (Figure 8) on July 19 (day 34). At this point, par-
tially blocked by a low off Nova Scotia, the high slowed and turned eastward,
the center passing just north of the study area. Upon reaching the New England
coast, the anticyclone again accelerated and continued up the coastline
finally dissipating south of Labrador.
The residence time modeling analysis (Figure 9) exhibits some rather
interesting milestones. Although total suspended particulate peaked early
(day 35), TSP levels remained near that peak through day 38. This EPE did
not terminate due to a frontal passage, but rather due to anticyclolysis on
day 38. On day 37, the "back side" of the high was still quite well defined.
Both the fairly reactive pollutants, ozone and sulfur dioxide, reached
peaks. As dissipation set in on day 38, the system became disorganized with
pressure dropping rapidly in the study area and some rain shower activity
noted. Ozone and sulfur dioxide concentrations decreased sharply while
sulfate reached a peak. Visibility dropped again significantly and the
sulfates increased on day 38.
August 4-8 Case Study
This broad anticyclone formed just south of Victoria Island the evening
of July 31 moving south and southeast until it was centered over northern
Nebraska the morning of August 4. It then turned 90 degrees and headed
almost due east, dissipating off the New England coast on August 8 (Figure
10). The residence time analysis is shown in Figure 11. The sulfur dioxide
peaked on day 52, and the other parameters on day 53 (Figure 12).
AIR MOTION DYNAMICS, RESIDENCE TIME, AND ELEVATED POLLUTION EPISODES
Figure 13 shows idealized trajectories of air parcels in the boundary
layer in high-pressure systems. Two out of the three trajectories show air
parcels that leave the boundaries of the high-pressure system very quickly.
One of the air parcels, that which originates in the northeastern quadrant
158
-------
of the high-pressure system, travels hundreds of miles in the domain of the
high-pressure system due to the circulation within that system. It is
interesting to note that though that air parcel has traveled hundreds of
miles, it passes over its own trajectory only a few tens of miles away from
its origination point. It is this air parcel that has the large residence
time within the high-pressure system and is most likely to be associated
with elevated pollution episodes within the high-pressure system.
The northeastern quadrant of a high-pressure system is generally charac-
terized by northwesterly flow (Figure 14) and is the region of the high-
pressure system that characterizes post-cold-frontal conditions. Generally
low levels of pollutants are found in this air mass for two reasons: there
is rainout of the pollutants and deep mixing associated with the front. The
air mass behind the front is generally cold, dry Canadian air, which is
relatively free of air pollutants. Shenfeld (1977) has shown relatively low
concentrations of oxidants to exist in the forest lands of Canada where this
air mass originates. As this air moves southward, it is modified through
the injection of heat and moisture at the surface. Furthermore, the concen-
tration of primary air pollutants increases in the air mass due to injection
from each city and town, which act as isolated point sources. Chemical
reactions take place and some primary pollutants are gradually transformed,
creating secondary pollutants within the air mass. As these reactions
continue, the concentrations of the secondary air pollutants increase.
However, as the high-pressure system continues to move eastward, the
region comes under the influence of the air on the "backside" of the high,
which is generally flowing from the south (Figure 15). This southerly flow
is warm and moist and, depending upon its trajectory, may be relatively free
of air pollutants. Decker et al. (1976) has shown that air parcels that
have a long fetch over water are generally free of air contaminants. Their
study was in the Gulf of Mexico where the warm, moist air on the "back side"
of the high generally originates. As this air mass moves northward, it is
modified by two processes. First there is a transport of heat and moisture
at the surface; and also there is entrainment of the cooler, dryer, generally
more polluted air ahead of this air mass. The entrainment of the polluted
air and the injection of pollutants into the air mass at the surface from
159
-------
population centers in the region increases the potential for producing an
elevated pollution episode. It is important to note that elevated pollution
episodes generally occur in the warm, moist air mass (see Bach (1975) for a
correlation among moisture, temperature, and high levels of ozone). Water
vapor appears in chemical reactions associated with the production of ozone
and certain sulfur compounds. It is also an important factor in producing
larger aerosols that might impede visibility.
Because residence time in high-pressure systems is a function of system
speed (Figure 1), it is expected that the persistent pollution episodes
occur in relatively slow moving highs or stationary high-pressure systems.
In a stationary high-pressure system, there is generally very little exchange
between the air within the high-pressure system and that outside the high-
pressure system. Under these circumstances, the injection of primary air
pollutants and the production of secondary air pollutants would continue
unabated, particularly since a high-pressure system is relatively free of
cloud cover that would interfere with solar radiation reaching the surface.
REFERENCES
Bach, W. D., W. J. King, and F. M. Vukovich. 1976. Nonurban Ozone Concentra-
tions in Transient High Pressure Systems. Presented at the Symposium on
the Non-Urban Troposphere, Hollywood, Florida, November 14.
Decker, C. E., L. A. Ripperton, J. J. B. Worth, F. M. Vukovich, W. D. Bach, J.
B. Tommerdahl, F. Smith, and D. E. Wagoner. 1976. Formation and Trans-
port of Oxidants Along the Gulf Coast and in the Northern United States.
Environmental Protection Agency, Report No. EPA-450/3-76-033, Research
Triangle Park, NC.
Husar, R. B., D. E. Patterson, C. C. Paley, and N. V. Gillani. 1976. Ozone
in Hazy Air Masses. Presented at the International Conference on Photo-
chemical Oxidant Pollution and Its Control, Raleigh, NC, September 12-17.
Environmental Protection Agency. 1974. Investigation of Ozone and Ozone
Precursor Concentrations at Non-Urban Locations in the Eastern United
States. Report No. EPA-450/3-74-034, Research Triangle Park, NC.
Environmental Protection Agency. 1975. Investigation of Rural Oxidant Levels
as Related to Urban Hydrocarbon Control Strategies. Report No. EPA-450/
3-75-036, Research Triangle Park, NC.
Shenfeld, L. 1977. Report on Oxidants and Their Precursors in Canada.
Proc. Int. Conf. on Photochemical Oxidant Pollution and its Control.
EPA-600/3-77-001A, Research Triangle Park, NC, pp. 917-927.
160
-------
Vukovich, F. M., W. D. Bach, B. W. Crissman, and W. J. King. 1977. On the
Relationship Between High Ozone in the Rural Surface Layer and High
Pressure Systems. Atmospheric Environment. 11:967-983.
Vukovich, F. M. 1979. A Note on Air Quality in High Pressure Systems.
Atmospheric Environment, 13:255-265.
161
-------
-i
cx»7.5ms
"'
BM«»
IOms"1
1977 by Pergamon Press. This figure originally appeared in an article
by F. M. Vukovich in volume 11 of Atmospheric Environment, p. 973.
Figure 1. The number of days air parcels in high-pressure systems have spent
(residence time) within the system versus the speed of the system, for
a circular symmetric high-pressure system with a radius of 1,000 km.
162
-------
CENTRAL
DAY SPEED PRESSURE
(m-s-i) (mb)
1023
1025
1024
1024
1024
Disp.
25 9.9
26 7.7
27 6.9
28 5.4
29 4.1
30 —
v
"H--
../
_>-O*.
. • N -! , U ^ T;
Figure 2. Location and speed of movement at 1700 EDT each day for the July 9-15 case study.
Also tabulated are 24-hr average speed for the system and central pressure at 1700 EDT.
163
-------
DAY
RESIDENCE TIME (days)
MILESTONES
25
<1
26
27
Pressure Peak
28
<3
<4
TSP Maximum
29
30
VSBY Minimum *
504 Maximum
63 Maximum
$62 Maximum
Secondary
VSBY Minimum
Figure 3. Residence time (days) for each day of the July 9-15 case study in the
500 X 500 km Columbus study area.
164
-------
Figure 4. July 14 (day 29} surface weather analysis showing visibility isopleths (miles) and the
location of pressure centers and fronts. Also shown are 8-hr average ozone values (jug- m~ 3)
from the RTI network and 24-hr average values for SC>2, TSP, and 864 (/ig- m~ 3) from the
EPRI SURE network.
165
-------
r
CENTRAL
DAY SPEED PRESSURE
(m-s-1 ) (mb)
30
31
32
33
34
6.4
8.6
11.8
10.8
12.3
1022
1023
1026
1025
1024
•X
±f •
^
X
\
Figure 5. Location and speed of movement at 1700 EOT each day for the July 15-19 case study.
Also tabulated are 24-hr average speed for the system and central pressure at 1700 EDT.
166
-------
DAY
RESI DEIMCE TIME (days)
MILESTONES
30
31
S02 Max! mum
32
Pressure Peak
33
VSBY Minimum
50)4 Maximum
63 Maximum
TSP Maximum
34
Figure 6. Residence time (days) for each day of the July 15-19 case study in the
500 X 500 km Columbus study area.
167
-------
Figure 7. July 18 (day 33) surface weather analysis showing visibility isopleths (miles) and the
location of pressure centers and fronts. Also shown are 8-hr average ozone values (ng- m~ 3)
from the RTI network and 24-hr average values for SO2, TSP, and 864 (jug- m~ 3) from the
EPRI SURE network.
168
-------
CENTRAL
DAY SPEED PRESSURE
(m-s-i ) (mb)
34
35
36
37
38
12.7
6 A
5.9
10.3
16.4
1027
1025
1024
1024
1025
\ ' -^-"X ,
\ rr \ *v
^-•\ i ..!--Ax'
\ .'-^-x. ; x \
• '^. • ,
\ • ' "^
Figure 8. Location and speed of movement at 1700 EDT each day for the July 19-23 case study.
Also tabulated are 24-hr average speed for the system and central pressure at 1700 EDT.
169
-------
DAY
RESIDENCE TIME (days)
MILESTONES
34
35
Pressure Peak
TSP Maximum
36
37
S02 Maximum
0«3 Maximum
38
VSBY Minimum
804 Maximum
Figure 9. Residence time (days) for each day of the July 19-23 case study in the
500 X 500 km Columbus study area.
170
-------
CENTRAL
DAY SPEED PRESSURE
(m-s-1) (mb)
50 7.1 1022
51 9.0 1022
52 9.9 1023
53 6.4 1023
54 — Disp.
\
Figure 10. Location and speed of movement at 1700 EDT each day for the August 4-8 case study.
Also tabulated are 24-hr average speed for the system and central pressure at 1700 EDT.
171
-------
DAY
RESIDENCE TIME (days)
MILESTONES
50
51
52
SOo Maximum
53
Pressure Peak
VSBY Minimum
50)4 Maximum
Og Maximum
TSP Maximum
54
Figure 11. Residence time (days) for each day of the August 4-8 case study in the
500 X 500 km Columbus study area.
172
-------
-Hi-
35°-
Figure 12. August 7 (day 53) surface weather analysis showing visibility isopleths (miles) and the
location of pressure centers and fronts. Also shown are 8-hr average ozone values (/ag- m~ 3)
from the RTI network and 24-hr average values for SC>2, TSP, and 804 (/zg- m~ 3) from the
EPRI SURE network.
173
-------
/
2
\
1 J
/ 3
1
\
1 J
f/' I
J
j
Figure 13. Three idealized trajectories of air parcels in the boundary layer of a
moving high-pressure system.
174
-------
Figure 14. Leading edge of a high-pressure system.
175
-------
WARM MOIST
RELATIVELY
CLEAN
AIR
Figure 15. Trailing edge of a high-pressure system.
176
-------
TABLE 1. THE DAYS VARIOUS PARAMETERS REACHED PEAKS (JUNE 16 = DAY 1)
Station pressure
4
18
22
26
32
35
40
47
53
56
61
66
67
71
Restricted
visibility
5
18
24
30
33
38
42
48
53
56
62
67
69
73
Sulfate
5
18
24
29
33
38
42
47
53
56
61
66
69
72
Ozone
5
18
25
29
33
37
41
48
53
56
61
66
70
73
TSP
5
18
24
28
33
35
42
47
53
56
61
66
69
72
S02
7
18
24
29
31
37
41
-
52
55
61
66
69
72
TABLE 2. COMPARISON OF THE DATE MAXIMUM RESTRICTED VISIBILITY OCCURRED WITH
THE DATE OF OCCURRENCE OF PEAKS IN SULFATE, OZONE, TOTAL SUSPENDED
PARTICULATE, SULFUR DIOXIDE, AND STATION PRESSURE
Number of times maximum restricted visibility occurred
Peaks
Sulfate
Ozone
TSP
S02
Pressure
2 days 1 day Same
before before day
9
2 7
8
1 3
3
1 day
after
5
5
4
8
5
2 days
after
1
1
4
3 days
after
1
1
4 days
after
1
177
-------
APPLICATION OF NUMERICAL MODELS TO
PROLONGED ELEVATED POLLUTION EPISODES
C. Shepherd Burton* and Mei-Kao Liu
Systems Applications, Inc., San Rafael, California
Abstract
Many field experiments have been carried out recently to study the problem
of pollutant transport over long distances under prolonged pollution episode
conditions. The use of a numerical model designed to examine this problem is
discussed in this paper. This model is composed of a plume module for treating
the transport, dispersion, and chemical transformations of a limited number
of plumes near major point sources, and a grid module for accommodating the
advection, diffusion, chemical reactions, and surface deposition of pollutants
from upwind boundaries as well as urban, and other widely scattered, sources.
This model can be employed as a tool in designing a monitoring network, or it
can be used to analyze and interpret data collected in the field measurement
program.
INTRODUCTION
During the past few years, considerable effort has been devoted to the
understanding of the problem of long distance transport of air pollutants from
either major point sources or urban areas under prolonged pollution episode
conditions. Several comprehensive field measurement programs have been carried
out to characterize the problem. For example, the 1975 Northeast Oxidant
Experiment was conducted to examine the role played by ozone and its precursors
from upwind sources on many high oxidant concentrations observed in rural areas
in the Northeast Corridor (Dimitriades and Altshuller, 1977). Another example
is the Sulfate Regional Experiment, which was deployed to assess the relation-
ships between sulfur oxide emissions and ambient air quality on regional scales
(Hidy et al., 1976). More recently, the Midwest Interstate Sulfur Transformation
^Speaker.
178
-------
and Transport (MISTT) study was performed to study the detailed distribution of
sulfur oxides downwind of a large coal-fired power plant (Wilson, 1978).
The MISTT study has provided a rich data base to aid in the understanding
of the various atmospheric processes that dictate the transformations and
distributions of sulfur oxides emitted from a coal-fired power plant. For
example, Husar et al. (1978) have found that transport of pollutants is governed
during daytime hours by the wind field in the mixed layer and by the nocturnal
jet during nighttime. They have also found that the S02/sulfate conversion
rates are higher during daytime hours. In addition, they have observed that
plumes from elevated sources begin to mix downward early in the morning. Ap-
parently, to analyze these phenomena, one must take into consideration diffuse
sources from upwind urban areas.
In this paper, the use of a numerical model that is designed to simulate
transport, diffusion, chemical reactions, and surface deposition is discussed.
In the first section, the various components of the model are described. Next,
the application of this model is delineated. In the last section, potential
use of this model to the prolonged elevated pollution episodes (PEPE) is dis-
cussed.
DESCRIPTION OF THE MODEL
As stated earlier, the distribution of air pollutants to distances far
downwind of sources is a result of the interplay between emissions from major
point sources and the upwind background. Because of the differing character-
istics that govern the transport and diffusion processes for point and diffuse
sources, a hybrid modeling approach is proposed. This model consists of a
plume module for treating a limited number of point sources in their initial
phases, and a grid module for accommodating areal and diffuse sources. The
major components of this mode! are briefly described in the next section.
The Grid Module
One of the intrinsic difficulties of regional scale modeling of the inter-
action between emissions from a major point source and those from diffuse
sources emanating from upwind cities is the problem of spatial resolution.
Obviously, grid spacings that are large enough to cover the region of interest
are not sufficient for resolving the evolution of a point-source plume. Further-
more, according to classical diffusion theory, the spread of a pollutant cloud,
179
-------
as measured by the horizontal dispersion coefficient a , initially grows as
the first power of time and eventually approaches the square root of time.
The short time limit, a valid description of the dispersion of a point source,
is the basis for the Gaussian formulation; the large-time asymptotics, ade-
quate for describing areal or diffuse sources, is the basis for the gradient
transfer approach used in the grid model. For these reasons, separate approaches,
as shown in Figure 1, have been adopted for modeling the plume and upwind diffuse
sources.
The grid module, with the horizontal region of interest being divided
into an equal-spaced grid, has two full vertical layers with which to handle
advection, diffusion, and chemical reactions in the inversion layer and the
mixed layer, and a surface layer embedded in the mixed layer for treating verti-
cal diffusion and dry deposition.
The time-dependent multiple-species atmospheric diffusion equation in two
dimensions is used as the model equation for the inversion layer and the mixed
layer:
9C: n 9C| n C; 3 n C; \ J) / p OC: \ n g 0
— + u*_L-H vK_L= JL K* __L)+-?_[K* Lj^. D* • f(Dg) + Rf + s , (1)
9t 3x 9y
i = 1, 2, ...., N
I: Inversion Layer;
& =
II: Mixed Layer ,
o US,
where c. is the vertically averaged concentration of species i; u , v , and
o a
k , k are wind velocities and turbulent diffusivities in the x- and y-
S, 9, S. 2
directions, D is the two-dimensional divergence (= du /dx + dv /dy), R. and
$, 1
S. are chemical reaction and volumetric source terms in the £-th layer, res-
1 o
pectively. The value of f(D ) denotes the corresponding concentration from
influxes and outfluxes.
The major feature of this model equation is that the pollutant distribu-
tion is assumed to be uniform in the vertical direction for both layers. With
this assumption, the model equation can be shown to be formally derived from
180
-------
the more general atmospheric diffusion equation. The compelling reason for
using this assumption is that, for the scale of interest, the vertical diffusion
term has been found to be one or two orders of magnitude greater than the
transport term and the horizontal diffusion term (Durran et al., 1979).
Because of the relatively large spatial scale employed, however, the pseudo-
diffusion associated with the numerical solution of Eq. (1) can be overwhelming.
Consequently, an accurate scheme must be used for the simulation of the trans-
port term. On the basis of consideration of both predictive accuracy and com-
putational speed, the SHASTA method (Boris and Book, 1973) was chosen for use
in this study. This method consists of two computational stages: a transport
stage that is conservative and nonnegative, and an antidiffusion stage de-
signed to cancel, in part, the pseudo-diffusion generated in the transport stage.
A comprehensive testing of this and other numerical methods indicates that,
for the present application, the SHASTA method yields realistic results with
a reasonable computational burden (Durran et al., 1979).
The Plume Module
To resolve the spatial-resolution problem, a multiple-box reactive plume
model has been used to track the first phase of the plumes emanating from major
point sources. At a sufficiently downwind location, pollutants contained in
the plume module will be released into the appropriate grid in the grid module.
The multiple-box reactive plume model (Liu, Stewart, and Roth, 1978) assumes
that the width and height of cell j can be described by two parameters:
wj = wj (s) , (2)
hj = hj (s) , (3)
where wj denotes the width of cell j, fr1 is the vertical thickness of the
plume, and s represents the distance downwind of a point source measured along
the plume axis. On the basis of a mass balance, the governing equation along
a wind trajectory, s, can then be expressed in the following manner:
181
-------
dcj /dc|
dt \ dt /chem \ wj ds
i = 1,2,..., N, j= 1,2,..., M,
where c1? denotes the concentration of chemical species i in cell j, u is the
wind speed along the plume trajectory, and t is time since emission. Thus,
the model can be envisioned as a series of temporally varying, expanding puffs
released at fixed time intervals.
As envisioned in this model, the total mass of each primary pollutant is
divided equally among M cells. By allowing the volume of each cell to expand
as dictated by atmospheric dispersion, cell boundaries can be viewed as being
impermeable for unreactive pollutants. Thus, only the diffusion of reactive
and entrained pollutants must be parameterized. For example, within the frame-
work of a Gaussian formulation and for an elliptically shaped plume, the radii
(in transformed polar coordinates) of concentric rings containing equal pollut-
ant masses can be shown (Freiberg, 1976) to be represented by the following
equation:
(5)
where r. is the radius of the j-th concentric ring, and a and a represent
J J
the horizontal and vertical dispersion coefficients for the plume in transformed
polar coordinates given by r2 = y2/a2 + z2/a2 Plume distributions other than
Gaussian can also be incorporated into this model.
A unique feature of this model is the manner in which the plume disper-
sion is treated. For an inert species emitted from the source, the rates of
expansion of these cells are determined by horizontal dispersion such that
there is no net mass flux across cell boundaries. For a background or a
reactive pollutant, the Fickian diffusion approach is adopted: the flux
across the interface between cells j and j+1 is assumed to be equal to the
product of an equivalent diffusion coefficient, K., and the concentration
J
gradient of the background or reactive species relative to that of an inert
species. That is,
182
-------
where c and c denote the concentration distributions of the background or
reactive pollutant and the inert species, respectively. The equivalent dif
fusion coefficient can be determined by considering the case of a constant
background concentration, CB. Since, after each time step, CB should be
unchanged, a simple balance of mass fluxes for cell j yields
(7)
i-V
Assuming local symmetry at the centerline of the plume, one may impose
KQ = 0 . (8)
Then the equivalent diffusion coefficient K. can be solved by using the
recurrence formula, Eq. (7), and appropriate finite difference approximations
(Figure 2).
The Chemical Kinetics Module
The application of both the grid module and the plume module requires a
kinetic mechanism describing pertinent chemical reactions. This mechanism must
include a specification of the elementary kinetic steps and their associated
rate constants.
Extensive efforts have been made in the past to develop a valid kinetic mech-
anism for photochemical smog, focusing primarily on systems of hydrocarbons and
nitrogen oxides. One recent result of these studies is carbon-bond mechanism.
When the carbon-bond mechanism (CBM) was first formulated, it represented a
condensation of existing explicit mechanisms, mainly for propylene and butane.
The original carbon-bond mechanism has been recently updated (Whitten et al.,
1978) for the following reasons:
1. Changes in one reaction may require compensating changes
in other reactions to maintain the overall predictive
accuracy of simulations using the mechanism.
183
-------
2. Testing of the CBM on a large set of smog chamber data should
be carried out subsequent to any changes to ensure that no
special problems have been created.
3. Documentation of changes is necessary to keep all users of
the CBM informed.
The formulation of the new version of the CBM reflects the following changes
to the original CBM:
1. Inclusion of the reactions of the intermediate Criegee species
formed from the ozone-olefin reactions, and surrogate species for
the hydroxyperoxyalkyl products.
2. Inclusion of a new formulation for carbonyl photolysis and
oxidation.
3. Treatment of alkyl radicals in long-chain paraffins, and
internal olefins as carbonyls.
4. Explicit treatment of ethylene, and incorporation of a new
aromatic chemical reactions scheme.
Each of these changes to the CBM is discussed in detail by Whitten et al.
(1978). The full version of this updated carbon-bond mechanism is implemented
in the plume module, and a simplified version is incorporated in the grid
module.
The Surface Module
The importance of dry deposition of S02 and sulfate in long-range trans-
port problems necessitates treatment of this process in the model. For pollut-
ants originating from either elevated or distant ground-level sources, most
of the pollutant mass is contained in the mixing layer. The removal processes
consist of the diffusion of the pollutants through the surface layer to the
ground, followed by absorption or adsorption at the atmosphere-ground interface.
A unique feature of the surface layer is its diurnal variation in temperature,
which is a result of daytime heating and nighttime cooling of the surface.
This variation affects the vertical pollutant distribution through atmospheric
stabilities and, consequently, the rate of surface uptake of pollutants.
The surface layer can be viewed as being composed of two layers: the tur-
bulent layer and the viscous sublayer. With this simplification, the pollut-
ant flux to the ground surface due to dry deposition, F, can be obtained
through the following equation:
184
-------
,.F+7-1/a.F1/a_
(9)
where
ku^z
dz
(10)
and where k is the von Karman constant, u* is the frictional velocity, p is
the Stanton number, (J) is the dimensionless wind shear (Thorn, 1972), and ZQ and
h are the surface roughness and the height of the surface layer, respectively.
The parameters y and a represent the reaction rate and reaction order, respec-
tively, for the uptake of air pollutants by chemical reaction or catalytic
decomposition at the ground surface. For a = 1, 2, and 3, closed-form solutions
can be found for F:
F =
1
—
7
a= 1
1
— + 4lc
y I
/2
(11)
21
a = 2
a=3
where
= ±r^2+_
I [ 21 277|
11/2
1/2
(12)
These formulas reduce to that of Thorn (1972) for the special case of a first-
order surface reaction and a neutrally stratified atmosphere.
PRIOR APPLICATIONS OF THE MESOSCALE MODEL
The first exercise of the mesoscale model was its application to the
Northern Great Plains. An early version of this model, without the cloud
185
-------
layer and using only an inert plume module, was used to assess the regional
air quality impact of emissions from local coal development (Durran et al.,
1979). Figure 3 shows the computed S02 and sulfate concentrations. It is
interesting to note that the surface deposition predicted by this model does
exhibit strong diurnal patterns, shown in Figure 4, as was observed by Husar
et al. (1978).
The development of the reactive plume module has also been completed.
As shown in Figure 5, a preliminary test run (without the grid module) demon-
strated that this model appears to retrieve qualitatively many features of
the observed concentration profiles, namely, the Gaussian distribution for
an inert species and the depletion of ozone near the source.
A more recent application of one version of the mesoscale model was in
the simulation of the 1975 Northeast Oxidant Experiment (Lamb et al., 1977).
Although preliminary results (Figure 6) appear to be encouraging, problems
such as the prescription of the background concentrations upwind of the
modeling region remain to be resolved.
APPLICATION OF THE MESOSCALE MODEL TO PEPE
The Mesoscale Air Quality Model described earlier can be used in one or
more of the following three modes in the PEPE experiments:
1. Planning mode. In the initial phase the Mesoscale Model can
be used as a planning tool for designing and scheduling the field
measurement program. Meteorological scenarios leading to pro-
longed elevated pollution episodes in the area of interest would
first be identified using climatological data. The Mesoscale
Model can then be exercised to determine the temporal and spatial
distribution of high pollutant concentrations as a result of
transport of precursors, typically at the trailing edge of a high-
pressure system. These predictions should be useful for (a) the
siting of ground-level monitoring stations and the specification of
location and density of concentrations; and (b) the development
of airborne measurements protocols.
2. Analysis mode. The Mesoscale Model can also be used to analyze
and interpret data collected in the field measurement program.
For example, by comparing the HC/NO ratios, as computed by the
/\
model, with the corresponding measured values, one may be able
186
-------
to unravel the differences between smog chamber experiments and
secondary pollutants. One may also be able to examine the ade-
quacy of emissions inventories. Through a series of well-designed
sensitivity studies, one could also employ the mesoscale model
to examine the interaction of large point-source plumes with upwind
urban emissions under prolonged elevated pollution episodes as
well as to determine the role played by different meteorological
scenarios (pressure contours, wind trajectories, S02/sulfate con-
version rate, relative humidities, and so on) in the formation of
sulfate, ozone, and other secondary pollutants.
3. Prediction. After the Mesoscale Model has been subjected to a
careful and thorough verification exercise, it can be used to predict
prolonged elevated pollution episodes. It can be utilized in fore-
casting the worst conditions under prolonged elevated pollution
episodes and in formulating control strategies on a regional basis.
ACKNOWLEDGMENT
Support for portions of this work from the U.S. Environmental Protection
Agency is gratefully acknowledged. In particular, we acknowledge the guidance
and assistance provided by D. Henderson (currently with the National Park
Service), and K. Demerjian, M. Dodge, and R. Lamb of ESRL.
REFERENCES
Boris, J. P., and D. L. Book. 1973. Flux-Corrected Transport--!: SHASTA,
A Fluid Transport Algorithm That Works. J. Computational Phys., 11:38-69.
Dimitriades, B., and A. P. Altshuller. 1977. International Conference on
Oxidant Problems: Analysis of the Evidence/Viewpoints Presented, Part I:
Definition of Key Issues. J. Air Poll. Contr. Assoc.t 27, No. 4:299-
307.
Durran, D., et al. 1979. A Study of Long Range Air Pollution Problems Re-
lated to Coal Development in the Northern Great Plains. To be published
in Atmos. Environ.
Freiberg, J. 1976. The Iron Catalyzed Oxidation of S02 to Acid Sulfate in
Dispersion Plumes. Atmos. Environ., 10:121-130.
Hidy, G. M., et al. 1976. Design of the Sulfate Regional Experiment (SURE),
Vol. I: Supporting Data and Analysis. EC-125, Environmental Research
& Technology, Incorporated, Westlake Village, California.
187
-------
Husar, R. B., et al. 1978. Sulfur Budget of a Power Plant Plume. Atmos.
Environ.. 12:549-568.
Lamb, R. G., et al. 1977. Development and Evaluation of a Mesoscale Photo-
chemical Air Quality Simulation Model. EI77-69, Systems Applications,
Incorporated, San Rafael, California.
Liu, M. K., D. A. Stewart, and P. M. Roth. 1978. An Improved Version of
the Reactive Plume Model (RPM-II). 9th International Technical Meeting
on Air Pollution Modeling and Its Application, 28-31 August 1978,
Toronto, Canada.
Miller, D. F., et al. 1978. Ozone Formation Related to Power Plant Emissions.
Science. 202:1186-1188.
Thorn, A. S. 1972. Momentum, Mass and Heat Exchange of Vegetation. Quart.
J. Roy. Met. Soc.. 98:124-134.
Whitten, G. A., et al. 1978. Modeling of Simulated Photochemical Smog
with Kinetic Mechanisms. EF78-121, Systems Applications, Incorporated,
San Rafael, California.
Wilson, W. E. 1978. Sulfates in the Atmosphere: A Progress Report on
Project MISTT. Atmos. Environ., 12:537-547.
188
-------
PLUME MODEL FOR MAJOR POINT SOURCE
• TRANSPORT
• DISPERSION
• CHEMICAL REACTIONS
00
BOUNDARIES FOR
GRID MODEL
(UPWIND BACKGROUND CONCENTRATIONS
GRID MODEL FOR< URBAN AND DIFFUSE SOURCES
(DOWNWIND PORTION OF MAJOR
POINT SOURCE PLUMES
LAYER I: CLOUD OR INVERSION LAYER
• ADVECTION
• HORIZONTAL DIFFUSION
• CHEMICAL REACTIONS
LAYER II: MIXED LAYER
• ADVECTION
• HORIZONTAL DIFFUSION
• CHEMICAL REACTIONS
LAYER III: SURFACE LAYER
• VERTICAL DIFFUSION
• DRY DEPOSITION
Figure 1. Schematic for the SAI mesoscale air pollution model.
-------
TIME AFTER EMISSION
S 20
60 MINUTES
40 MINUTES
20 MINTUES
200 400 600 800
Distances from plume center!1ne (meters)
1000
Figure 2. Equivalent diffusion coefficients as a function of time
after emission and location.
190
-------
100 km
(a) S02; 1700-2000 MST 5 April 1976
100 km
(b) Sulfate; 1700-2000 MST 6 April 1976
Figure 3. Concentrations predicted by the mesoscale air pollution
model. Isopleths show concentrations at 2, 4, 8, and
16M9/m3.
191
-------
v» sii s* 7i all ei iii* itii x 10 km
(a) 1400-1700 MST.4 April 1976
«* S« 6* 7« M 9» !«• 119
I I 0-2
I I 2-4
f~~i 4-6
I I 6-8
> 8
1* 2* 3»
>•« lit c. 0 kn
(b) 200-500 MST 5 April 1976
Figure 4. SC>2 deposition velocities (in mm/s) calculated by the
mesoscale air pollution model.
192
-------
10
00
(a) Measured Profiles
NO CONCENTRATION (ppb)
- -150
1000 800 600 400 200 0 20° 400 600
DlsUnce froa plune centerllne (neters)
60 MINUTES
1000
BACKGROUND OZONE
x^ CONCENTRATION
! i '
1
1 !
1 1
1 i
ji
L- 1
L- 1_
i i i i i
OZONE CONCENTRATION (ppb)
•20
^-60 MINUTES
1 1 1
i 1
•is j !r-~"~ 40 NINUT1
j 1
4- 1
| ;
i i
" 10 | -^ 20 HJHUTES
r— '
5
1 1 1 1 1
1000 800 600 400 200 0 200 400 600
Distance from plune centerllne (neters)
800 1000
" 1978 by the American Association for the Advancement of Science.
This figure originally appeared in an article by D. F. Miller in
volume 202 of Science, pp. 1186-1188.
(b) RPM-II Predicted Profiles
Source: Liu et al. (1978).
Figure 5. Comparison of measured and predicted profiles for a reactive plume.
-------
too
200
300
too
500
600
700
800
100
200
300
400
500
600
700
BOO
Figure 6. Preliminary results from the simulation of the 1975
northeast oxidant experiment (surface ozone
concentrations in ppb).
194
-------
EPRI'S AIR QUALITY STUDIES
Glenn R. Hilst
Electric Power Research Institute
Palo Alto, California
INTRODUCTION
Air quality studies and research, including the forthcoming PEPE project,
which this Workshop addresses, have taken on new and larger degrees of impor-
tance as our society wrestles with the problems of restoring and preserving
equitable environmental quality while keeping the costs of environmental protec-
tion in balance with the benefits to be gained. The complexities of atmospheric
transport and dilution of airborne contaminants coupled with the chemical trans-
formation and depositional losses of pollutants along the way from source to
receptor have been the subject of scientific inquiry for decades. To be sure,
these research studies have provided a great deal of insight, in particular, into
the individual components of atmospheric processes that control pollution ex-
posure patterns once the source strengths and configurations have been estab-
lished.
Unfortunately, however, these bits and pieces of scientific wizardry are not
adequate for present and future requirements for rational and cost-effective
pollution control strategies for at least two reasons:
1. Either parts of the puzzle are very poorly understood or our
present predictive capabilities are severely handicapped by
erroneous and inadequate input information. Atmospheric trans-
port on the regional scale and chemical transformation mechanisms
and rates are prime examples. Another example is the depositional
loss of gases and particles, also of fundamental importance in
regional pollutant behavior.
2. The linkage of the transport-dilution/mixing chemistry-sources-
sinks modules of air pollution models is still in a primitive
state and represents a serious deficiency when we address such
basic requirements as source attribution for pollutants origi-
nating from multiple sources. This problem is partly scientific
195
-------
and partly technological, i.e., computer cost and size limita-
tions. However, it is now very clear that we are dealing with a
coupled, nonlinear system in which the consequences of feedback
from one module to another can have drastic impacts on the
accuracy of our predictions. Diffusive mixing limitations on
fast chemical reactions in inhomogeneously mixed turbulent flow
reactors, of which the atmosphere is one, are examples.
These limitations on our present ability to predict the transport and fate
of airborne pollutants, and the very practical and urgent need for improved
predictions, suggest two things to me:
1. We must accelerate research in this area significantly, but in a
thoughtful and coordinated way, in order to produce results in
a timely and efficient manner. Air pollution research, as we
atmospheric scientists know it, is emerging into an era of "big
science" with costs and management requirements to match. Large
field measurements programs are essential, and our appetites for
computer size and time are insatiable. If these expensive opera-
tions are to be efficient, we must learn from and synthesize all of
the results of related programs that have gone on before and are
going on around us now.
2. We must maintain a balance between the individual research studies
that can improve our understanding of the inner workings of the
atmospheric pollution system and the large, integrated measure-
ments and model construction systems. This may seem self-evident,
but it is my experience that accomplishing this objective is far
from easy. It is all too convenient to focus on one or the other
and lose sight of the fact that "small" science provides the tools
that "big" science uses, and the societal problems that "big"
science addresses are in fact the best rationale for a strong and
vigorous "small" science program. As we think in terms of mega-
bucks for our PEPE's, SURE's, STATE'S, MAPSS's, and the like, we
must also keep their life support system, the hundred kilobuck
research projects, vigorous and healthy.
196
-------
THE EPRI AIR QUALITY STUDIES
Faced, as it is, with the prospect of regulatory requirements for spending
hundreds of billions of its rate payers' dollars on air pollution control
equipment and operations during the decade ahead, to say nothing of highly
restricted power plant siting options, the U.S. electric power industry has
a very lively interest in establishing the factual base for the requirements
and the strategies to be adopted for air pollution control. As a result, the
research arm and center for the industry, EPRI, has developed a large (by
industrial standards) research subprogram within the Physical Factors Program
of its Environmental Assessment Department. This subprogram is called, aptly
enough, the Air Quality Studies. At present the Air Quality Studies are strongly
oriented toward pollutants associated with fossil fuel combustion, but the
general thrust of the program is directed to three key questions: (1) where
does the pollutant go, (2) in what concentration does it arrive, and (3) what is
its chemical form when it arrives. We are equally concerned with primary and
secondary pollutants, gases and particles, time scales from minutes to days,
and travel distances from kilometers to continental dimensions. Naturally, our
prime concern is with pollutants emanating from tall stacks, but the need to
separate out utility contributions to ambient pollution levels from those attrib-
utable to other sources blurs this distinction quickly.
The EPRI Air Quality Studies have also been responsive to various regulatory
forcing functions. Secondary formation of sulfates from S02, first as a poten-
tial health hazard and more recently as a factor in visibility impairment, gave
rise to our Sulfate Regional Experiment (SURE). Visibility impairment and pre-
vention of significant deterioration (PSD) as embodied in the 1977 Clean Air
Act Amendments have been forcing functions as well. However, these immediate
problems have not given rise to a "fire fighting" type of research. Rather,
timely but thoughtful and thorough understanding and fact finding have been
called for. Our typical research plan calls for a 3- to 5-year effort, or "time
to pay off" if you will.
The organization of EPRI's Air Quality Studies can best be described under
four major headings:
197
-------
Large-scale field measurements
Integrated analyses and model development
Supporting special studies
Instrumentation development
As a measure of the level of effort called for in these categories, our
1979 budget allocates approximately $5,000,000 to the large-scale field measure-
ments projects, $500,000 to integrated analyses and model development, $750,000
to supporting special studies, and $750,000 to instrumentation development.
This allocation reflects our deep concern for the need to establish the facts
concerning present baseline values of the magnitude and variability of air
pollutants across the country, particularly on a regional basis. Prior moni-
toring and field experimentation have concentrated on pollution "hot spots" for
obvious reasons; these areas have required immediate regulatory attention and
they have provided the best targets for research studies. But do these measure-
ments represent anything but the local situation? Through the SURE project,
now nearing completion, and its sequel, the Regional Air Quality Studies (RAQS),
we have set out to document the answer to that critical question. These activ-
ities will continue to dominate our program, from a budgetary point of view,
for some time to come because they come closest to establishing the factual
basis for regulatory actions and they provide the basic phenomenological record
that scientific research must explain. However, we at EPRI see the "bottom line"
of all of this as models that incorporate the scientific know-how and predict,
within the irreducible uncertainty that nature herself provides, the environ-
mental consequences of man's activities, including electric power generation.
The project content of EPRI's Air Quality Studies and the possible inter-
actions with the EPA PEPE project can best be described under the major program
categories listed above.
Large-Scale Field Measurements
As noted before, EPRI's first venture into large-scale field measurements
projects is the Sulfate Regional Experiment (SURE), which has been underway since
1976 and will be completed in 1980. The details of this project and some of the
early results will be discussed by Dr. George Hidy of Environmental Research and
Technology, Inc., later in this Workshop. To say the least, the SURE is highly
relevant to the PEPE study because we have measured pollution episodes over the
198
-------
eastern United States, as well as the much more frequent low-level pollution
behavior characterizing the western United States. These data and analyses
should provide a gold mine of information for fine tuning the PEPE's field
measurements program, and I hope they will be used that way.
The SURE project, as the name implies, emphasized the S02/S04 problem, but
it also gathered extensive information on the regional behavior of NO/NO , 03,
TSP, RSP, and various elements contained in airborne particulates. During the
latter phases of the project we have also incorporated precipitation chemistry
measurements. During 1979 we propose to initiate a two-station visibility study
in this area. Although the basic data gathering for SURE has been completed, we
are extending the lifetime of an augmented Class I station network into 1980.
This network will consist of twelve surface aerometric stations and fourteen
precipitation chemistry sites, as well as the two visibility field sites. This
sequel to the SURE has been incorporated into our general RAQS as the eastern
network (ERAQS). Depending upon the timing of PEPE and budgetary considerations,
there is, of course, a large potential for coordinated data gathering between
ERAQS and PEPE.
RAQS is also being extended to the western United States (WRAQS), but with
somewhat different emphases. The regional air quality problems of the western
United States clearly focus on the formation and transport of fine particles
and their impact on visibility impairment. These are the problems we are address-
ing in WRAQS. During 1979 we propose to set up a visibility/fine particle
network complementary to the EPA/NPS VIEW network. In 1980 and 1981 this net-
work will be extended throughout the intermountain and high plains region. Again
our intent is to gather the data necessary to define the regional behavior of
air pollutants and develop regional models.
On another front and time and distance scale, during 1979 EPRI plans to ini-
tiate its Plume Model Validation project (PMV). This project has been described
in a recent publication (EPRI EA-917-SY, "Plume Model Validation," October 1978).
In brief, very extensive sets of measurements of plume behavior in various
terrain and meteorological conditions are to be made in order to validate the
modular performance of plume models as well as their general operational utility.
These results are expected to pinpoint improvements in plume model performance
and reliability; again, an analysis and model development component is also
planned for this project.
199
-------
The relevance of the PMV project to PEPE is not so obvious, but better
understanding of the dilution and chemical behavior of plumes during their
first few hours may shed some light on the regional episodic behavior of pollut-
ants as well as their potential for long-distance transport.
Integrated Analyses and Model Development
Our primary effort in this area during 1979 will be the completion of the
SURE analyses and modeling. However, as noted above, ERAQS, WRAQS, precipitation
chemistry, and visibility are also headed toward extensive analyses and modeling
efforts in our Air Quality Studies. For planning purposes, we anticipate that ap-
proximately one-third of our future efforts will be directed toward this area.
However, we have also noted the need for closer liaison and coordination with simi-
lar efforts in other agencies and we welcome suggestions along these lines.
Supporting Special Studies
This category includes a veritable pot pourri of projects with limited but
important objectives. For example, we have funded a number of plume studies
designed to measure chemical conversion rates. We are funding two chamber
studies, one for dark reactions of S02 as a function of temperature and humidity,
including liquid water, and another to determine nitrogen chemistry in plumes
under photochemical conditions.
In another area we are supporting a very fundamental study of turbulent
transport as it affects dry deposition of gases and particles under various
terrain and vegetation conditions. We have also pioneered in developing measure-
ment techniques and assaying biogenic emissions of sulfur compounds in the
eastern United States. Similarly, we have conducted a feasibility study for
comparisons of various instrumental methods used to measure visibility in the
southwest United States.
Instrumentation Development
Although our extensive field measurements projects have provided a natural
forcing function for improved aerometric measurements techniques, EPRI has con-
centrated on the development of a new generation of remote sensing devices. At
present we are funding the final design and construction of a differential
absorption lidar system (DIAL) that shows great promise. This system is designed
200
-------
to measure concentrations of S02, N02, or 03 out to distances of 3 to 4 km and
with an axial resolution of 5 to 10 m. When field-proven, this instrument should
revolutionize the study of close-in plume behavior.
On another front, we are supporting the final design and construction of
an automated telephotometer that will systematize measurements of visual range.
When completed, we propose to use this instrument extensively in our western
visibility studies.
Finally, we are also supporting further development of remote sensing sys-
tems designed to resolve the three-dimensional wind and temperature fields through
the depth of the planetary boundary layer. Such instrumentation is essential
to plume behavior studies in complex terrain.
These descriptions of EPRI's Air Quality Studies are necessarily brief.
Also, this program is evolving rapidly in its young life. However, I think it
is evident that the electric power industry has made a substantial commitment
to research and development in the area of air quality.
CONCLUSION
Beginning with the SURE project and its coordination with the DOE MAP3S and
the EPA MISTT and STATE programs, as well as informal coordination with other
work, the desirability of coordinated programs has been well demonstrated. It
is our hope that this is only the beginning. To be sure, each of us will have
our own peculiar goals, objectives, and timetables that must be met. Within
this framework, however, there is ample room for making certain that the total
results of our efforts are greater than the sum of the parts provided by each of
us. We welcome the opportunity to coordinate our research with PEPE and other
air quality studies.
201
-------
ENSEMBLE TRAJECTORY ANALYSIS OF SUMMERTIME SULFATE
CONCENTRATIONS IN THE EASTERN UNITED STATES
Perry J. Samson
New York State Department of Environmental
Conservation, Albany, New York
INTRODUCTION
Much of our present knowledge about the formation, movement, and eventual
decay of regional scale episodes of high summertime sulfate concentrations
has been obtained from spatial and temporal examination of individual episode
periods. Trajectory analysis is a useful tool for tracing the movement of
such systems either forward or backward in time. This technique has been
employed by several authors including Hall et al. (1973), Husar et al.
(1976), Samson and Ragland (1977), and Wolff et al. (1977) to demonstrate
the advection of large areas of low visibility, high ozone concentrations,
and/or high sulfate concentrations in the eastern United States.
There are, however, several potential problems inherent in constructing
individual trajectories that add to the uncertainty of analysis of small
data sets. Observational errors in the input winds can cause considerable
error in trajectory location. If there exist significant large-scale verti-
cal motions, then the sampled air may have its origin in a layer of the
atmosphere different from the layer being traced. The net motion over a
predetermined layer may not adequately describe the motion of the air column
being studied if large vertical shears exist over the layer.
If we assume that these effects are random along the path of the trajec-
tory then it would be useful to use ensembles of trajectories over a large
number of days to smooth the random "noise". It is then possible to obtain
general relationships between resultant sulfate concentrations and the
history of the sampled air.
This paper describes the use of ensemble trajectory analysis for relat-
ing resultant sulfate concentrations to the upwind emissions burden, wind
202
-------
speed, and mixing height. The general technique is presented and prelimi-
nary results relating sulfate concentrations to upwind wind speed at several
locations in the eastern United States are presented.
APPROACH
The trajectory model of Heffter et al. (1975) as modified by Heffter
and Ferber (1977) has been used in this study. The model computes the
trajectory of an air column within the mixed layer of the atmosphere. All
motion is assumed horizontal and winds are integrated from the top of the
surface inversion to the top of the mixed layer as determined from wind and
temperature profiles. The endpoints of the trajectory segments of 3 hours'
duration are stored in the computer.
This output can be used with a graphics plotting program labeled
TRAJPLOTTER which, for each day of interest, draws a base map (designated by
the user) and overlays the trajectories for that day. But the ensemble of
trajectory endpoints can also be used with corresponding pollutant concen-
tration to arrive at estimates of interrelationships.
For a well-mixed boundary layer, the conservation of mass equation can
be solved for equilibrium conditions (see Lettau, 1970) to be of the form
C = -- - F(x;Ri) (1)
0 h
where C is the concentration (g/m3), Q is the areal source strength per
length of integration (g/m-sec), U is the wind speed average over the layer
of interest (m/sec), and h is the mixing height. The function F describes
the horizontal dispersion as a function of downwind distance from the source,
x, and stability (Richardson Number, Ri).
For the purpose of this investigation, assume that the net effect of
the dispersion function is small over a long time series (i.e., the horizon-
tal gradient of sulfate concentrations is small). Then the resultant concen-
tration of sulfate can be related to the upwind emissions burden, wind
speed, and mixing height using the formula
N (a.Q.)
C = C + I - !-! - (2)
° 1 (P
203
-------
where C is the natural background concentration (assumed less than 1 ug/m3
and ignored in the present study) and the subscripts refer to each of the N
trajectory segments included in the analysis. The parameters a., p. and Y,-
represent the weight of that particular variable at that particular trajec-
tory segment on the resultant concentration. These values can be determined
from least squares analysis of the sulfate and trajectory data. Figure 1
shows a hypothetical trajectory calculation backward in time from a point A.
The wind speed estimated for segment II upwind of point A would be the
distance traversed between points B and C divided by the time increment for
the segments. The emissions burden calculated for segment II is determined
by an objective analysis scheme from the midpoint between points B and C.
The area included in estimating the areal source strength increases with
distance from the origin (because of the increased scatter of endpoints
intuitively expected with increasing time from the origin). Further, the
weighting of sources is elongated along the axis of the segment, the ratio
of major to minor axes increasing with increasing wind speed.
The height of the mixed layer can be ascertained from analysis of
vertical temperature and wind profiles. In their model, Heffter and Ferber
(1977) use a technique whereby the observed vertical temperature profiles
are examined for inversion levels and the wind profiles are examined for
significant wind shear. This technique is to be tested against other, less
stringent criteria such as those proposed by Meyers et al. (1976). Once a
reasonable approach is determined, then the results will be included in the
total analysis.
RESULTS
The relationships between resultant sulfate concentrations and upwind
wind speed have been completed for several monitoring stations in the east-
ern United States. Figure 2 shows the location of the sites used thus far.
Table 1 lists the sites along with the dates of operation and the investi-
gators responsible.
In an earlier paper (Samson, 1978) it was shown, using the 1975 and
1976 sulfate data for the three sites in upstate New York, that no correla-
204
-------
tion existed between the measured sulfate concentration and the wind speed
measured through the mixed layer at the same time. However, using estimated
wind speeds from trajectory segments greater than 24 hr upstream produced a
significant correlation. That is, the resultant concentration of sulfate
measured at these rural locations could be explained, in part, statistically
by stagnation occurring more than 24 hr upwind.
The analysis of an additional year of summertime sulfate values at
those three stations and analysis of other sites in the eastern United
States has verified this finding. Figure 3 shows the analysis between the
inverse of wind speed and the resultant sulfate and nitrate concentrations
as a function of hours upwind. The sulfate produces the same pattern as was
reported previously for the upstate New York stations. For this short
period the sulfate could be explained, in part, by stagnation occurring more
than 48 hr upwind. The nitrates, on the other hand, have a high correlation
nearer the point of measurement. Thus, more local stagnation appears to
have affected nitrate concentrations rather than stagnation occurring well
upwind.
Figure 4 shows these relationships for four other locations. In three
of the four, the previously discussed relationship is true. Stagnation
upwind of Keysville, VA, McKee, KY, and Allegheny Mountain, PA, was signifi-
cantly correlated with resultant sulfate concentrations. The data from
Brookhaven did not produce the pattern. There are a few possible explanations
for this apparent discrepancy, the most likely being that due to Brookhaven's
location near the sea many of its trajectories that tracked out over the
ocean would have been ignored. As written, the trajectory model will termi-
nate trajectory analysis if insufficient upper air data are available within
a certain radius. Thus, the trajectories available for analysis for Brook-
haven would be biased toward those of continental origin. Other possible
explanations are the nearness of the New York metropolitan area, which could
be masking upwind effects, and the importance of mesoscale wind features
such as the sea breeze to pollutant concentrations.
Finally, the categorization of trajectories into those corresponding to
low resultant sulfate concentrations (<5 ug/m3) and high resultant sulfate
concentrations (>15 ug/m3) gives a graphic description of the general history
205
-------
of the sampled air. Figures 5, 6, 7, and 8 show the trajectories arriving
in the middle of the sampling period for Holland, NY, Allegheny Mountain,
PA, Keysville, VA, and McKee, KY, respectively. The duration of travel
between arrowheads is 6 hr, each trajectory being computed backwards for 72
hr (or until insufficient wind data were available to continue).
The dichotomy between the high and low samples is striking at these
stations. In Figure 5, Holland receives its lower concentrations from the
west to north with relatively high winds in the boundary layer. The highest
concentrations occur with winds from the south to west with several cases of
upwind stagnation observable. The classification of Allegheny Mountain
trajectories in Figure 6 shows a similar dichotomy.
The trajectories for Keysville, VA, show a markedly different pattern.
The lowest concentrations occur with airflow out of the south or with fast
moving air from the northwest. The higher concentrations occur with slower
moving air from the north and west and stagnant conditions to the southwest.
One of the most striking dichotomies occurs in the classification of
trajectories at McKee, KY, shown in Figure 8. Over the period of this
study, the lowest concentrations were associated with airflow from the north
to northwest moving at a moderately fast speed. The highest concentrations
occur with a consistent pattern looping anticyclonically from the east.
Since the trajectories to a point in Pennsylvania are similar to those shown
in Figure 6a for low concentrations at Allegheny Mountain, PA, it would
appear that production and/or accumulation south of that point must be
responsible for the high concentrations.
DISCUSSION
o
This is a preliminary report on research intended to investigate the
sensitivity of measured sulfate concentrations to upwind emission burden,
wind speed, and mixing height. A consistent pattern has emerged between
sulfate concentrations and upwind wind speed with a significant correlation
with the inverse of wind speed occurring more than 24 hr upwind of the sam-
pling point. This pattern could be due to:
1. The lag time between accumulation of sulfur dioxide and its trans-
formation to high concentrations of sulfate, or
206
-------
2. The relative location of these rural samplers to large source
regions (i.e., stagnation near the sampler will not produce high
concentrations in the absence of anthropogenic sources).
It is hoped that future work, including emissions burdens, will clarify
the reason for this pattern and give an indication of how sensitive sulfate
concentrations are to upwind emissions.
ACKNOWLEDGMENTS
The author would like to thank the investigators who offered their data
for this analysis: Dr. Peter Coffey and Mr. Phil Galvin, New York State
Department of Environmental Conservation; Dr. Paul Lioy, New York University;
Dr. George Wolff, General Motors Research Lab; Dr. Roger Tanner, Brookhaven
National Lab; and William Pierson, Ford Motor Company.
REFERENCES
Coffey, P. E., P. J. Galvin, J. Cline, and P. J. Samson. 1979. Results of
the 1977 New York State Sulfate Sampling Network. New York State
Department of Environmental Conservation Report (in preparation).
Galvin, P. J., P. J. Samson, P. E. Coffey, and D. Romano. 1978. Transport
of Sulfates to New York State. Environ. Sci. Tech, 12:580-584.
Hall, F. P., C. E. Duchon, L. G. Lee, and R. R. Hagan. 1973. Long-Range
Transport of Air Pollution: A Case Study, August 1970. Mon. Wea. Rev.,
101:404-411.
Heffter, J. L., A. D. Tayler, and G. J. Ferber. 1975. A Regional-Continental
Scale Transport, Diffusion and Deposition Model. NOAA Tech. Mem.
ERL-ARL-50, Air Resources Lab., Silver Springs, MD.
Heffter, J. L and G. J. Ferber. 1977. Development and Verification of the
ARL Regional-Continental Transport and Dispersion Model. Proc. Joint
Conf. on Appl. of Air Poll. Meteor. 29 November-2 December 1977, Amer.
Meteor. Soc., Boston, Mass., pp. 400-403.
Husar, R. B., N. V. Gillani, J. D. Husar, C. C. Paley, and P. N. Turcu.
1976. Long-Range Transport of Pollutants Observed Through Visibility
Contour Maps, Weather Maps and Trajectory Analysis. Proc. 3rd Symp.
Atmospheric Turbulence, Diffusion, and Air Quality. 19-22 October 1976.
Lettau, H. H. 1970. Physical and Meteorological Basis for Mathematical
Models of Urban Diffusion Processes. In Proc. of Symp. on Multiple-
Source Urban Diffusion Models (edited by A. C. Stern). U.S. Environ-
mental Protection Agency, AP-86, Research Triangle Park, NC, pp 2-1 to
2-26.
207
-------
Lioy, P. J., G. T. Wolff, J. S. Czachor, P. E. Coffey, W. N. Stasiuk, and D.
Romano. 1977. Evidence of High Atmospheric Concentrations of Sulfates
Detected at Rural Sites in the Northeast. J. Environ. Sci. Hlth.. A12:
1-14.
Lioy, P. J., G. T. Wolff, K. A. Rahn, D. M. Bernstein, and M. Kleinman.
1979. Characterization of Aerosols Upwind of New York City, II:
Aerosol Composition. Annals of the New York Academy of Sciences (in
press).
Meyers, R. E., R. T. Cederwall, W. D. Ohmstede, and W. aufm Kampe. 1976.
Transport and Diffusion Using a Diagnostic Mesoscale Model Employing
Mass and Total Energy Conservation Constraints. Proc. 3rd Symp. Atmos-
pheric Turbulence, Diffusion, and Air Quality. 19-22 October 1976.
Pierson, W. R., W. W. Brachaczek, T. J. Truex, J. W. Butler, and T. J.
Korniski. 1978. Ambient Sulfate Measurements on Allegheny Mountain
and the Question of Atmospheric Sulfate in the Northeastern United
States. Research Report, Scientific Laboratory, Ford Motor Company,
Dearborn, MI.
Samson, P. J. 1978. Ensemble Trajectory Analysis of Summertime Sulfate
Concentrations in New York State. Atmospheric Environment, 12:1889-1893.
Samson, P. J., and K. W. Ragland. 1977. Ozone and Visibility Reduction in
the Midwest: Evidence for Large-Scale Transport. J. Appl. Meteor.,
16:1101-1106.
Tanner, R. 1979 Sulfate, Nitrate and Related Ionic Species in Airborne
Particles. New York Acad. of Science Conf. on Aerosol; Anthropogenic
and Natural Sources and Transport, 9-12 January 1979.
Wolff, G. T., P. J. Lioy, G. D. Wight, R. E. Meyers, and R. T. Cederwall.
1977. An Investigation of Long-Range Transport of Ozone Across the
Midwestern and Northeastern United States. Atmospheric Environment,
11:797-805.
Wolff, G. T., P. R. Monson, and M. A. Ferman. 1978. On the Nature of the
Diurnal Variation of Sulfates at Rural Sites in the Eastern United
States. Presented at American Chemical Society Meeting, 11-14 September
1978, Miami, Florida, (also General Motors Research Rpt. GMR-2772, ENV
#46, General Motors Research Laboratries, Warren, MI, July 1978).
208
-------
Figure 1. Hypothetical trajectory
calculation.
Figure 2. Location of stations used in this report. A listing of the station
names along with dates of operation and investigators responsible
appears in Table 1.
209
-------
0.6
0.2
0.0
-0.2
-0.4
[S04]vs /v/
12 24 36 48 60
High Point, N J
August 01-15,1977
-0.4
Figure 3. Correlation between the inverse of wind speed and resultant sulfate
and nitrate concentrations as a function of hours upwind.
0.6-
0.4
0.2
0.0
-0.2-
-0.4
> 24 36 48 60
Keysville, Virginia
June 19-July 21,1976
0.4
0.2
o.o
-0.4J
jsojvs/v/'1
/12 24 36 48 60
Allegeny Mtn., Pa.
July 26-Aug. 11,1977
0.6
0.2
I °
-0.4
[S04]vs/V/"'
12 24 36 48 60
MeKee, Kentucky
August O3 - Sept. 13,1976
o.e
0.4
0.2
-0.4J
[sojvs/v/'1
Brookhaven
National Laboratory
July 18-Aug. 30,1977
Figure 4. The correlation between the inverse of wind speed and resultant sulfate
concentrations as a function of hours upwind.
210
-------
INS
Figure 5. Trajectories arriving at Holland, NY, at
1400 EOT on days of low and high
sulfate concentrations.
Figure 6. Trajectories arriving at Allegheny Mountain,
PA, at midpoint of sampling period (12 hr)
for low and high sulfate concentrations.
-------
ro
i—'
ro
Figure 7. Trajectories arriving at Keysville, VA, at
midpoint of sampling period (4 hr) for low
and high sulfate concentrations.
Figure 8. Trajectories arriving at McKee, KY, at
midpoint of sampling period (4 hr) for
low and high sulfate concentrations.
-------
TABLE 1. SOURCES OF SULFATE DATA AND PERIODS OF STUDY
Site
WFC
GLB
HOL
HPT
BNL
KEY
MCK
ALG
Location
Whiteface Moun-
tain, NY
Gil boa, NY
(Schoharie, NY)
Holland, NY
High Point, NJ
Brookhaven
National Labora-
tories, NY
Keysville, VA
McKee, KY
Allegheny Moun-
tain, PA
Investigators
Lioy et al. (1977)
Galvin et al. (1978)
Coffey et al. (1979)
Lioy et al. (1979)
Tanner (1979)
Wolff et al. (1978)
Wolff et al. (1978)
Pierson et al. (1978)
Period
1-31 July
1975
6-31 July
1976
15 June -
15 Aug.
1977
1-15 Aug.
1977
(6 hr)
18 July -
30 Aug.
1977
(6 hr)
19 June -
21 July
1976 (4 hr)
3 Aug. -
13 Sept.
1976 (4 hr)
26 July -
11 Aug.
1977 (12 hr)
213
-------
SOME RESULTS OF THE SULFATE REGIONAL EXPERIMENT
(SURE) AND THE PEPE EXPERIMENT DESIGN
G. M. Hidy,* P. K. Mueller, and T. F. Lavery
Environmental Research and Technology, Inc.
Westlake Village, California
INTRODUCTION
Because of the emerging results of studies of long-range pollutant
transport in Europe (e.g., OECD 1977), researchers have begun to suspect that
the occurrence of regional scale phenomena is also important in North America.
Early evidence suggested that regional pollution over scales exceeding 1,000 km
occurred in the northeastern United States. The occurrence of transport of
pollution over such distances, as compared with multiple source interactions
over distances of hundreds of kilometers, could not be distinguished. Inves-
tigation of such processes in the Northeast led to the design and implementa-
tion of the Sulfate Regional Experiment (SURE) (Hidy et al., 1976), as well
as other studies including the Multistate Atmospheric Power Production Pollu-
tion Study (MAP3S), and the Ohio River Basin Study (ORBS). The planned
investigation of Prolonged Elevated Pollution Episodes (PEPE's) represents a
second generation of regional projects in the Northeast. Operationally, we
define a PEPE as an occurrence of pollution for more than one day over a
geographical extent of several hundred kilometers. A PEPE is signalled by
the occurrence of high concentrations of particulate sulfate and ozone at
ground level and aloft. These events are believed to be associated with
mesoscale and synoptic-scale meteorology, and the interactions of sulfur or
nitrogen oxide emissions from tall stacks and from urban plumes.
To assist in the design of the PEPE experiment, consideration should be
given to the experience of programs nearing completion. This paper is intended
to provide a perspective on the occurrence of PEPEs as identified in the SURE
results.
THE SURE AND ITS RESULTS
The SURE is a major study that involved the operation of a ground network
*Speaker.
214
-------
of 54 monitoring stations in the region shown in Figure 1. The stations
are located by number as indicated. The first nine stations (Class I stations)
were heavily instrumented for sampling pollutant gases (S02, NO, N02, and 03)
and airborne particulate matter and were run continuously from August 1977
through October 1978. The additional 45 stations monitored only S02 and sus-
pended particulates in six center-season months: August 1977, October 1977,
January-February 1978, April 1978, July 1978, and October 1978. The SURE
ground measurements were supplemented by sampling aloft using two aircraft
during a 12-day period in each of the center-season months. The aircraft were
instrumented in a manner equivalent to the Class I stations; in addition, they
included an integrating nephelometer, a condensation nuclei counter, and
turbulence instrumentation. The experiment also involved extensive collection
and use of emission data for S02, NO , particulate, and hydrocarbons, as well
as National Weather Service meteorological observations. The experimental
concept of the SURE and its relationship to MAP3S and other programs is
discussed in detail by Mueller et al. (1978).
The SURE field program has been completed, and the data are currently
being analyzed. The results to date have revealed certain important features
of regional scale pollution patterns in the northeastern United States. Some
of the results relevant to the PEPE studies are summarized below.
The occurrence of elevated sulfate concentrations is characterized from
data taken at the nine continuously operating stations (Class I) over the 1-
year period between August 1977 and July 1978. The cumulative frequency of
occurrence of different sulfate levels is shown in Figure 2. The differences
among these rural stations are relatively small throughout the region. The
annual median level of sulfate is about 6 ug/m3 for all nine stations.
Concentrations greater than 20 ug/m3 occurred less than 6 percent of the time,
and concentrations of 10 |jg/m3 less than 25 percent of the time. Using the
20-ug/m3 level as a criterion for a PEPE, Figure 2 shows that PEPE's occur
only a small fraction of the time.
Looking at the regional events in a different way, we estimated the sul-
fate concentration in each of the 690 80 km x 80 km grid squares in the SURE
region by extrapolating the 54 station measurements (Mueller et al., 1979).
The extent of coverage of the region by different sulfate levels of XLO ug/m3,
>15 u g/m3 , and >20 ug/m3 is shown in Figure 3. These estimates cover the
215
-------
summer months of August 1977 and July 1978. For the purpose of discussion, we
have defined a regional event in terms of grid squares and land area covered:
40-50 percent in the >10-ug/m3 isopleth, accompanied by 25 percent in the >15-
(jg/m3 isopleth, and 10 percent in the >20-ug/m3 isopleth. We see that two
events of a regional nature took place in each of the summer months and each
lasted for at least 3 days. From these two samples, a PEPE is likely to occur
not more than twice during the duration of a planned summer experiment of
6 weeks.
The geographical extent of elevated pollutant levels may be characterized
by the number of days of >10 ug/m3 sulfate or by >20 pg/m3 calculated from
extrapolation of the network data. (See the isopleth maps for August 1977
in Figure 4.) The estimated heaviest regional concentration of sulfate in
summer is focused on the middle to upper Ohio Valley and Pennsylvania. On
this basis, and taking into account the heavy concentration of moderate to
large-si zed cities and the intense industrialization in this region, the PEPE
experiments should be conducted here.
It is interesting to compare the distributions of monthly geometric mean
sulfate and nitrate for August 1977, as given in Figure 5. Here, the sulfate
concentrations are estimated to be regionally widespread, but nitrate is more
localized around urban or isolated source complexes. Furthermore, the nitrate
levels in August 1977 at the ground are nearly an order of magnitude lower
than sulfate concentrations, despite a much smaller difference in regional S02
and NO precursor emissions. Such observations underscore the differences in
A
the behavior of the two anion species, and emphasize the need for future work
on nitrogen oxide behavior on an urban and regional scale.
The SURE aircraft sampling provided a useful indicator that, indeed, the
occurrence of sulfate involves significant activity aloft. Observations were
taken over the ground station at Duncan Falls, OH, during all four seasons.
Duncan Falls is roughly in the center of the zone of high regional pollution.
In Table 1 the average values of sulfate and nitrate are listed for the six
flights in each month for Duncan Falls. Data reported for the ground, for 500
to 5,000 ft, and for 5,000 to 10,000 ft suggest that sulfate concentration is
higher in the 500- to 5,000-ft layer than at the ground for all sampling
months except January-February 1978. Furthermore, the July 1978 observations
216
-------
and others given in the table suggest that (total) nitrate is considerably
higher in concentration aloft than at the ground (see also Hidy et al., 1979).
Thus, it will be crucial to the PEPE program to provide as much sampling aloft
as practicable within budgetary limitations.
From the early analysis of the SURE data, a distinction was made in the
character of PEPE's between the buildup of pollution associated with regional
scale air mass stagnation and that associated with the long-range transport
situation. Lavery et al. (1979) have identified long-range transport with the
regional scale "ducting or channeling" of air flow between the west side of
the Appalachian Mountains and east-west oriented summer cold fronts. The
differences in these events are summarized in Table 2. Here, the transport
winds are identified as those at 300 m. Zones of influence are defined qual-
itatively in terms of the extent to which elevated sulfate concentrations are
identifiable with a fixed source complex (an 80 x 80 km square in the SURE
grid).
The structure of the ducting situation is illustrated for August 3
and 5, 1977, in Figure 6. The hatched squares denote the zones of high S02
emission density. The two graphs show the isopleths of sulfate estimated from
24-hr data of the ground network. In this case, the isopleths are elongated
with the 300-m winds. Also shown are the isobars in millibars. Moist, warm
air flows at moderate speed northeastward with the anticyclonic flow, accumu-
lating pollution behind the large high-pressure area over the high emission
density zones of the midwest, and passing along the western part of the Appa-
lachians. Unlike the stagnation situations, the zone of maximum sulfate
concentration "detaches" from the areas of high-emission density and moves
"downwind" into areas to the east with lower emission density. The apparent
movement of the zone of high pollution passed eastward with the winds over a
3-day period before precipitation and other processes associated with the
southward drifting of the front dissipated that event. To date, such situa-
tions have been identified only three or four times in the SURE set, and only
in summer when the cold fronts are more east-west oriented than in winter.
Thus, it is expected that such conditions for a regional event would be acces-
sible on at least a 24-hr forecast basis over the proposed PEPE study area.
Examination of the SURE results and results from the National Aerometric Data
217
-------
Bank, with meteorological data, should provide a basis for such forecasts,
perhaps extending well beyond 24 hr.
RECOMMENDATIONS
The SURE served a very useful purpose by providing a comprehensive set of
basic measurements of air quality at rural locations in the eastern States.
As a second generation study, EPA's proposed PEPE program offers an opportun-
ity to investigate several aspects of regional pollution events in detail
using more sophisticated techniques. Some of the areas that need additional
investigation are suggested in the following list.
Rural hydrocarbon vapors, including oxygenates, in the PEPE region
need quantification both at the ground and aloft for improvement of
regional scale photochemical models.
Some of the chemical links require characterization, including the
rate of production of OH radicals, H202, PAN, and HN03. The vertical
and horizontal distribution of these constituents is of interest in
the PEPE. ' -
Knowledge of the chemical processes taking place in clouds is needed
to clarify certain aspects of the S02 and N0v oxidation processes.
^
The linkage to cloud chemistry requires elucidation between chemical
reactions and wet and dry removal processes in regional pollution
events.
These four factors combined with other meteorological activity (e.g.,
transport and mixing), undoubtedly can explain the "sudden" genesis and death
of PEPE1s. The details of the dynamics of these events are unclear after
initial analysis and interpretation of the SURE data. We recommend that the
PEPE studies attempt to focus on these four areas in extending knowledge of
regional scale pollution processes.
ACKNOWLEDGMENT
This work was sponsored in part by the Electric Power Research Institute
under Project 862-2.
REFERENCES
Hidy, G. M. et al. 1976. The Design for the Sulfate Regional Experiment.
Report EC-125, Electric Power Research Institute, Palo Alto, CA.
218
-------
Hidy, G. M., et al. 1979. Air Chemistry in the Rural Northeastern United
States. To be submitted to Environ. Sci. & Technol.
Lavery, T. F., et al. 1979. Occurrence of Long Range Transport of Sulfur
Oxides in the Northeastern United States. In Proc. 4th Symposium on
Turbulence, Diffusion and Air Pollution, American Meteorological Society,
Boston, p. 330.
Mueller, P. K., et al. 1978. Implementation and Coordination of the Sulfate
Regional Experiment (SURE) and Related Research Programs. Electric
Power Research Institute Report, Palo Alto, CA, in press.
Organization for Economic Cooperation & Development (OECD) 1977. The OECD
Programme on The Long Range Transport of Air Pollutants Measurements and
Findings. OECD, Paris, France.
219
-------
90-
70-
IVS
ro
o
IS 16 17 18 19 2O 21 22 23 24
1234
9O'
Figure 1. SURE region and deployment of Class I (triangles) and II (circles)
Quality Monitoring Stations.
2} 26 27 26 \29
EPRI/SURE Air
-------
ro
IVJ
ra
V)
MONTAGUE
SCRANTON
INDIAN RIVER
O DUNCAN FALLS
ROCKPORT
GILES CITY
FT. WAYNE
RESEARCH TRIANGLE PARK
LEWISBURG (GREENBRIER AIRPORT)
0,01
10 30 50 80 90 95
% OF TIME LESS THAN VALUE
99 99,8
99,9
Figure 2. Frequency of occurrence of 24-h average sulfate concentrations in the SURE region
based on August 1977 to July 1978 results.
-------
300
AUGUST 1-31, 1977
DAYS
Figure 3. Extrapolated grid coverage of 24-h average sulfate concentra-
tions > 10 fig/m3, > 15 itg/m3, and > 20 /*g/m3. Extrapolation
was based on inverse distance squared from the SURE ground
stations over the SURE grid.
222
-------
DAYS £ 10
AUG, 1977
SO4 DAYS 2 20
AUG, 1977
Figure 4. Calculated dosages of sulfate based on extrapolation of the
SURE network data for August 1977. Isopleths are number of
days > 10 jig/m3 (top) and >20 jig/m3 (bottom).
223
-------
N03(pg/m3)
AUG, 1977
250 Km.
Figure 5, Distributions of monthly geometric mean values of 24-h
average sulfate and nitrate concentrations extrapolated from
the SURE network data.
224
-------
ro
en
Figure 6. Evaluation of an elevated sulfate event in early August 1977. Conditions suggest
long-range transport across the Northeast associated with a channeling or ducting of
moist air flow. Pressures associated with highs and lows are in millibars; isobars are
drawn in with the sulfate isopleths. The sulfate levels are in two digits compared
with the isobars in four digits.
-------
TABLE 1. AVERAGE AIRCRAFT OBSERVATIONS* FOR SULFATE
AND NITRATE FOR DUNCAN FALLS, OH
Period
August 1977
Jan-Feb 1978
April 1978
July 1978
October 1978
Mean
(Duncan Falls)
Maximum
height (10~3 ft)
Surface
5
10
Surface
5
10
Surface
5
10
Surface
5
10
Surface
5
10
5
10
S04 (ug/m3)
9.1
12.5
7.5
10.1
5.6
2.0
3.9
4.8
1.9
10.6
18.0
3.1
6.3
2.7
9.4
3.4
N03 (ug/m3)
2.6
2.4
1.1
0.09
0.7
0.1
0.18
1.0
0.6
0.9
.0.1
1.3
0.7
*Excludes horizontal traverses; average of all flights per sampling period.
226
-------
TABLE 2. METEOROLOGICAL REGIMES CONDUCIVE TO REGIONAL SULFATE
ACCUMULATION AND TRANSPORT
Type
Stagnation
Channeling
or
ducting
Synoptic
pattern
Dominant high-
pressure cell
(surface and
aloft)
Stationary
front along
Great Lakes
Seasonal
preference
Any season
Summer
Transport
winds
Light (~5 m/s)
Moderate
SW wi nds
(7-12 m/s)
Afternoon
mixing
heights
Low <1,000 m
Moderate
(1,000-1,500
m)
Zone of
influence
100-300 km
Long-range
transport
(>500 km)
Thermo-
dynamics
Any temperature;
generally moist
Warm; moist
and/or westward
extension of
Bermuda high
PO
-------
PERFLUOROCARBON TRACER SYSTEM FOR ATMOSPHERIC
TRANSPORT AND DISPERSION STUDIES
Gilbert J. Ferber
National Oceanic and Atmospheric Administration
Silver Spring, Maryland
INTRODUCTION
A variety of atmospheric transport and dispersion models are being
developed under the auspices of EPA, DOE, and others to simulate the be-
havior of air pollutants and estimate regional air concentrations under
various scenarios. With increasing concern over regional and international
aspects of air pollution, reliable model calculations are required at dis-
tances out to 1,000 km from pollution sources.
Attempts to verify model calculations, using air quality data, are
complicated by the presence of multiple sources and imprecise knowledge of
emission amounts. There is a need for nonreactive, nondepositing tracers
that could be released at precisely controlled rates and measured accurately
at very low concentrations. This will allow us to conduct tracer experi-
ments that isolate atmospheric transport and dispersion from the other
complexities and provide data for verification of this basic aspect of all
model calculations. Regional scale model verification experiments require
tracers that can be unambiguously identified and measured out to hundreds of
kilometers or several days' travel time. Sulfur hexafluoride, SF6, has been
used out to about 100 km, but its relatively high and variable background
concentration militates against its use to much greater distances.
PERFLUOROCARBON TRACER SYSTEM
Investigations by J. E. Lovelock in England indicated that a perfluoro-
carbon tracer system could be developed that would be ideal for long-range
dispersion studies. The perfluorocarbons are extremely stable, nontoxic
compounds, measurable at very low concentrations by electron capture gas
228
-------
chromatography. Atmospheric background concentrations are well below that
of SF6. Under DOE funding, the NOAA Air Resources Laboratories (ARL) con-
tracted with Lovelock to develop three different prototype samplers as the
first step in the development of a perfluorocarbon tracer system.
The first instrument, a sequential sampler, consists of a pump and a
cassette of 24 tubes containing molecular sieve material to trap the tracer.
Air is pumped through each sampling tube in a preset automated sequence.
The cassette is returned to the laboratory and inserted into an analyzer
unit, which automatically heats each tube in turn, to desorb the sample into
a catalytic reactor that destroys unwanted components. The sample then
flows through a chromatograph column that separates the perfluorocarbon
tracer from other surviving compounds and then tracer concentrations are
determined with an electron capture detector.
Another prototype instrument combines the sampling and analysis func-
tions into a single unit. The unit contains two sampling tubes so that one
is sampling while the other is being analyzed. This instrument provides
readout of concentrations every 5 minutes at the sampling site.
A third instrument was designed to be flown in a small aircraft.
Ambient air is drawn through a catalytic reactor that reduces the 02 and
other electron-absorbers, leaving the perfluorocarbon and nitrogen. This is
passed directly to an electron capture detector providing continuous concen-
tration readout with a 3-second delay.
Prototype instruments were delivered by Lovelock in June 1976. Since
then ARL has been working closely with the DOE Environmental Measurements
Laboratory (EML) and Brookhaven National Laboratory (BNL) in a cooperative
o
effort to develop a practical perfluorocarbon tracer system.
Comparative data on SF6 and perfluorocarbons (PDCH, PMCH and PDCB) are
shown in Table I. The atmospheric background concentration of PDCH is about
-14
0.02 ppt by volume (2 x 10 x ), about 1/25 of the SF6 background. Background
of the other two perfluorocarbons (PMCH and PDCB) is an order of magnitude
lower. The amount of tracer released in any experiment must be sufficient
to distinguish the plume from background at the maximum sampling distance.
The required release rate for PDCH is about 10 percent of that for SF6; for
PMCH and PDCB it is about 1 percent of the SF6 rate. Taking the higher
229
-------
price of the perfluorocarbons into account, the PDCH required for an experiment
would cost about 70 percent as much as SF6; the cost of PMCH and PDCB would
be about 5 percent of the SF6 cost.
Perfluorocarbon release, sampling, and analysis techniques were tested
in a field experiment in Idaho in April 1977. Three perfluorocarbons were
released simultaneously with SF6 and samples were collected along arcs out
to 90 km from the release point. The experiment demonstrated that release
techniques worked well with all tracers released uniformly over a 3-hr
period as planned. Five sequential samplers performed well on the 50-km
arc. About 60 half-hour samples were taken and all data were consistent.
As the plume passed the arc, PDCH concentrations were measured from the
background of about 0.02 ppt to a peak near 10 ppt and down to background
again as shown in Figure 1. Perfluorocarbon measurements from these sam-
plers agreed well with SF6 measurements from whole-air samples taken at the
same locations.
The real-time continuous sampler did not operate properly in this test.
Subsequently, the BNL version of this instrument was successfully flown in
the MAP3S-AMBIENS experiment in October 1977.
IMPROVED SEQUENTIAL SAMPLER
The Idaho experiment established that the sequential samplers worked
well and that PDCH can be measured reliably down to its background level of
about 0.02 ppt. However, the molecular sieve traps used in these instruments
do not collect PDCB or PMCH, both of which have background concentrations
about 1/10 that of PDCH. In exploring perfluorocarbon trapping and analysis
techniques at BNL, it was found that charcoal efficiently traps all three
tracers but also collects many unwanted components that interfere with
perfluorocarbon analysis. A laboratory analysis scheme has been devised
that effectively eliminates interferences and achieves a sensitivity better
than 0.001 ppt (1 x 10 ). This achievement has led to a new concept for a
sequential sampler, using charcoal traps, that is smaller, lighter (under
20 Ib with self-contained batteries), simpler to operate, and less costly
than the Lovelock samplers. The sampler will automatically start at a
preselected time and take a preset number of samples (up to 24) of pre-
selected duration (up to 24 hr per sample). Two prototype units are sched-
230
-------
uled for delivery in April 1979. Tests should be completed by June and we
plan to contract for production units to be delivered early in 1980.
DEMONSTRATION EXPERIMENT
We plan to conduct an atmospheric tracer experiment in the spring of
1980 to demonstrate the capabilities of the perfluorocarbon tracer system.
Two perfluorocarbons will be released simultaneously with at least one other
tracer and the new sequential samplers will be operated on several arcs out
to about 500 km from the release point. Measurements of the different
tracers will be compared to establish the reliability and precision of
release, sampling, and analysis techniques. The BNL real-time continuous
monitor will be used to measure the vertical distribution of tracer and
cross-wind concentration profiles. Data will be used to estimate diffusion
parameters and to test model calculations of plume dispersion.
The perfluorocarbon tracer system will provide a capability, for the
first time, to perform long-range atmospheric transport and diffusion experi-
ments, at reasonable cost, for verification and improvement of air quality
models. Other applications of interest to EPA include:
1. Tracer studies of atmospheric transport and dispersion from pro-
posed plant sites;
2. Simultaneous release of several perfluorocarbons at different
sites to determine their relative contributions to the gross
pollutant burden at locations of interest; and
3. Study of long-range pollutant transformation and deposition rates
by injection of tracer into the stack and measurement of pollutant/
tracer ratios as a function of plume travel time.
APPLICATION TO PEPE STUDIES
The feasibility of using perfluorocarbons in studies of prolonged
elevated pollution episodes (PEPE) can be examined with a computer-simulated
tracer release. The ARL Regional Scale Transport and Diffusion Model1 2 was
used to calculate tracer concentrations for 5 days following a release of
PMCH. The simulation assumes a release rate of 100 Ib/hr for 3 hr (1,030-
1,330 GMT) at Nashville, TN, on August 18, 1976. A stagnating high-pressure
231
-------
system dominated the weather over the eastern half of the United States
during this period. The calculated trajectory, shown in Figure 2, travels
clockwise around the high, arriving over Michigan on the fifth day. Calcu-
lated tracer concentrations are shown for 12-hr air samples collected on the
third and fifth day after release. Contours are drawn for concentrations at
the limit of detection (0.002 ppt) and ten times the detection limit (0.02 ppt),
The simulation indicates that the tracer should be measurable over a large
area up to 5 days after a total release of 300 Ib of PMCH.
REFERENCES
1. Heffter, J. L., A. D. Taylor, and G. J. Ferber. 1975. A Regional-
Continental Scale Transport, Diffusion, and Deposition Model. NOAA
Tech. Memo. ERL ARL-50, Air Resources Laboratories, Silver Spring, MD
20910, 29 pp.
2. Telegadas, K., G. J. Ferber, J. L. Heffter, and R. R. Draxler. 1978.
Calculated and Observed Seasonal and Annual Krypton-85 Concentrations
at 30-150 km from a point source. Atmospheric Environment, 12:1769-1775.
232
-------
LOVELOCK CASSETTES (50km ARC)
10
5
^ 2
"Q.
Q.
o '-o
<
H as
2
UJ
O
0
0 0.2
X
o
o
°- o.i
0.05
0.02
0.01
• \
317 —
319 —
r 323--
~ 'aoK u
—
"
—
—
-
—
•
_
—
f L
"~ 1
| .*"" '
i
"I
1 / / / /
^™
i 1
1 | 1 | 1 | i | I | I
-
••
_• ____ • • ' _
PJn^^'i-fiT^ :
.1 ~^~"1 L4 '_
• i ^
• Ijcl
~™ ' i i *.
I Ln i ,» Ijj
— — Hi =
1 1 1 *
1 ?! • —
" IpU :
* i Jr?
1 G1
• i **|!*~r
l-"[,
x — x x— *i
_i «
* 1 -
}
BACKGROUND i-x-=i
/•///////////////^* <^* /
. —
i 1 i I i 1 i 1 i 1 i
1300 1400
1500 1600 1700 1800 1900 2000
TIME(MST)
Figure 1. Perfluorocarbon (PDCH) concentrations measured in %-hr sequential samples
at five locations on an arc 50 km from the release site.
233
-------
\
TRACER (PMCH)
Concentration after 3-hr
release (100 Ib/hr)
Figure 2. Calculated tracer concentrations (ppt) in 12-hr samples
taken 3 and 5 days after release at Nashville, TN.
-------
TABLE 1. COMPARATIVE DATA ON SF6 AND PERFLUOROCARBONS
Tracer
Sulfur-
hexa-
fluoride
Perfluoro-
dimethyl-
cyclohexane
(PDCH)
Perfluoro-
methyl-
cyclohexane
(PMCH)
Perfluoro-
dimethyl-
cyclobutane
(PDCB)
Formula
Mol. wt.
Background
(pptv)
Cost/kg
Relative
release
rate
Relative
cost/
release
SF6
146
0.5
$7
100
100
400
0.02
$45
11
70
C7F14
350
0.002
$45
0.9
300
0.002
$45
0.8
235
-------
MAP3S PROGRAMS APPLICABLE TO PEPE'S
Michael C. MacCracken
University of California, Livermore, California
SUMMARY
The MAP3S goal is to improve the capability to simulate the behavior of
energy-related pollutants in the atmosphere. This requires improving the
understanding of atmospheric behavior and of the processes affecting pollutant
evolution in the atmosphere. MAP3S has conducted a wide variety of field
experiments and characterization studies to improve understanding of wet and
dry deposition, pollutant transformation, planetary boundary layer evolution,
aerosol speciation, long-range pollutant transport, precipitation chemistry,
and pollutant budgets. While MAP3S will be continuing in FY 1980 to focus on
its own program objectives, the objectives of PEPE (Persistent Elevated
Pollution Episode) are sufficiently related to those of MAP3S to permit con-
sideration that some MAP3S resources could be diverted to PEPE studies during
limited intensive periods in the summer of 1980. Resources and capabilities
that may be available can be divided into the following three areas, with
subareas also listed.
A. Network Support
1. Precipitation chemistry network data from the eight MAP3S and nine
Electric Power Research Institute event-basis sites can be used to
measure the final removal stages of PEPE's.
2. Aerosol acidity of fine particles, in terms of the molar ratio of
hydrogen ion to sulfate ion, with a 4-hr time resolution at five
sites.
3. Turbidity data may be available if the network that operated from
the summer of 1977 to fall of 1978 is reestablished so as to inter-
face with the National Oceanic and Atmospheric Administration (NOAA)
network.
236
-------
B. Field Capabilities
1. Pacific Northwest Laboratories and Brookhaven National Laboratory
(BNL) instrumented aircraft may be available for episode studies by
diverting them from ongoing MAP3S studies.
2. MAP3S funded tracer capabilities, developed by a team from BNL,
Environmental Measurements Laboratory, and NOAA, are expected to be
ready for major field implementation in 1980. (Refer to paper by G.
Ferber in this volume.)
3. Special sampling instruments, including the diffusion processor, BNL
filter packs, and the BNL-developed real-time sulfate/sulfuric acid
monitor.
C. Modeling and Analysis Capabilities
1. Data from the August 1977 and July 1978 studies of elevated $04
concentrations can serve as the basis for PEPE preplanning.
2. Source emission data are available that may indicate likely PEPE
source regions.
3. MAP3S trajectory models are expected to be able to operate using
National Weather Service forecast trajectories as a means of esti-
mating the evolution and movement of the polluted air mass.
Because of uncertainty in the planning of MAP3S activities for 1980 and
beyond, resources in addition to those now available to MAP3S participants may
be required.
ACKNOWLEDGMENTS
This work was supported by the U.S. Environmental Protection Agency
through interagency agreement with the Department of Energy. The work was
performed by the Lawrence Livermore Laboratory under contract no. W-7405-Eng-
48.
237
-------
BROOKHAVEN
LEWES DE
ILLINOIS
OXFORD OH
figure 1. Locations of MAP3S precipitation chemistry sites.
238
-------
Figure 2. Location of MAP3S aerosol acidity (X) and (former)
turbidity network (•) sites.
239
-------
MAP3S CHARACTERIZATION FLIGHT BNL 8/9/77 1300-2100
ro
-p>
o
SO2 =
SO; =
5.4
6.6
XX SURE/Class I Station
31.3
13.4
SO2 = 29.5
SO; = 14.3
= 37.6
SO! = 13.6
Figure 3. Typical result from horizontal aircraft flights in the well-mixed layer (/ig/m3).
-------
20
PO
16
I
L-
**—
• "5
i 8
o
o>
0
19
O = Real-time sulfate, valid zero
D = Real-time sulfate, marginal zero
A = Filter pack sulfate
20
21
Date July, 1978
22
23
Figure 4. Comparison of BNL real-time and filter pack measurements of sulfate.
-------
i-ff
AM
r— --
PM
1 "" i ~ 7" ^—
figure 5. Aircraft sulfate measurements during SURE/MAP3S
intensive study on July 19, 1978 (/*g/m3).
242
-------
REGIONAL SCALE AIRCRAFT MAPPING OF HAZY AIR MASSES
DURING PROJECT MISTT, 1975 and 1976
D. E. Patterson, J. M. Holloway, and R. B. Husar*
Washington University, St. Louis, Missouri
1.0 INTRODUCTION
The measurements reported in this document were made as part of the
1975 and 1976 field programs of the Midwest Interstate Sulfur Transformation
and Transport (MISTT) study. Project MISTT is an integrated interdisciplin-
ary research program of field, laboratory, and theoretical studies sponsored
by the U.S. Environmental Protection Agency. The principal objective is to
obtain a physical understanding and quantitative estimation of the nature
and fate of atmospheric sulfur in large power plant and urban-industrial
plumes.
Although MISTT focused on plume studies, there were a few flight days
in which measurements were made on a regional scale. Four such missions
have been chosen for analysis in the context of this PEPE seminar. Each
mission characterizes the air mass of interest on a scale of hundreds of
kilometers over a period of at least 8 hr, and each represents a different
sampling approach. It is hoped that these missions can serve as examples
(good or bad) for use in designing future regional scale sampling programs.
The analysis of each day consists of four levels of information:
1. The meteorological setting is described through isopleth maps of
noon surface temperature, dewpoint, sea level pressure, relative
humidity, and wind field over the eastern United States. The
observed noon visibilities (corrected to 60 percent RH) are dis-
played as light extinction coefficient, b. .. The extinction
coefficient, or haziness, has been shown to be a qualitative
surrogate for sulfate concentrations.
*Speaker.
243
-------
2. The flight path, time of day, and altitude of sampling are de-
scribed in detail. Locations of traverses and vertical soundings
are shown on an appropriate regional map.
3. Pollutant parameters are displayed by superimposing the measured
values on the flight path for ease of interpretation. The contin-
uous measurements usually include S02 and ozone in addition to
three aerosol parameters: condensation nuclei count (CNC), aerosol
charge, and the light scattering coefficient (b .). Particulate
sulfur (as sulfate) was obtained from filter samples by the flame
photometric detection (FPD) method.
4. Vertical profiles of gaseous and aerosol parameters add the third
dimension to the picture of the atmosphere. Profiles of tempera-
ture, turbulence, and dewpoint are also provided where available.
Sulfate concentrations are averages over each spiral and are
listed in the corner of each plot.
The information is primarily in graphical form, so that much of the
report is to be found in the figure captions.
244
-------
2.0 JULY 29-30, 1975
2.1 Flight Description
A large high-pressure area dominated the eastern portion of the country
from the Missouri Valley to the east coast. The center of the ridge, both
aloft and on the surface, was situated over the eastern Great Lakes. To the
west, a rather well-defined trough, extending from the surface to 9,145 m
MSL, was located over the western plains. Since the Mississippi Valley was
between these two features, a strong pressure gradient pattern favoring
south to southeasterly winds affected Missouri, Iowa, and Illinois through-
out the air sampling period. Since the flow pattern tended to be more
southerly closer to the trough, winds favored a southeasterly direction east
of the Mississippi River and a southerly direction west of it.
Skies were clear throughout the period but considerable haze caused
daytime visibilities to be 7 to 8 km and nighttime readings to average near
5 km. Lower visibility readings were particularly evident east of the
Mississippi and south of Iowa and Wisconsin.
Temperature soundings displayed nearly adiabatic lapse rates from the
surface to 1,830 m MSL. Above that level, neutral conditions capped a
shallow +1° C inversion. The diurnal surface inversion started to establish
itself by midnight on the 29th, extending up to about 460 m MSL by sunrise
on the 30th.
Wind direction variation with height was more than 10° to 15° throughout
the period of sampling. Along the east of the Mississippi River, directions
were from 120° to 150° while directions from 170° to 190° became apparent
west of the river, especially during the early hours of the 30th. Wind
speeds at all the sampling locations were between 10 and 20 m/s at all
levels.
The flight consisted of traverses at several discrete altitudes oriented
crosswind to the air mass trajectory. In addition, spirals were made at
each downwind distance. The mission began at 1300 CDT July 29 and ended at
1130 CDT on July 30.
2.2 Summary
The July 29-30, 1975, mission sampled a very hazy air mass as it passed
over St. Louis, covering over 20 hr from beginning to end. Sulfate concen-
245
-------
trations in excess of 60 ug/m3 and light scattering coefficient near 10 x
-4 -1
10 m were encountered northeast of Quincy, IL.
The sampling pattern consisted of multiple traverses at discrete down-
wind distances complemented by spirals over a fixed location. This technique
provides excellent resolution of the three-dimensional structure of the air
mass, but requires too much time for truly regional scale sampling.
246
-------
ro
JUL 25, 1975
RH CORRECTED BEXT
NOON TEMPERRTURE
TEMP
NOON OEUPOINT
JDENP
16.3
SURFflCE
MIND FIELI
JJIND/L
SEfl LEVEL PRESSUR
RELATIVE HUMIDITY
0. 45. SO. 79. 100.
IOH- 1010
1016- 1010
1018- 1020
> 1020
;;;; 40-60 m 80-90
= 60-80 • >90
Figure 2-1. Stagnating winds from the Ohio River Valley south to Florida are associated with the beginning
of this episode. Noon surface readings indicate temperature near 28° C, dewpoint 20-24° C. RH
of about 60-80 percent with extinction coefficient of ~3*10~4m~1 (visibilities of ~8 mi).
-------
JUL 26, 1975
RH CORRECTED BEXT
BEXT
NOON DEHPOINT
JDEHP
DEC C
;:;; 12-16 m 20-24
= 16-20 •
INS
4*
00
SURFACE MIND FIEL
» x//J7 11tn,.
\ tfftt /
0. ZS. SO. 75. 100.
SEfl LEVEL PRESSURf
PRES
RELATIVE HUMIDITY
1018- IOM
> IOM
iiii 40-60
= 60-80
80-90
>90
Figure 2-2. The region with bext >3»10~4m~1 has expanded into the MO-IL-IO area. Note the stagnant
region in the upper IN-OH region associated with the high-pressure cell.
-------
ISJ
JUL
RH CORRECTED BEXT
BEXT
= U-6
SLIRFflCE MIND FIEL
0. it. 90. 75. 100.
1975
iiii 20-2U
= 214-28
28-32
>32
SEfi LEVEL PRESSURl
.PRES
lots.;
:::: IOU- 1016
= ioie- 1010
ioia- 1020
> 1020
NOON OEMPOINT
;;;; 12-1 6 m 20-24
= 16-20 • >24
RELfiTIVE HUMIOITT
iiii yo-eo
= 60-80
80-90
>90
Figure 2-3. The region of haze covers a large part of the east-central United States by July 27.
-------
JUL
RH CORRECTED BEXT
BEXT
1975
NOON DEWPOINT
;;:i 12-16 m 20-24
= 16-20 • >24
PO
en
o
SEfl LEVEL; PRESSUR
0. 2S. SO. 75. 100.
RELRTIVE HUMIDITY
:::: 10iv- I0i« s 1018- 1020
EE 1010- iota • > 1020
;;;; 40-60
= 60-80
80-90
>90
Figure 2-4. Continued stagnation in the Ohio River Valley is associated with extinction coefficient
>4»10~4m~1. Noon surface temperatures are 28°-32° C with relative humidity typically
40-60 percent. The area around St. Louis, MO, is stagnant.
-------
tn
JUL 29, 1975
RH CORRECTED BEXT
BE XT
10MM-4/M
:::: 3-4 M 6-8
= 4-6 • >8
SURFRCE NINO FIEL
NOON TEMPERRTURE
SER LEVEL: PRESSUR
0. 29. SO. 79. 100.
IOI«- IBI6
1010- 1010
lOlt- IOM
> 1020
NOON DEWPOINT
;;;; 12-16 m 20-24
= 16-20 •
RELRTIVE HUMIDITY
;;;; 40-60
= 60-80
80-90
>90
Rgure 2-5. This is the first day of the sampling mission. The wind and pressure fields have changed
dramatically over the past 24 hr; the St. Louis area is in the region of lowest visibility.
-------
JUL 30, 1975
RH CORRECTED BEXT
kBEXT
NOON TEHPERRTURE
JEMP
DEC C
iiii 20-24 M 28-32
= 24-28 • >32
NOON DEMPOINT
JDEWP
20-24
>24
on
ro
JRFRCE MIND FIEL
0. 89. SO. 75. 100.
SEfl LEVEL PRESSURI
J>RES
joie,
RELflTIVE HUMIDITY
IOI«- 1016
1010- IOIS
1018- 1020
> 1020
;;•• 40-60
= 60-80
80-90
>90
Figure 2-6. The MO-IL-IA region is the area of lowest visibility; the aircraft have moved to the north along
with the hazy air mass. Temperatures have increased since the 29th along with dewpoint. The
black area on the bext plot represents noon visibility less than 3 mi.
-------
JUL 31, 1975
tn
co.
RH CORRECTED BEXT
.BEXT
RFflCE HIND FIEL
NOON TENPERRTURE
JEMP
NOON DENPOINT
DEC C
;;;; 20-24 m 28-32
= 24-28 • >32
SEfl LEVEL-' PRESSUR
•. ts. so. n. too.
1011- !•!•
101*- !•!•
ioia- 1020
> IOW
20-24
>24
RELRTIVE HUHIDITY
iiii 40-60
= 60-80
80-90
>90
Figure 2-7. The hazy air is now moving out of the St. Louis area toward the Canadian border.
-------
10
ST. LOUIS HRZINESS
8
T I I I I T
JULY/flUGUST 1975
ii
I I
I I
252627282930 31123
Figure 2-8. Noon extinction coefficient in St. Louis, MO, calculated from surface visibility.
The haze episode peaked on July 29, dropping off sharply by the 31st.
254
-------
TABLE 2-1. FLIGHT OUTLINE, JULY 29-30, 1975
Altitude Time
(m MSL) Route (CDT)
451 Point A to Brighton 1335-1342
(tuned to St. Louis VOR)
610 Brighton to Point A 1345-1351
762 Point A to Brighton 1354-1400
915 Brighton to Point A 1403-1409
1460-410 10 km NE of Point A (spiral) 1418-1427
INOPERATIVE
762 Point C to Point D 1502-1511
915 Point B to Point C 1525-1540
1460-290 Point E (spiral) 1546-1556
451 Hettick to Point F 1620-1628
457 Pearl to Point G 1643-1657
(tuned to Maples VOR)
INOPERATIVE
***Refuel***
180-2590 Jacksonville (spiral) 1854-1920
1061 Pearl to Point G 1949-2009
(tuned to Maples VOR)
1475-390 Point H (spiral) 2024-2043
915 Point I to Point J 2045-2053
915 Point K to Point L 2056-2119
756 Point L to Point K 2124-2146
610 Point K to Point L 2148-2210
1650-450 Point M (spiral) 2222-2231
***Refuel***
1900-400 Point N (spiral) 0014-0028
604 Monroe City to Havana 0042-0135
(tuned to Maples VOR)
909 Havana to Monroe City (partial) 0140-0233
INOPERATIVE
***Refuel***
762 Peoria VOR to Macomb 0519-0543
(tuned to Peoria VOR)
255
-------
TABLE 2-1 (continued)
Altitude
(m MSL)
756
756
1067
451
1061
1990-290
756
604
Route
Macomb to Point 0
Point 0 to Point P (partial)
(tuned to Macon VOR)
Point P to Point R
Navoo to Point S
Point S to Kirksville
***Refuel***
Cannon Memorial Apt. (spiral)
Kirksville VOR to Chillicothe
Chillicothe to Point T
Time
(CDT)
0544-0612
0613-0640
0648-0719
0739-0832
0837-0852
0901-0912
1034-1052
1054-1125
256
-------
PEOWA
-~^*=S=^«V>>^: *.
July 29-30, 1975
1330 - 1130 CDT
- Spiral
Rgure 2-9. Map showing location of traverses and spirals for July 29-30, 1975. The numbers 1
through 18 indicate the positions of the spirals displayed in Figures 13-18.
257
-------
OZONE
JULY 29-30, 1975
Rgure 2-10. Crosswind profiles of ozone at 760 m MSL. Peak ozone concentration is
approximately 140 ppb, with levels generally near 100 ppb ozone.
258
-------
SO
JULY 29-30. 1975
Springfield) .
Jacksonville
<*&
Rgure 2-11. Crosswind profiles of SO2 at 760 m MSL on July 29-30, 1975. At 1400
COT (the closest traverse) the easternmost peak corresponds to the Wood
River, IL, power plant/refinery complex; the western SO2 peak is due to
the Portage des Sioux power plant. Both sources may be detected at the
next two downwind distances as well.
259
-------
B
SCflT
JULY 29-30, 1975
Figure 2-12. Average suffate concentration at 760 m MSL ranged from 24 to 61
pg/m3. The highest 804 was associated with extreme haze of bgcat
10*10~4m~1. Light scattering coefficient ranged from 5 to 10*10~4m~1,
showing evidence of the contribution of many area and point sources.
260
-------
1.1
1.0
i ••«
*
§ '•«
*
O.I
•KM IIO-Vlll
it, lt.f.l.>
« IW/MI
e»«»eilW....
1.1
1.0
« 1.1
*
t—
a!
O.I
lot Iio-Viil
50, If.f.O.I
M \\vnn
emmet «v....
i.l
1.0
1 ,.l
§ '•'
5
O.i
.... | ....
Mtf I OT-JO-7J | N1MM .
TIM i 11.31,11 . I1.J1.SI .
1 -
_ - —
•
- -
ff^ j& ''•
f '•• S** i
01 10 IS 10 11 10 11 10 11 S
0 10 U M M IM IM IM IM IM *
. • . . | . • •
Mil 1 O'-JO-W 1 HIMM .
IINC I IlilOiM - UllJ.St .
i 2 -
• 1
iJizsL__ 1
• . . . . I....'
g 1 10 19 10 19 10 U 10 IS S
0 10 M M M IM IM IM IM IM K
Hire i OJ-M-W i HTMM I
L 3 -
^^fer "
- > C^b J.
": '• _J> ^»
r^ ; ^ ~
: sou - qa.u :
i ....
I.O
i "
X
jjj 1.0
<*
o.s
Q 0
) !(•> 1^1
t IUM ICM/KSI
o OCM r ifcj
M OIOM !'..»..»."
1.1
1.0'
5 1.1
I
a I 0
§ '••
5
O.I
0.0
) !«»» JJI
1 IIMO ICIll'1/ll t
i M» r ffi e
M ItOtt if,f..(..l 1
2.0 •
*' '•' :
•x.
I •••!
_i .
IT
O.i •
0.0 •
.........
MIC i OI-M-7S i liniM
UK i u. ii. si • ii.u.ss .
1
-. _
-
r . ] ^ , . \i . . i
•10 1 10 II 10 n 10 » M 1!
0 SO IM ISO 100 ISO WO IM IM HO SOI
0 10 IM ISO IM MO MO' IM 100 IM U!
.........
onrc i 07-M-Tt . IITMM
. UK i K.IO.M - IlilTiSI .
r 2 -
- \ :
LX \4 "i
r* * " *
• . . . . i . . . . •
SO S 10 IS 10 11 10 n 10 IS
SO 100 ISO IM ISO MO ISO IM IM SOO
U IM IM IM MO m IM IM «* SOO
Wit i 07-JO-TS i N7MH '
f IKff • !•• 1 9i4l • !•• >8i^S
3 j
• * \ : ~
- ^^\ \ -
"^^^ \ ': •
^" ^ ' ~
•
1 . . . .
C« \l.
-------
o.o
'Mil l 07-M-7I l MTMM J
IINC i ISHOiM - IftMiiO J
In iio-Vm 0
u> IF.F.I.I a
« wean »
CMMIM".... 0
l.t
o.o
lu 11 at/ft- a
Ul IF.F.I.I 0
:• IWVtfl o
.. o
0.0
dm no*tin" o
Ut IF. P. 1.1 0
c« iW/jft 0
CMMIIIRI.... t
'(MIC l 07-JO-7S | M7MM
III* i ISi«0.« • ISnSiSO
I
I
t
M
10
ii
it it
M W
M
IH
IM IM Mt
IIMI
«•» ft).... I
.F.I.I 0
i s 10 .11 10 is u n <• «
111*1171111
M IM IM Ml IS* Ml Ml Ml IM Ml
M IM IM IM Ml Ml Ml «M «M Ml
'(»« i 07-JO-7S
tlMC i ISiMi* •
I H7WM
ISiMtS
88.5
I.I
* I.I
'MIC i 07-jo-n i DIMM
IINC i ISi««.I - IS.SI.S
5
I
I
I
M
a
10
j
7
IS
10
IM
I
I
«
90
10*
OK F ffl 0
OIUM tF-.F.-.l...l 0
ISO 200 ISO MO ISO 100
-------
ho»
Ml
u
CMM
SO,
u
CMM
•am
SO.
U
CMM
1.1
1.0
* I.I
X
I "'
a
o.s
' \ ootc i 07-io-n i *7nss ;
i \ IIW i ItiMlT - l*lUi«l
-V 7 -
: ^^
?J
— —
- -
• . . . . 1 ....
1.0
i '••
X
I •••
0.0
:f \
~\ \
\\ • . \
-S >
' >
: x 1
••
-
•
i.O
i ••»
-
0 1.0
O.I
! 0*TC • 07-30-71 i N7WU
tIMC i MiMiU - IOiUil7
8 _
- -^
:l ^
t^ 4^,
:^==- " ' ^ — ^ ;
H ^ •:
: sou « no. 1
t.i
t.o
* l.l
at
n
»-
"""\
•
•
•90 9 10 II
0 90 IM 110 ni
I.O
£ .1.1
| Ul
a
0-i.
Wit I 07-JO-Tt t «7Mn
ttIC I Ml3S«M - JOlMll*
9
-
- ,
s >
i ....
2.0
£ l.l
c
3
O.S
I1I-./V i1 i1 i1 ,' i' •' ,' •' i' ,g '«» flJ—
O.F.I.! 0 1 J 1 , J 1 7 1 I 10 '«" «^"
Wall 01 10 ii M n n n M u ii "** ™-;-
I0W'.... 0 M M M M IM l» IM IM IM no (TiMJ
-
-
" \ <
-j
•
-90 1 10 IS
i u 100 iso no
1 M IM IM IM
IJMC I 07-JO-n i H7M»
flit i l«iMi7 - III 111 tl
; 7
1
\
\ i
\
™*
-
1 ....
n n M II M «i
m MI m MI ui MI
Mft I 07-30-71 i N7MSS
(IMC I MlJllM - Ml»ll7
8
-
\ *
\--jl
\ \
V
1 . . I .
20 29 10 H 40 n
2SO MO SU 400 110 (ft
MFC i 07-SO-7S i -TM55
TIK i 20iS)*U - 20iMit2
9 _
-
\
\
i ....
20 n 10 19 W 19
291 Ml IM «M «M 900
211 IM IM Ml ISO 900
Figure 2-15. Spiral #7 indicates a "lid" at 1,700 m, capping the mixed layer with bscat
of 8.5. Here SO2 is less than 1 ppb, with ozone 100 ppb. One hour
later, at 2030 CDT, a fresh power plant plume was encountered at 850 m
with peak SO2 of 60 ppb; bscat shows a marked vertical inhomogeneity.
263
-------
J.I
MIC i 07-JO-7S i H7MM
IIMC I ttitiM - Ml III*
10
.... I . ..."
in MO-VIM oiaiis«>««n
u, ir.f.i.i *iii«iiii»i*
u uV/m o » 10 is n M M n M «i u
CMKI(W.. .0 M M M 0* IH IM «• IM IM «••
0.0
Will i 07-JO-W I HIMM
10
1
..
III* I'll -SO S 10 IS 20 2S M IS 10 IS
Timl (CJJ/I/VO I21HSIIOIIO
om r ff\ o so 100 iso 200 2so 100 iso not ito soo
UMf H..M..I • M IM IM m IM Ml IM 100 IS* SOO
oarc i 07-u-n i nnun
IIW i M.JliM - UiMi»
11
= 70.7
0.0
lui UQi/m 0121
so, if.r.t.i o 2 « •
CM UQ.Vp.'l o s 10 is
CHWCI(I1«> 0 20 HO iO
S I ' > I 10
10 12 11 It !• 20
2S M IS 10 ft SO
100 120 110 IM ISO 200
0.0
onrc i OT-tt-n
TIM i Mill i U - HiMiII
n
I€M Itl -50 S 10 IS 20 ii 10 IS 00 IS
IUM IIH2'1/SIO !2llS61«flO
OCU f '.'O.... 0 SO 100 ISO 200 *SO 100 ISO 100 ISO SOO
oiom IP.F.O.I o so 100 iso 2M no 100 >so 100 «so soo
MIC l OT-30-H i IHMM
IIM i III Sit 11 - OOtliil
0.0
(MTt i 07-SO-TS i RTMM
IIW I ZliSJill - OOilJil
12
n» O) -so s 10 is 20 n 10 is 10 is
IUM ICm/1/SIO I21«SI1ISIO
om f (K) o so 100 IM 200 2» 100 IM «oo iso soo
UM r.-r.-«.o • M IM IM IM >M m >M M* «M soo
Figure 2-16. Spiral #11 was obtained immediately after and - 10 km east of Spiral
#10. Both the vertical structure and the levels of bscat are substantially
different. At Quincy, IL (spiral #12), bscat is 6.
264
-------
1.0
out i oi-M-n i MMM
UK i oiIIIIV • Ot»i«
I
SOU > 39.2
no* n» oilltSI'llll
IP.IM.I OI211S«»««IO
Wf/ffi. 0 » IOIIM2IMMMWM
cwiwfvn>..
I M M M
IM »0 IM IM IM 2M
Mil I OT-M-71 I RIMM
II* i Olllllt - Ol»iJ»
13
' . . . . I
Illf I«C> -10 S 10 IS 20 21 10 IS «0
tUM ICM/1/lie 1211'SIIIIIO
om t if! o so IM IM 200 no MO iso .1 0 s
W.... i M
1.1
0.0
Mi If.f.t.l 0
u UV/."fl o
CMHKI W»l.... I
sou
ii to it n ii ii M
II XI n 10 39 II It U
MM IM IN l«0 IM IM Ml
MIC I 07-90-11 i KKW
IIM I fllOiM - *il»iM
15
CJ
I21VSI7IIIO
1 I I II II II 21 M 27 M
S II II 20 21 M II M «1 SO
M M M M IM IM IM IM IM 2M
MIC i 07-U-n i KIMM
IIMC >
10 IS 20
IUM ICH2/1/SI0 I
«» ' «!.... 0 SO
IM ISO 2M 2SO MO ISO «M «0
IM IM 2M 2M MO HO WO ISO
-------
O.f
Mff i OT-aO-n I OTMM
TIM i Oits.u -
16
• . . . . I . . . .
0111*91711 10
Ut O.f.l.l 01 11«I070IIO
" Wttffi. at n ii » M M n «* u u
CWKICI wi.... a n M M M IM in IM IM IM no
i.l
O.I
0.0
MK i oJ-w-n i *nni»
nm i iiMiM -
\
16
I
ig ii n n M n «o
1IM> Itl -SO
IMS ICM'l/ilO I J I « I • > I « 10
mttr ffi o so 100 ito too no MO no uo «to too
OIOM l»..f..(."l 0 M IM IM no »0 MO MO M« IM MO
0.0
den inrvm Oill
Mt If.f.O.) OI11
" IWVlft •• IIH
0.0
io
IO
M
oo too in IM iu 100 no
MIC I 07-JO-W i II7MM
riw i tOiMiw • lo.ii.n .
18
I
IIO-VIII OIIJ«t0700IO
if.ro.i oiii«to7ioio
Wf/t* 01 lOllnMMMMMM
«•>.... • . n W M M IM in IM IM IM 3M
1.0
0.0
Mil i 07-M-rt i H7MM
IIM i tlOiTO - «llli«l
17
-so 5 10 is so M u n oo us
IUKO icm/i/iio i j ) » 5 6 i • t 10
on r (u o M IM IM no tM MO »• uo i i 10
our M o M IM iu no MO MO iu too MO too
OIOH if..r..»..i o M IM IM no no MO wo IM no son
Figure 2-18. Spiral #17 indicates bscat of 7-7.5 within the 500-m mixed layer at
9 a.m. on July 30, with elevated haze layer of4-5*10~4m~1.
Average sulfate concentration from surface to 2,000 m was 48 /tg/m3.
By 10:30 a.m., bscat is more nearly mixed in the lowest 800 m.
266
-------
JULY 29-30, 1975
2.5
2.0
t.S
o 1.0
0.5
0.0
UQ-VH) 0
T i
5 6
8 8 10
Figure 2-19. Overlaid soundings from July 29-30, 1975. This period was very hazy,
with bscat ranging from about 4 to 8 in the lowest 1,300 m over a
24-hr period. Sulfate concentrations varied from 25 to above 80 /ig/m3.
267
-------
3.0 AUGUST 10, 1975
3.1 Flight Description
A large flat-surface high-pressure area dominated the entire eastern
portion of the country from the Mississippi Valley to the coast. A strong
zonal flow pattern did exist aloft over southern Canada but it only extended
as far south as the northern plains. The southern two-thirds of the country
was influenced by a very featureless upper and midlevel pattern. Stagnant
air conditions were common from St. Louis to New York. A well-defined
surface low moved across southern Canada, just south of James Bay bringing a
cold front behind it. However, the front never moved farther south than
central Nebraska and Iowa before it became stationary. The more northerly
portion of the front did follow the Canadian border eastward enough to clear
Minnesota and Wisconsin. A large outbreak of air-mass thunderstorms occurred
in Ohio and eastern Indiana while a line of showers and weak thunderstorms
moved out of Iowa and southeastern Minnesota into Wisconsin and northern
Illinois.
Sky conditions were generally clear from Kansas to Indiana through most
of the day. Visibility was only 3 to 7 km over eastern Missouri, Illinois,
and western Indiana during the morning with improving visibility in the
afternoon. The area of lowest visibility moved into Ohio, southern Michigan,
and northeastern Indiana by 1300 CDT. Visibility improved significantly
about 150 mi east of the cold front throughout the day.
Wind flow patterns were south-southeasterly during the early morning,
becoming southerly—then south-southwesterly by late morning, through the
afternoon. This pattern had a tendency to move the area of lowest visibility
toward the northeast.
The main leg of the flight consisted of a pass at 750-m altitude from
Chillicothe in Northeast Missouri to Indianapolis, with few spirals taken.
The main path began at 1000 CDT and ended about 1400 CDT; the overall mission
covered 0500-1800 CDT on this date.
3.2 Summary
The main traverse of August 10, 1975, was at a constant 750-m altitude.
The aircraft found a wide haze plume in central Illinois and a sharp gradient
268
-------
of b . near the Illinois-Indiana border. Both aerosol charge and condensa-
tion nuclei count also increased sharply from eastern Illinois into Indiana.
The flight plan provided a good "snapshot" of the atmosphere across the
central midwest, but provided relatively little vertical information. It
appears that on this day there were several hazy layers aloft, even at
midday, and thus the evolution of a hazy air mass may well involve varia-
tions in day-to-day mixing height.
269
-------
fsj
^J
o
RUG 09, 1975
RH CORRECTED BEXT
.BEXT
10MN-4/M
;;;; 3-4 •. 6-8
= 4-6 • >8
SURFflCE MIND FIEL
NOON TEHPERRTURE
JEMP
SEfl LEVEL; PRESSURE
PRES
0. 2S. 90. 75. 100.
:::: IOIH- 1016 M IOIB-1020
EE 1010- ioi» • » 1020
;::; 12-16 m 20-24
= 16-20 • >24
RELflTIVE HUMIDITY
iiii 40-60
= 60-80
80-90
>90
Figure 3-1. A high-pressure system was present over the entire eastern United States. Noon surface
temperatures were in the range off 28°-32° C. The center of the hazy air mass was in south-
central Illinois.
-------
RH CORRECTED BEXT
BEXT
10MN-4/M
3-4 m e-a
4-6 • >8
1975
NOON DEWPOINT
IENP
20-24
>24
IVi
URFflCE MIND FIEL
0. 25. SO. 75. 100.
SEfl LEVEL. PRESSURI
PRES
IOU- 1018
1010- 1010
IOI«- 1020
> 1*20
RELflTIVE HUMIDITY
RH
40-60
60-80
Figure 3-2. The haziest portion has moved into the Indiana-Ohio region, with expansion of the region of
reduced visibility into all of Missouri. Winds from St. Louis to Indianapolis are out of the
South. The high-pressure system still prevails over the eastern United States.
-------
RUG 11. 1975
RH CORRECTED BEXT
BEXT
SURFACE HIND FIEL
»r 111\\«. * tMTlo -4—..
*' /VJXV vJf-P'V *%•"•*
\ \ (A J t«/xai^r^3^^r
^>VOl ' • -N.-Wx-X*, ,^,>
'/\\\ » • • -**«<''''•" - >.r*T
20-24
21-28
LEVEL: PRESSURI
.PRES
tots.;
0. 25. SO. 75. 100.
IOI«- 1010
1010- 1010
1018- IOJO
> 1020
NOON PENPOINT
:::; 12-16 m 20-24
= 16-20 • >21
RELflTIVE HUMIDITT
RH
;58.6.
yo-eo
60-80
80-90
>90
Figure 3-3. The pressure and wind fields have become disorganized; St. Louis is still in a hazy region.
-------
TABLE 3-1. FLIGHT OUTLINE, AUGUST 10, 1975
Altitude Time
(m MSL) Route (COT)
590-2570 Wentzville (Spiral) 0512-0534
762 Wentzville to Boonville 0542-0637
(Tuned to Maryland Heights VOR)
770-2570 Boonville (Spiral) 0639-0653
762 Boonville to Blue Springs 0659-0745
(Tuned to Blue Springs VOR)
756 Cameron to Chillicothe 0812-0832
***Refuel***
310-1960 Chillicothe (Spiral) 1012-1030
762 Chillicothe to Hannibal
(Tuned to Quincy VOR)
730-1960 Hannibal (Spiral) 1125-1138
756 Hull (Hannibal) to Tuscola 1142-1259
(Tuned to Capital VOR, changed
to Decatur VOR at 1247:00)
820-2110 Tuscola (Spiral) 1259-1310
762 Camargo (Tuscola) to Indianapolis 1317-1353
***Refuel***
756 Green Castle to Brazil 1530-1541
(Tuned to Lafayette VOR)
756 Terre Haute to Casey 1550-1609
(Tuned to Terre Haute VOR at
1601:36)
756 Springfield to Hillsboro 1707-1729
(Tuned to Capitol VOR)
756 Litchfield (Hillsboro) to Brighton 1733-1750
(Tuned to St. Louis VOR)
273
-------
August 10, 1975
0500-1800 CDT
A- Soiral
Figure 3-4. Map showing the flight path and location of spirals for August 10, 1975. The numbers 1
through 10 indicate the positions of the spirals displayed in Figures 3-10 and 3-11.
274
-------
OZONE
cn
HUGUST 10, 1975
0.20-
0. 16-
0. 12-
0.08-
O.OH-
0.00-
Figure 3-5. Ozone was slightly higher in western Missouri than in the Mississippi Valley, with a small
gradient from Illinois into Indiana.
-------
ro
-j
en
flUGUST 10, 1975
Figure 3-6. Particulate sulfur was highest in western Missouri and western Illinois.
-------
CMC
INS
RUGUST 10, 1975
Figure 3-7. Condensation nuclei were present in high concentration in western Missouri and central
Illinois, wth a strong consistent peak east of the Illinois-Indiana border.
-------
CHflRGE
flUGUST 10, 1975
00
0. 10-
0.08-
0.06-
o.oq-
0.02-
0.00-
Figure 3-8. Aerosol charge shows the influence of several point and area sources in Missouri and
Illinois, with a strong and consistent increasing gradient from the Illinois-Indiana border
heading toward Ohio.
-------
B SCRT
RUGUST 10, 1975,
ro
^j
UD
5.00-
4.00-
3.00-
2.00-
1.00-
.00-
Figure 3-9. Light scattering (haziness) behaved much like CNC and charge, peaking in central Illinois
with a strong gradient east of the Illinois-Indiana border.
-------
0.0
MIC i M-IO-T*
Il«€ i tiliM - SitiU
-so t 10 u 20 n M n u
TIMO icm/i/iio i>i«titoiio
M" ' tfpj o u io« iu no iu wo HO too na MO
OIOM if.f..«..i o w tot iu no no MO IM wo IM ua
0.0
lor i«ei -jo s 10 n u n . 10 n to n
1UU lCIU/1/ll 0 IlllSllltIO
«» ' l*fl 0 SO 100 IU 100 IU 100 MO MO MO UO
HUM (f,f,|.l 0 U 100 IU no 1U M IU tOO 4U too
0 I 1 I < I I t I t i.
0.0
mil i o*-it-n i wwzi
til* I •iMlIf - tl»7iM
3
1.1
. . . . I ....
-10 t 10 IS 10 IS M 1S «0 US
TIM lOUIlnt 0 IlltSITOflO
attir ffj o u 100 iu m IM MI iu too tu wo
oiow ip..r..n..i o u loa iu in iu MO no too ltl0
«« ' >.fl.... * 90 IM IM 200 210 100 ISO 400 ISO (00
OIOM (r.r.^.i o JO IM iu 200 iu MO iu too tso wo
I 1 1 « I < > 0 • 10
0.0
OIOM ic.%i!v(si o
•m IIQ-Vlll 0
IIM> Itl -S
tVM ICM/»/ll 0
OM r fti.... o
OIOM V.f..*..\ 0
0
I
U
U
I
S
1
100
10
t
IU
IU
100
too
20
i
150
IU
100
wo
IU
IU
IS
0
100
100
i
10
I
«u
MO
II
10
UO
UO
10
Figure 3-10. The bold solid line is bscat. At 5 a.m. there are several haze layers, with
bscat 4-5. Spiral #2 indicates peak bscat from 1,700-2,200 m. Spirals
and #5 found bscat of 2-2.5 from 600-1,700 m with hazier layer above.
At 10 a.m. (#6) there are distinct well-mixed regimes with bscat of 3 (to
400 m), 2 (500-1,400 m). aadl -2 (1,500-2,000 m).
-------
0.0
MIt i M-IO-1S I MMIt
II«C I UlMlM • II.J7.J1
I
HUT ill -10 I 10 19 JO IS Ji M «0 M
TIM c^BJ/J/510 | 1 ) « t Itl • tlO
ou' ftj.... o so 100 IM MO no MO a* «» «• MO
UOM if.MO o H lOoiMMonoiooiMWotuMO
On, ltf*H» 0 I 1 I « 1 I 7 I t 10
0.0
MFC i at-io-n i »n»it
I IK i IIiUill - IIiliM
8
tn» ££i -so s 10 is jo n M n «o
-------
flUGUST 10, 1975
2.S
2.0
l.S
a i.a
13
0.5
0.0
I I
1
I I
BKIIT nQ-VM) 01 23H5678910
Rgure 3-12. Overiaid soundings from August 10, 1975. Even at midday, the soundings
in the Illinois-Indiana region show marked layers of haze with peak bscat
aloft.
282
-------
4.0 JULY 30, 1976
4.1 Flight Description
A large high-pressure cell with related stagnant air was predicted to
occur in the region to the east and north of south-central Illinois. Also,
during the night preceding the sampling missions, winds from the Chicago
region were expected to deliver an accumulation of pollution into the region
along the Illinois-Indiana border. Consequently, a "long-range" flight
aimed at characterization of a large stagnant air mass along the Illinois-
Indiana border was planned.
The flight plan called for an approach to the target area from the
south. Thus, the first leg of the mission consisted of a flight from St.
Louis south along the Mississippi River to Paducah, Kentucky, then east to
Bowling Green, Kentucky. In the second leg, which was expected to be the
principal component of the sampling mission, the flight consisted of zigzag-
ging northward, up along the Illinois-Indiana border (also Wabash River
Valley). The flight was broken up into an alternating ferry-dolphin-ferry...
sequence in order to obtain horizontal gradients going into the stagnant air
mass and to resolve the structure of the atmosphere. The northernmost point
reached was north of Danville, Illinois. A strong gradient of visual range
was observed along the Illinois-Indiana border. A sharp increase in haziness
also existed across central Illinois. The predicted meteorology did prevail,
with nearly stagnant flow conditions in the Wabash River Valley. Several
power plant plumes were encountered along the flight path.
Figure 4-1 shows a schematic of the ferry-dolphin flight pattern used
on the July 30 and August 2, 1976, missions. The ferry is a conventional
traverse at constant altitude and provides purely horizontal information.
The dolphin pattern of ascent-descent-ascent is designed to measure the
vertical structure within and above the mixed layer. It combines the hori-
zontal progress of the ferry with the altitude variation of an aircraft
spiral.
In an operational sense, an attempt is made to schedule the dolphin
segments of the flight at locations where there is reason to expect several
layers of pollutants. As seen in the illustration, a ferry at a poorly
chosen altitude could provide a misleading measurement. Dolphin measure-
283
-------
ments, on the other hand, can be difficult to interpret because there is no
natural axis (horizontal or vertical) to the pattern. This approach, ideally,
can provide vertical resolution over a large horizontal extent.
4.2 Summary
For the July 30, 1976, flight the ferry-dolphin sampling technique
provided a large number of vertical soundings over a range of about 50 km.
A strong gradient of b t, sulfate, and ozone was seen along the
Illinois-Indiana border with higher concentrations in the northern loop of
the flight. The atmosphere became layered by the evening hours, with strong
plume effects in central Illinois.
FERRY
.••.•:•••.. .
. • •.";.•'.•.:•: •
•7 /
//////
.
t rtftftHS If/ fff tfff
GROUND
Figure 4-1. Ferry-dolphin flight pattern schematic for July 30 and August 2, 1976.
284
-------
(VJ
00
tn
JUL 29, 1976
RH CORRECTED BEXT
SURFACE MIND FIEL
t I I f . ,
t t f » „ ~
» » > ...-x--*.....»..«.
t f t t
NOON TEMPERATURE
JEMP
DEC C
•Hi 20-211 m 28-32
= 24-28 • >32
SEfl LEVEL: PRESSURE
PRES
o. as. so. 75. 100.
JON- tote
1010- 1018
1018- 1020
> IKO
NOON DEWPOINT
.DENP
RELfiTIVE HUMIDITY
;;;; UO-60
= 60-80
80-90
>90
Figure 4-2. July 29, 1976—winds were stagnant over Missouri and Illinois the day before the flight.
-------
ro
oo
JUL 30, 1976
RH CORRECTED BEXT
.BEXT
,2.
lOMM-y/M
!•;• 3-4 m 6-8
= «l-6 • >8
SURFACE WIND FIEL
NOON TEMPERATURE
SEfl LEVEL PRESSURf
.PRES
1013.1
a. ss. so. 75. too.
MB
:::: 1014- 1016 ^ 10IB- 1020
— loio- loie • > IKO
NOON DEHPOINT
DEHP
RELflTIVE HUMIOITT
RH
40-60
60-80
80-90
>90
Rgure 4-3. July 30, 1976—continued stagnation over the central midwest, with the region of lowest
visibility in Indiana. Noon surface temperatures along the path were near 28° C, with dewpoint
20° -24° C. A high-pressure system is moving in from the Northern Great Plains.
-------
00
JUL 31, 1976
RH CORRECTED BEXT
SURFACE MIND FIELI
WIND
NOON TEMPERATURE
.TEMP
DEC. C
;;;; 20-2H m 28-32
= 24-28 • >32
SEfl LEVEL;PRESSURE
.PRES
1013.1
0. Z5. SO. 75. 100.
NOON DENPOINT
.DENP
::;; 12-16 m 20-24
= 16-20 • >24
RELATIVE HUMIDITY
RH
U.I
;;;; 40-60
= 60-80
80-90
>90
Figure 4-4. July 31. 1976—the high-pressure system induced a major change in the wind field,
pushing the hazy air mass toward the east coast.
-------
TABLE 4-1. FLIGHT OUTLINE, JULY 30,
(Cartridge No. 021)
1976
Event
No.- Type
1. Dolphin
2. Ferry
3. Dolphin
4. Ferry
5. Dolphin
6. Ferry
7. Dolphin
8. Ferry
Landing/T.O.
9. Ferry
10. Dolphin
11. Ferry
12. Dolphin
13. Ferry
14. Dolphin
15. Ferry
16..' Dolphin
17. Ferry
Landing/T.O.
18. Ferry
19. Dolphin
20. Ferry
21. Dolphin
22. Ferry
23. Dolphin
Landing
ALT
(m MSL)
450-1650
600
450-1800
600
600-1650
1050
450-1500
1050
1050
540-1500
900
480-1500
600
600-1650
600
640-1600
600
-
600
480-1800
G
600
600-1500
600
600-2300
-
Route
B - C - D
D - E - F
F - G - H - I
I - J
J - K - L - M
M - N - 0
P - Q - R
R - S - T
Evansville, Ind.
U - V - W
W - X - Y
Y - Z - A1
A'-B'-C'-D'
E1 - F1 - G1
G1 - H1 - I1
I1 - J' - K1
K' - L' - M'
M' - N1 - 0'
Terre Haute
pi - Qi - Rl
R1 - S1 - T1
T1 - U1
U1 - V - W
W - X1 - Y1
Y1 - Z1 - A1
Spirit
Time
1250 -
1310 -
1330 -
1351 -
1402 -
1416 -
1431 -
1443 -
1457 -
1637 -
1651 -
1701 -
1712 -
1722 -
1733 -
1746 -
1759 -
1813 -
1828 /
1923 -
1931 -
1947 -
1958 -
2013 -
2024 -
2052
(CDT)
1308:30
1330
1350
1401
1416
1430
1443
1456
1632
1649
1659
1711
1721
1732
1745
1758
1812
1820
1915
1930
1946
1957
2012
2022
2042
288
-------
AIRCRAFT SAMPLING
ROUTES, July 30, 1976
Cartridge No. 021
Rgure 4-5. Map showing flight path and location of spirals for July 30, 1976. The
numbers 1 through 27 indicate the positions of the spirals displayed in
Rgures 4-18 through 4-22.
289
-------
1976
JULY 30,
Rgure 4-6. During the flight several SO2 plumes were encountered, especially along
the southern leg. Winds are light, predominantly from the southwest. Note
that concentration is displayed perpendicular to the aircraft path and does
not represent prevailing wind directions. Note: The systematic decrease-
increase in concentration is due to the dolphin pattern of flight.
290
-------
JULY 30, 1976
•OWll NO
GttEM
Rgure 4-7. Continuation of SO2 concentrations. Note: The systematic decrease-
increase in concentration is due to the dolphin pattern of flight.
291
-------
JULY 30, 1976
OZONE
Rgure 4-8. Ozone concentrations were somewhat higher along the northern loop of the
flight. Note: The systematic decrease-increase in concentration is due to
the dolphin pattern of flight.
292
-------
JULY 30, 1976
OZONE
•OWIINC
GCIIM
Rgure 4-9. Continuation of ozone concentration. Note: The systematic decrease-
increase in concentration is due to the dolphin pattern of flight.
293
-------
JULY 30,1976
Figure 4-10. Sulfate concentration was much higher along the northern loop of the
flight. Note: The systematic decrease-increase in concentration is due to
the dolphin pattern of flight.
294
-------
JULY 30, 19-76
S04
•OWLI NO
can*
Figure 4-11. Continuation of sulfate concentrations. Note: The systematic decrease-
increase in concentration is due to the dolphin pattern of flight.
295
-------
JULY 30,1976
CMC
Figure 4-12. Condensation nuclei count. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
296
-------
JULY 30, 1976
CMC
•OWUNG
ORIIN
Figure 4-13. Continuation of condensation nuclei count. Note: The systematic
decrease-increase in concentration is due to the dolphin pattern of flight.
297
-------
JULY 30, 1976
CHARGE
Figure 4-14. Aerosol charge reveals the contribution of major urban areas as well as
power plant plumes. Wide charge plumes are seen along the northern
loop and in central Illinois. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
298
-------
JULY 30, 1976
CHARGE
•own NO
CKUM
Figure 4-15. Continuation of aerosol charge. Note: The systematic decrease-
increase in concentration is due to the dolphin pattern of flight.
299
-------
JULY 30, 1976
BSCAT
Figure 4-16. Haziness, as measured by bscat, was also much higher on the northern
Illinois-Indiana border region than farther south, with an additional peak
in central Illinois shown on the following page. Note: The systematic
decrease-increase in concentration is due to the dolphin pattern of flight.
300
-------
JULY 30. 1976
B
'SCAT
CREIN
Figure 4-17. Continuation of bscat concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
301
-------
.0.1
MIC I OT-JO-lt7« i
IIXC i liiUt* - H
I
sm « ii.it :
II
•M iif/m 01 l i < t I )l
M, If.f.l.l II I t IJ It II II » U M
t* Wf/lH 0 t 10 It M M M It HI W M
l" 0 M M M M IM IM IM IM IM Ml
MIC i •»-»•-11* i *Mtti
I IK i IIiMiif • IJillil
o.l
I*
IBM llir'/m I I
Mt If.P.I.I I « I II II M M II U M U
U I'V.'fl It II II M It M It M «t tO
dMMiffU... I II M U M IM IM IU IM IM IM
,.0
MIC I •J-JO-II7* i IIMMI
I IK i lliMill -
L , SOU = 6.7
•«• ll«>OIO
SO, |».P.|.I • <| • 12 !• 20 2« 21 32 )• «0
CD no.'/"'! It 10 It M » M n 10 «t tO
. I M •• M M IM IM IM IM 110 100
* I.I
" MIC i 07-jo-iiTi i nnoai
IIMC i Hit.I - KilliO
SOU
•a tar'na 01 1 I « t I 7 I III
M, ir.f.i.i at « i I u 11 t« ii ii 10
<• U&tfl 0 t 10 II M It M I* «• M U
... 0 M M M M IM IM IM • IM IM MO
0.1
MIC I 07-U-IOTt i HTOOai
UX i Ilillil* - IlilSill
I
Ml If.f.l.l OIII«SI7II||
" UVtftl • < 10 It II 2S M M M M M
CMMIIIRI.... I » 10 U 10 100 120 l«0 IM IM IN
Figure 4-18. The first two dolphins encountered two elevated plumes. The third,
along the southernmost leg of the flight, indicates a rather well-mixed
layer up to 1 km with bscat of 2. Dolphins 4, 5, and 6 were taken in the
vicinity of Bowling Green, with bscat 3-3.5.
302
-------
MIC i OT-U-IITt i MMMI
Ill* i IlitttU - Hi Mi W
1
IIO-VM 01 lll-ll ) I I 10
if.f.tJ 01 I I I 10 II II II II M
Wt/fl •• III«M»MHMMM
«>.... » >* M U M IM IM IM IM IM Ml
KR
r
«•
0.0
MIC i 0»->0-ll»l i lllOil
UK i MiUilO - lliMiU
8
SOU • 6.5
•*« nirVm 0121111)11 10
MI If.f.l.l It 1 I I II It !• 10 10 M
en Wtntl II ii it M n M H M u M
UMUWI 0 II 10 M M IM IM IM IM IM 2M
1.0
i l.l
X
1 '•'
»*
5*
O.I
•MM uyvm
M, if.f.ki
>.i
1.0
C 1 «
K I.I
§ '••
^
.J
a
O.I
0.0
U, If.f.l.l 1
" IWVSfl 1
CWWGI mi. ... 1
i MIC i 07-io-itn i tntai ~
IINC i UiM.IJ - lliUlll .
9 j
- ' -
1 H-£ ~^_ -
'. i-Sj-r "^^I :
- 1 -.. 7 -
' . . . . 1 . . . . '
01 it ii 10 a n n 10 11 so
0 20 M U M IM IM IM IM IM »
i MIC I 07-M-ltM i MM1I )
IINC i 11.11. <0 - II. Mi 91 !
11 J
-
1 , ,.^ -
;';;:: f^__ :
]/?%^ '
i it ii » n M n M M u
20 M M M IM IM IM IM IM 101
I.I
1.0
« I.I
1 '••
i-
i
O.I
0.0
o '.n''»*i
* inn* u(»i <
1.1
1.0
K '
K
S 1.0
O.I
oo
liu. no^/m
SO, If.f.l.l
" MttH
> CWWSI ««....
MIC i o>-M-itH i miaii '.
'. IIMC i II.II.M - IliHiM .
10 .
-
H: -rr<%
: 1 '""r;\ ^
- %:.Vv ' B:
: son n e.i.
S 10 IS 20 2S M IS 10 11 SO
20 10 M M 100 120 110 IM IM 200
MIC i OJ-M-IIJI i mecai '
UK i II.SO.2I - lliU.II .
L ' 12 _
*-^^ "^— — _"
L^ "^- ~~~1
.
— —
i ! SOU = 8.9 .
01 10 II 21 21 10 II 10 11 SO
0 20 10 10 10 100 120 110 110 IM 20
Figure 4-19. Dolphins 7-9 occurred along a line heading northwest from Bowling
Green, with 10 and 11 (separated by 2 hr) at Evansville, Indiana. Within
the 1,250-m mixed layer the characteristic levels of bscat increased from
2.5 to 3.0 along this path.
303
-------
MIC i o7-*o-it7i i «7*«i
13
9m na-»A'i' o
, SOU » 8.9
Ui If.f.l.l Illlllllllll
<• U*?/M 01 II II M M JO H M « I*
.... 0 M M U iM IM IM IM IH IM M*
MIC I 07-10-1*1* i N7M1I
(IK i HiUiM -
14
llfl-Vm • I J I « t I 1 I t
If.f.l.l Illlltllti,,
lUVtt II 10 II M M U M «• «t M
IW.... I M « U M. 10* IM IM IM IM Mt
* I.I
l.l
wic i 07-jo-1 in
III* I IJ.II.17 - I7.U.S7
15
,
I
SOU 3 11.7
M! If.f.l.l 01 » I I II II l« II II M
" wf/rn oi ii ii M n 10 ii M it M
»'.... 0 a M M M IM IM IM IM IM Ml
0.0
MK i I7-JO-II7I i H7KQI
IIW I 17.H.I • I7i»i*
16
. SOU » II.7
HO" 'HI OII1H1I7IIII
If.f.l.l 1-1 1 1III7IIII
i < 10 ii M n M n « « M
>.... 0 » 10 U M IM IM ' IM IM IM Ml
MIC I 07-10-1*7* i M7MBI
UK I IllUlW - I7||7(M
: . sou • 16.2 :
Id If. P.I. I 01
l.tf/fJl 01 IIIIMnuiS
I M M M
IM IM IM IM Ml
MIC I f?-M-l*7* | U7IO3I
IIW i I7.17.W - I7.«.I7
18
SOU » 16.2
tmm llfrlif OI1I4II1IIII
U. If.f.l.l II1I«II7IIII
II 10 II M » M M M U M
I M « M M IM IM IM III IM Ml
Figure 4-20. Near point X on the map (spirals 13 and 14), light scattering coefficient
is near SOO"4™*1. Heading northwest toward point H', bscat is ap-
proximately 4 at spiral 16, increasing to 5 at spiral 17. The northeast
path of spiral 18 indicates bscat of 6 near the surface at 1730 CDT.
304
-------
1.1
MI
lll-«/m 0
MIC i 17-jo-1 in i HTM*)
MM i l7.Uill - Itilil*
19
, SOU » 17.5
« IWVtfl « « '• '•
. 0 H «0 U
t t i • i ii
ID II II II II M
» II If M »« H .
IM IM IM IM IM tM
0*IC i OT-X>-lt7l i H74OJI
flM i ItlllM - III 111*
n M w «o M
IM IN IM IM IN
1.0
MTC i OT-9O-IIN i UMOII
IIMt I lllMiW • HilTiU
21
IV \
1
SOU = 22.3 .
oi 2 i i i i i i • 10
Id If.f.l.l 0 I i 1 1 i | 7 • I 10
0 I II II MM M M M II U
I.... 0 20 M M M. 100 IM IM II* III 200
0.0
Mfr I 07-SO-IITI i II7U1I
TIM i lliMiM - III Mi >
22 j
SOU = 15.8
0121111)1110
SO, If.f.I.I 01 « I I 10 12 II It It 20
Cl I.l1'/»'l 0 > 10 II 20 K M » HO It to
C«*W M»l.... I 20 10 M M 100 IM IM ItO ItO 200
a. a
-------
0«f i «T-W-lt7« I
IM i XhtiM - MtlllM
25
SOU = 11.9
*u> iif/li" o i a
MI tUJJ II>
c» IWVlft o t ia
CMMClmi.... 0 20 W
t 0 I 0 0 10
I * » 0 0 10
M M M M t% M
IM IM IM IM IM MO
5 '-•
X
I '•'
a
o.(
0.0
onrc i oi-M-ttn
.'•-• IIW i MiUiH - Mi«llS
27
i
SOU n 10.3
> « ( 0 10 II H 10 It M
u WffUl o t 10 i* M as M n it n n
aunt m>.... 0 M M M M IM IM IM IM IM MO
Rgure 4-22. In south-central Illinois (25-26) at 2000 CDT bscat is approximately
3.5 in the lowest 1,200 m.
306
-------
JULY 30, 1976
2.5
2.0
1.5
UJ
a 1.0
0.5
0.0
I I
I I I
I
12:51-m:56
Bic« MQ-VM) 01 23X5678910
figure 4-23. Overiay of soundings 1-10. This leg of the flight ran southeast from St. Louis
to Paducah, east to Bowling Green, and northwest to Evansville.
Highest sulfate and haze were near Bowling Green. Bscat is typically 2-3 in
the southern portion of the flight.
307
-------
JULY 30, 1976
2.5
2.0
1.5
o 1.0
0.5
0.0
I i i
I
I 1
I2s51-ms56
BSMT UQ-VH) 01 2 34 56 7 8 910
Figure 4-24. Overlay of soundings 1-10. This leg of the flight ran southeast
from St Louis to Paducah, east to Bowling Green, and north-
west to Evansville. Highest sulfate and haze were near Bowling
Green. B,.~,t is typically 2-3 in the southern portion of the
flight.
308
-------
JULY 30, 1976
2.5
2.0
1.5
UJ
Q 1.0
0.5
0.0
i i
r r
16^31-18:27
UO-Vm 01 2 3 II 5-6 7 8 910
figure 4-25. Overlay of soundings 11-21. In late afternoon the IL-IN border region has
highest bscat of 5-6 along the northern loop, with sulfate concentration
15-20/ig/m3.
309
-------
JULY 30, 1976
2.5
2.0
1.8
Q 1.0
0.5
0.0
i i
19:30-20:41
I
B»c« IIO-VH) 01 2 3 M 56 7 88 10
Rgure 4-26. Overlay of soundings 22-27. Across central Illinois, about 2000 CDT, several plumes
of haze are encountered, with bscat 3-4 near the surface.
.310
-------
5.0 AUGUST 2, 1976
5.1 Flight Description
Once again, the objective was to collect data over a long-range flight.
Under predicted northerly wind conditions, the decision was to make a box
flight enclosing the Ohio River Valley approximately between Louisville and
Paducah, Kentucky.
In the first leg of the mission (1100-1300), an eastward flight from
St. Louis ended in Louisville, Kentucky; the route included the northern
border of the "box." The second flight traveled down the eastern edge of
the box, then proceeded east almost to the Virginia border. Three traverses
were then made along the south edge of the box, being close to and parallel
to the Kentucky-Tennessee border. The return to St. Louis completed the
western edge of the box. Several power plant plumes from the Ohio River
Valley and from central Kentucky were picked up in the east-west traverses
along the southern leg.
The entire flight lasted about 10 hr, and was carried out by the Wash-
ington University aircraft alone. The entire mission was carried out in an
alternating ferry-dolphin-ferry...sequence.
5.2 Summary
The August 2, 1976, sampling mission consisted of a box flight enclos-
ing the lower Ohio River Valley. The region was reasonably clean, with peak
-4 -1
b . of approximately 2 x 10 m and maximum sulfate concentration of
approximately 8 ug/m3.
The horizontal traverses indicated a north-south gradient with increased
haze along the Kentucky-Tennessee border area. From St. Louis, Missouri, to
Louisville, Kentucky, the levels of b . were near 1, with little variation
either horizontally or vertically. The southern leg encountered higher
levels of all aerosol parameters, with a more pronounced vertical structure,
indicating the effect of several urban and point sources. Thus, this mission
yields information about the isolated contribution of emissions within the
box.
311
-------
RUG 01, 1976
RH CORRECTED BEXT
BEXT
10MM-H/M
:::: 3_q m Q-Q
= 4-6 • >8
NOON TEMPERflTURE
JEMP.
DEC C
I;!: 20-24 H 28-32
= 24-28 • >32
NOON DENP01NT
DEWP
DEC C
::.: 12-16 M 20-24
= 16-20 • >24
CO'
SURFflCE MIND FIEL
SEfl LEVEL? PRESSUR
0. 25. SO. 75. 100.
iota- toie
1010- 1010
1010- 1020
> 1020
RELflTIVE HUMIDITY
RH
iiii 40-60
= 60-80
Figure 5-1. August 1. 1976—a large hazy air mass was centered south and east off St. Louis, Missouri.
Temperature and dewpoint were relatively low, whh northerly winds.
-------
CO
I—"
oo
RUG 02, 1976
RH CORRECTED BEXT
BEXT
SURFACE MIND FIEL
0. 23. SO. IS. 100.
NOON TEMPERRTURE
;;;; 20-24 m 28-32
= 232
SEfl LEVEL:PRESSUR
:::: 1014-1016 Hi 1018- 1020
= ioi»- 1010 • > torn
NOON DEHPOINT
DENP
DEC C
... 12-16 m 20-24
= 16-20 • >24
RELATIVE HUMIDITY
;;;; 40-60
= 60-80
80-90
>90
Figure 5-2. August 2. 1976-the region of haze moved south of the flight pattern by August 2nd.
The Ohio River Valley region had good visibility with low RH. Winds are out of the
north to the northeast.
-------
CO
RUG 03, 1976
RH CORRECTED BEXT
BEXT
10MM-4/M
:::: 3_4 m 6_Q
= 4-6 • >8
JRFRCE HIND FIEL
NOON TEMPERATURE
JEMP
25.1
•iii 20-24
= 24-28
28-32
>32
SEfl LEVEL PRESSUR
o. as. so. 79. 100.
ION- toie
101ft- 1010
1018- 1020
> 1020
NOON DEHPOINT
.DEWP
;;•: 12-16 m 20-24
= 16-20 • >24
RELATIVE HUMIDITY
;;•; 40-60
= 60-80
80-90
>90
Figure 5-3. August 3, 1976—a large clean cool dry high-pressure system covers the entire
eastern United States on the day after this flight.
-------
TABLE 5-1. FLIGHT OUTLINE, AUGUST 2,
(Cartridge No. 022)
1976
Event
No.- Type
1. Ferry
2. Dolphin
3. Ferry
4. Dolphin
5. Ferry
6. Dolphin
7. Ferry
Landing/T.O.
8. Ferry
9. Dolphin
10. Ferry
11. Dolphin
12. Ferry
13. Dolphin
14. Ferry
15. Dolphin
16. Ferry
17. Dolphin
18. Ferry
Landing/T.O.
19. Dolphin
20. Ferry
21. Dolphin
22. Ferry
23. Ferry
Landing
ALT
(m MSL)
690
420-1200
540
420-1350
570
450-1410
810
-
600
640-1890
600
600-1800
540
600-1650
600
450-2220
600
450-1950
750
-
390-2400
690
450-2160
600
810
-
Route
A - B
B - C
D - E
E - F
F - G
G - H
H - I - J - K
Louisville, Ky.
M - N
N - 0
0 - P - Q
Q - R - S - T
T - U - V - W
W - X
X - Y
Y - Z - A'-B'
B1 - C1
C1 - D1
D1 - E1
Bowling Green
F'-G'-H'-I1
I1 - C1
C'-J'-K'-L'
L'-M'-N'-O1
p. . Q.
Spirit
Time
1111 -
1128 -
1141 -
1151 -
1209 -
1222 -
1242 -
1310 /
1443 -
1502 -
1518 -
1532 -
1551 -
1606 -
1622 -
1631 -
1653 -
1708 -
1727 -
1751 /
1841 -
1910 -
1938 -
2005 -
2031 -
2056
(COT)
1128
1139
1151
1207
1221
1241
1303
1420
1501
1518
1531
1549
1606
1620
1629
1651
1707
1727
1747
1838
1909
1937
2004
2028
2047
315
-------
to
M
CT)
LOCATION OF AIRCRAFT SAMPLING ROUTES, August 2, 1976
Cartridge No.: 022
Figure 5-4. Map showing flight path and location of spirals for August 2. 1976. The numbers 1
through 30 indicate the positions of the spirals displayed in Figures 5-20 through 5-24.
-------
flUGUST 2, 1976
oo
Figure 5-5. SO2 concentration is generally higher along the southern edge of the box. with numerous
individual plumes encountered near St. Louis and by the Illinois-Indiana border. Note: The
systematic decrease-increase in concentration is due to the dolphin pattern of flight.
-------
RUGUSiT 2, 1976
co
M
00
Figure 5-6. Continuation of SO2 concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
oo
flUGUST 2, 1976
SO2
Figure 5-7. Continuation of SO2 concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
oo
N>
O
RUGUST 2,
SQ,
1976
Figure 5-8. Sulfate concentrations are low over the entire area, with peak of about 8 /*g/m3 in the vicinity
of Bowling Green. Note: The systematic decrease-increase in concentration is due to the
dolphin pattern of flight.
-------
flUGUST 2, 1976
co
Figure 5-9. Continuation of sulfate concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
GO
INS
NS
flUGUST 2, 1976
SQ,
Figure 5-10. Continuation of sutfate concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
GO
ro
co
RUGUST 2, 1976
CMC
Figure 5-11. Condensation nuclei concentration is highest in central Illinois; apparently the southern region
contains relatively more aged aerosol. Note: The systematic decrease-increase in concentration
is due to the dolphin pattern of flight.
-------
co
IV3
RUGUST 2,
CMC
1976
Figure 5-12. Continuation off condensation nuclei. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
co
ro
01
RUGUST 2, 1976
CMC
Figure 5-13. Continuation of condensation nuclei. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
nUGUST 2, 1976
CHARGE
Figure 5-14. Aerosol charge also peaks along the stretch between Paducah and Bowling Green. Note: The
systematic decrease-increase in concentration is due to the dolphin pattern of flight.
-------
IVJ
HUGUST 2, 1976
CHARGE
Figure 5-15. Continuation of aerosol charge. Note: The systematic decrease-increase in concentration is due
to the dolphin pattern of flight.
-------
co
ro
00
flUGUST 2, 1976
CHARGE
Figure 5-16. Continuation of aerosol charge, fyote: The systematic decrease-increase in concentration is due
to the dolphin pattern of flight.
-------
co
fSi
flUGUST 2, 1976
B
'SCAT
Figure 5-17. Bscat is much higher along the southern leg of the flight. This was a relatively clean day,
especially in central Illinois, with bscat typically less than 1 »10-4m-1. Note: The systematic
decrease-increase in concentration is due to the dolphin pattern of flight.
-------
flUGUST 2, 1976
CO'
00
O
B,
SCAT
Figure 5-18. Continuation of bscat concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
flUGUST 2, 1976
CO
CO
B,
SCAT
Figure 5-19. Continuation of bscat concentrations. Note: The systematic decrease-increase in
concentration is due to the dolphin pattern of flight.
-------
0.0
Mt( i M-M-IBTt i Mltttt
TDK i Mill I* ' HilliM
0 I 2 > 1 » • 7 « I 10
SO, If.f.1.1 01 I I 11 II II 11 I* 17 M
OS 10 U 10 IS M it M M 10
0 M M M M IM IN ia IM IM JM
0.0
Mff • M-M-ltTt i M7KU1
flNf i HlJlK* - llilli«0
2.5 :
•KM UP**/*! Oll)i|SI7llio
SO, If.ri.l OI1)«S|7II 10
01 10 II » IS M II W> «I SO
0 20 10 U M IM 120 IM IM IM IM
.0.0
Mff i M- if.n.i 01 !>«siiiiia
«« U4VM 01 II IS M a M IS HI M M
CHMCII.HYI.... 0 20 M M M IM IM IM IM IM 2M
1.1
0.0
M» i M-M-KT* i tatatt ,
flK i lliUil? - IIilill
I
sou • 3.5 :
ha, no^/iii OI11III7IIIO
JO, If.f.1.1 OI11VSI7IIIO
« UV/!>fl 01 10 II M II M IS «0 »» SO
enact (i|»>. 0 10 M M M IM IM IW IM IM 100
Mfl i M-OJ-II7I | UTtOIl
TIM i lliUill - lliMi*
*KH no-*'"' oil)
u, if.f.i.i 0111
« i.iV.'*^. o < 10 ii
I««.... I M M M
1.1
0.0
1 I I 7 I I 10
I I I 7 I I 10
M » M IS «0 HI SO
M IM IM IW IM IM IM
Mt« I W-01-II7S i H7SUI
II* i ItilliK • Iliilil
6
i
sou « 1.7 :
01 1I4SI70IIO
SO, lf.f.0.1 01 111SI7IIIO
U ll.V.'f* 01 10 II 20 29 90 II 10 •» U
CIMMfllw 0 10 M M M 100 110 IKO l«0 IM 200
Rgure 5-20. Along the line from St. Louis to Louisville, near noon, bscat is relatively
uniform with respect to height and is below 1 «10-%i-1.
332
-------
1.0
i '•«
z
1 ...
_J
(I
O.S
0.0
CMMCf mi. . . .
I.Q
5 '••
a
o i.o
E
O.J
0.0
•m iio^/m
" IWVW
CMMI mi. . . .
1.0
«• l.k
ALTITUDE
B —
M B
CJI JJ Q*/N*l
e«*« M« . . .
• • i . 1 1 1 1 1
OflfC i :
L s .,-'r .*•.•::;:;-
i '''•• S ••'
.' SS . i . . . . :
) 1 10 II M M M M U HI M
1 10 U M M IM IM 140 IM IM 10
) MTt I M-«-l«I* 1 «««»
IIM i I5ll.ll - IfltllJ
11
;U -••-.
: '? '"; ^
I I f < j
SOU - 6.8
1 ....
01 II II M M M H U
-------
0.0
OHM i M-OJ-II7I | MTMH
IINC i ISiMi> -
13
I SOU » lt.3 -
O.I
0.0
MIC i M-Oi-IITI i HltMJ
M«t i ISiWi) - I Si«li}l
14
sou • 4.3 :
iii-vvr o . 2 i i « • > i i 10 •— UCtfl « i » i i i i i i i. 11
M. litjj o i J > * i • i • • 10 «* ci^J" o i 2 » « 5 i T i i 10
U MH»/.'t 0 « 14 IS » n M II M « M " '.«/!* » » II II M a H M II It U
CMMIIMI ' o 20 w «o M 100 120 n« IM IM no "««« we. . . . o i« u MM
121 i«o lu IM KM
0.0
MIC I Ot-tt-117* i MTMHt
»l« i lliliM -
15
SOU = li.S ]
0.0
o*rc i oi-oi-iin i H7MU2
tIK i IlilSiia -
16
SOU = U.5 !
IJ5156>tlll> In, IIO-«/m 0 I 2 ] < 5 6 7 | 1
SQ» tf.f.S.t 0 I 2 1 II S 6 T •• 10 Sit 1^. P.B.I OIZ}«%67t«|0
c» i.i1V.»'l o i to is 20 a 10 is «o is to ci '.Ml'/."?', o s 10 i: ;o a 30 is »o «s so
CHHBCI.H... 0 20 M M M 100 IM 110 IM IM 2M CMMU ffil. . . . 0 20 10 (0 M 100 I JO l«0 160 110 200
0.0
MIC i OI-M-II76 I H76022
IIHC i ll.29i«> - III Mill
17
7.2 I
OAK i 01-02-1171 i tntOll
MMC i llilliU - Ilitlill
SOU = 7.2 ;
0.0
I21«St7IIIO •«« I IO<»1II 0 I 2 3 1 S e » I I
JO, If.P.I.I 0 I > 1 1 S I 7 I I It SO, If.P.i.l 01 e t 12 IS II 21 2« 21 10
CH UtV"'! OS 10 II 2* 21 M H M M M CH I.IO.'/fl IS 10 II 20 2S 10 IS 10 IS SO
CMWI ll|>l.... 0 K «0 M M IM 121 IM IM IM Ml CIWW«« . .. I M 40 M M 100 120 110 ItO in 200
Rgure 5-22. On the eastern corner of the sampling path there is a well-mixed 1,600-m
layer with b^cat = 1- Spiral 17, near Bowling Green, exhibits bscat of 1.5.
Spiral 18 indicates an SO2 plume with associated bscat peak of 2.5.
334
-------
1.1
MIC
IINC i llitl.lt .
19
SOU « 7.2 i
0.0
MX i M-OI-lin i M7IO2J
tlW i I7i«ill . ITilOit
20
,
sou « u.u :
10
tm llO-«/m 012111171111 ••« IIO-VIll 0 I 1 I 1 S I 7 I
so. if.f.i.i o 2 « I I 10 12 11 if u 21 MI ir.r.i.i oi2j*si7ooio
« '.Wf./lft 0 I 10 II M II M IS 10 1» M CI IU*/**1 01 10 II 10 IS 10 n « 0$ M
CWMCI H»l.... 0 M M U 00 100 IN IU 101 IU Ml CKWC1 URI.... 0 20 10 U U IU 110 IU IU IU IU
1.1
0.0
Mil i M-M-ltM i tttttt
IIMt i I7.l0iil - I7H7.U
SOU > 6.5 i
1.1
On na-»/m 0 I t 1 t
M. ir.r.t.i oijll
« wean • • » » «•
CMWU M«l.... 0 2* . U U U
I I 1 I I 10
I I 7 0 I 10
11 10 M «0 if U
IU IM IU IU IU 2U
,.
o.s
.0.0
Ml( I M-M-IITO i M7IOU
HM i I7illil - !7iMl»l
sou = 6.5 :
OMI Mai/in o i
Ut if.P.O.i o i
en l.l.0//»?l 0 I
!«•.... 0 M
20 it 10 II M
U IU 120 IU 110
0.0
Ont no-Vlll
io> ir.r.t.i
Ml(
Ml* i I7i«7.1 . !7iSOll«
23
•••
'o.o
Mil I B*-OI-lt7t i U7IOM
TIM i IO.J7.JI - lOltllU
24
J • « || On IIITVIII OI2119I10IIO
T 0 I 10 ' ««t I'.f.t.l 111111(701 10
iiq'/rt 01 10 it 20 n to M u «l u u lUV.ufi o i 10 n 20 21 M n u « u
CMMUimi.... 0 20 10 U U 100 120 IU IU IU 200 CHMH ml.... 0 20 10 U M 100 120 110 IU IU 2U
Rgure 5-23. South and east of Paducah (spirals 20-22) the light scattering coefficient
has again dropped to 1 •10-4m-l with little vertical structure. Near Bowl-
ing Green (23 and 24), bscat is measured at 1.5-2.
335
-------
J.I
I ..
0.0
MTI i ooai-itn
TIM i I lilt i IS - HiMiU
25
SOU " 7.1
0.0
Mil i M-M-IOT* i «7M»
tine i ItiliM • I0i»il»
26
7.1
I*. no-v'iii' o i i i « i i i o < 10 •«« !U2i!l o i i j « i o i o i 10
M, li.PJJ 0 I 1 1 « t I 1 0 t 10 «* KiMJ 0 I I I « S 0 7 0 | |0
» lltf/fl 01 10 IS M IS M 3S M ««io
SOi If.f.O.I OI11«>OTOOIO *•» »•'•».< OIJ1»»«>0«IO
U WtftH 0 J 10 H M Jl JO M M M SO « '.'.V.'ffl « « 10 U M « JO l« «0 « M
CMMCtllRI.... 0 M U M M tOO IM IU IM IM MO CMMt ll|«l.... 0 » M M 00 100 l» IM 100 10* IQt
0.0
Oil)
u» o.f.i.i 0111
'» UV/Jfl o t i* i*
CHMCf (IRI.... 0 M «0 M
SOU a 3.4 :
< s > i o « 10
1 S 0 1 0 t 10
10 J» 10 H
-------
flUGUST 2, 1976
2.5
2.0
1.5
O 1.0
0.5
0.0
i i
I
MQ-VM> 01 2 3 456 1 8 9 10
Rgure 5-25. Overlay of soundings 1-9. Along the northern leg from St. Louis to
Louisville there is little vertical structure around noon; bscat is near 1
within the lowest 1,400 m, with sulfate 2-3 jig/m3.
337
-------
RUGUST 2, 1976
2.5
2.0
1.5
Q 1.0
0.5
0.0
I T
m:28-17:50
UQ-VH) 01 2 3 II 56 78 9 10
Rgure 5-26. Overlay of soundings 10-23. Along the southern leg, 1500-1800 CDT,
bscat is slightly higher at 1 -2, with evidence of several layers of haze and
sulfate 4-7 /ig/m3.
338
-------
RUGUST 2, 1976
2.5
2.0
1.5
UJ
O 1.0
0.5
0.0
18:37-20:49
1
iscni IIO-VH) 01 2 3 «l 56 7
9 10
Rgure 5-27. Overlay of soundings 24-30. Returning north in the early evening bscat
drops from 1.5 to below 1, with sulfate again near 2 /xg/m3.
339
-------
6.0 CONCLUSIONS
Four days' sampling missions from Project MISTT have been presented in
graphical form as an aid in designing future regional scale aircraft experi-
ments. The July 29-30, 1975, mission caught a major sulfate and haze episode
and was designed to provide detailed cross sections of the air mass at a few
discrete distances. The August 10, 1975, flight obtained measurements
across three Midwestern States at a single altitude. A new approach of the
alternate ferry-dolphin sequences was adopted for the July 30, 1976, mission,
which concentrated on characterizing the pollutant gradients along an east-
west and north-south radial. The ferry-dolphin technique was used again for
the August 2, 1976, experiment to "box in" a section of the Ohio River
Valley. -
These attempts to characterize large air masses encountered two major
problems: (1) balancing the need for large horizontal coverage versus the
vertical resolution, and (2) designing methods to conveniently examine the
data generated on a large spatial-temporal scale.
ACKNOWLEDGMENT
This report draws upon the extensive data base of project MISTT. We
wish to acknowledge the participants of the project for preparation of the
data volumes, with special appreciation to Meteorology Research, Inc., for
preparation of the daily weather summaries, flight outlines, and flight maps
for the July 29-30, 1975, and August 10, 1975, flights.
340
-------
GROUND MEASUREMENTS OF REGIONAL HAZE
Kenneth T. Whitby* and Peter H. McMurray
University of Minnesota, Minneapolis, Minnesota
Abstract
This report describes some of the first results from physical and chemical
measurements on the haze measured during the STATE study in Tennessee during
the summer of 1978. These results show that most of the haze aerosol mass can
be accounted for by sulfate in equilibrium with water. There is good agreement
between the size distributions measured by a low pressure impactor (LPI) and
those measured by the electrical aerosol analyzer (EAA). Also, aerosol volumes
computed by assuming that the aerosol is H2S04 in equilibrium with water agree
within 20 percent of the volumes measured by the EAA, and LPI sulfur values
agree within 30 percent of those measured by X-ray fluorescence. These data
also suggest that the aerosol is partly neutralized H2S04 and that the aerosol
is internally mixed.
Findings from a recent analysis of the aerosol formation and growth in
the St. Louis urban plume (Whitby, 1979) are reviewed. Analysis of aerosol
measurements in the St. Louis urban plume on July 18, 1975, showed nuclei
formation rates of up to several hundred cm hr . Nuclei formation rates and
ozone formation rates correlated well with the ground-level ultraviolet inten-
3 -3 -1
sity. Aerosol volume formation rates of up to 6.7 urn cm hr during the
middle of the day were observed. These also appeared to correlate well with
the ultraviolet radiation. Using a simple aerosol growth model based on the
experimental results, the total weight of secondary aerosol formed in the
plume during one irradiation cycle was found to be 1,000 metric tons day .
CHARACTERIZATION OF AEROSOL SULFUR IN PEPE'S
During August 1978, researchers on the University of Minnesota Mobile
Laboratory (UMML)t used several independent but complementary measuring tech-
*Speaker.
tBackground material on the University of Minnesota Mobile laboratories
is presented as an appendix to this report.
341
-------
niques to study chemical and physical properties of haze aerosols in Tennessee.
These measurements included X-ray fluorescence (XRF) of filter samples col-
lected with a dichotomous sampler, measurements of aerosol sulfur distributions
collected with a low pressure impactor (LPI), measurements of the hygroscopic
properties of aerosols with the tandem differential mobility analyzer (TDMA),
and measurements of aerosol physical properties with the electrical aerosol
analyzer (EAA) and optical particle counters (OPC).
The low-pressure cascade impactor used in this study, (design and calibra-
tion described by Hering et al., 1979) had five stages; the top stage had a
lower cut at 0.5 urn, while the lower cut point for the bottom stage was 0.05 urn.
This insured that virtually all of the accumulation mode aerosol mass was col-
lected. Samples were collected on small stainless steel strips and analyzed
by the flash volatilization technique described by Roberts and Fried!ander
(1976). The validity of the impactor results was checked by comparing the
total sulfur mass detected on all five impactor stages with the total sulfur
measured by XRF for samples collected during similar time periods. Results of
this comparison are shown in Figure 1. Note that the two independent tech-
niques agree to within about 30 percent.
If the aerosol is assumed to be sulfuric acid in equilibrium with ambient
humidities, the aerosol volume associated with measured sulfur values can be
calculated. Although the sulfate in the aerosol was probably partially neu-
tralized by NH3, this should still give a reasonable estimate for the volume.
While H2S04-(NH4)2S04 mixtures do exhibit hysteresis properties upon humidifi-
cation and dehumidification, humidities must be reduced to extremely low
values (-5-15 percent) to observe this effect. For summer haze in Tennessee,
where humidities typically vary from 90 percent to 50 percent during the
diurnal cycle, the particles are probably always supersaturated and hygroscopic
properties of H2S04 and H2S04-(NH4)2S04 mixtures will be similar. This hypothe-
sis is supported by Figures 2 and 3, taken from a report by McMurray and Liu
(1978). Note from Figure 3 that hygroscopic properties of H2S04-(NH4)2S04 solutions
do not appear to vary a great deal with composition as humidities are decreased
from 95 percent. With this assumption, total aerosol volumes measured with
the EAA and OPC can be compared with those calculated from impactor sulfur
data. Results of this comparison are shown in Figure 4. Note that impactor
values are, on the average, about 20 percent higher than those measured with
342
-------
the EAA and OPC. This is remarkably good agreement, and supports the hypothesis
that most of the aerosol volume in the haze is associated with sulfur. If
this were not the case, one would expect EAA and OPC volumes to exceed the
volume associated with the aerosol sulfur. Independent measurements to deter-
mine the volume associated with other constituents (e.g., nitrates and organics)
are required for a better understanding of the aerosol.
If the aerosol sulfur distributions are converted to volume distributions
by making the assumption discussed above, volume distributions measured with
the LPI can be compared with those measured with the EAA and OPC. Results of
a typical comparison are shown in Figure 5. Note that the agreement below
0.3 (jm is quite good, but that the impactor detects more volume in the 0.3- to
l-pm size range. A possible explanation for this discrepancy is that the
lower cut point for the top stage of the impactor was 0.5 urn. Therefore, any
coarse particles that penetrated through the sampling system would also be
collected on that stage. For future experiments, stages with larger cut points
will be added to the impactor to test this hypothesis. If this is true, how-
ever, it follows that there is a considerable amount of sulfur associated
with the coarse particle mode as well as the accumulation mode, which is gen-
erally assumed to contain most of the secondary aerosol volume.
It has been shown above that most of the aerosol volume in this continen-
tal haze can be accounted for by sulfur. Measurements of the hygroscopic
properties of these aerosols with the TDMA gave some insight into the compo-
sition of these particles. [The TDMA was described in detail by Liu et al.
(1978). Essentially, the instrument generates a monodisperse aerosol from a
polydisperse aerosol (e.g., from ambient aerosol), manipulates the monodisperse
aerosol (e.g., by humidification), and determines the new particle size after
this manipulation.] Data collected during the night of August 5 and 6, 1978,
are shown in Figure 6. In this figure, growth of the ambient aerosols with
humidity is shown to agree favorably with laboratory data for growth of internal-
ly mixed solutions of H2S04 and (NH4)2S04 in the molar ratio 1:3. Laboratory
data for pure H2S04 or for solutions that contained relatively more (NH4)2S04
did not agree as well. It was concluded that during these measurements, about
25 percent of the H2S04 was neutralized by NH3. The degree of neutralization
343
-------
did vary from day to day, but aerosols were generally found to contain more
H2S04 than (NH4)2S04.
The TDMA data can also be used to show that haze aerosols are internally
mixed. If the haze aerosol consisted of several distinct types of particles,
then it is likely that the hygroscopic properties of these particles would
vary. Thus, if a monodisperse, dry aerosol were humidified (as is done in the
TDMA), particles of several sizes would be expected. This was never observed.
Humidification of a monodisperse aerosol always resulted in another (usually
larger) monodisperse aerosol. This'result was true for all particle sizes
investigated (initial sizes at 10 percent humidity of 0.05, 0.07, 0.1, and
0.2 urn). These results are demonstrated by the data presented in Figure 7.
In conclusion, it appears that most of the haze aerosol volume can be ac-
counted for by the volume associated with measured aerosol sulfur levels, if
it is assumed that the sulfur is in the form of H2S04 in equilibrium with
ambient humidities. Comparisons of volume distributions measured by the EAA
and OPC with sulfur distributions collected with the LPI suggests that there
may be a significant amount of sulfur associated with coarse particles (larger
than 1 urn). Studies of the hygroscopic properties of these haze aerosols sug-
gest that the aerosol consists of partially neutralized H2S04 and that the
particles are internally mixed.
AEROSOL FORMATION IN URBAN PLUMES
Recently, Whitby (1979) did an analysis of the July 18, 1975, St. Louis
urban plume measurements with the goal of determining the rate of aerosol
formation during a daily irradiation cycle. Some of the more important results
from that analysis and the implications for regional haze studies are presented
here.
The principal conclusions from that analysis are as follows:
1. Nucleation rates exceeding 200 cm sec were found in the St. Louis
urban plume on July 18, 1975. The nucleation rate was found to
correlate with the ultraviolet intensity.
2. Using a simple model in which it was assumed that the accumulation
mode aerosol volume formation rate was proportional to the ultravio-
let intensity, aerosol concentrations similar to those observed
were predicted.
344
-------
3. Using the above model, it was estimated that St. Louis produces
about 1,000 metric tons per day of secondary aerosol in one irradia-
tion day.
4. Our results agree with those of White et al. (1976) and Isaksen et al
(1978) that the entire secondary aerosol can be accounted for by
assuming it is a sulfate such as H2S04.
It will be noted that nothing has been said about conversion rates. The
reason is that, except in the near-source region of plumes, the S02 concentra-
tions and the concentrations of other gaseous precursors are so low that con-
version rate calculations based on gaseous precursors are rather inaccurate.
Rather, the approach that must be taken in urban plumes and in hazy areas is
to determine the rates of formation of aerosol or other chemical species.
Also, it is likely that as the concentration of S02 or other gases reaches low
levels, the rates of aerosol formation are related in a more and more complex
way to; the concentrations of the gaseous precursors. Last summer we made an
attempt to determine the rate of aerosol formation in the hazes of Tennessee.
Although only a small part of the data has been analyzed, one episode in the
Great Smoky Mountains of Tennessee was of such a nature that the growth rate
could be calculated.
McMurray et al. (1978) found that the aerosol volume in the Smokies grew
3 -3
from 4 to 20 urn cm over a 2-day period. At the same time, the b . grew
— T
from 0.3 to about 1.3 m , or about 0.5 b . per day. Husar et al. (1976)
states that the average growth of b .in the most hazy parts of the smog
blobs he has been studying in the eastern United States is about 1 b . per
day. Since the calculated growth rate for the St. Louis urban plume from this
study is 36 urn cm day and 2.2 b . day , it appears that the urban plume
growth rates are about four times that observed by McMurray in the Smokies,
and about twice those in the hazier parts of the smog blob.
Table 1 compares the various measures of aerosol growth for the St. Louis
plume, the eastern haze, and the Smokies. The rate of aerosol sulfur forma-
tion has also been calculated for 70 percent relative humidity. Although com-
paring the aerosol growth rates in the urban plume with those in the general
haze is something like comparing apples and oranges because the urban plume
concentrations will be diluted by overnight travel, the data do suggest the
range that probably exists from plume to haze.
345
-------
Future haze studies should make an effort to determine the rates of for-
mation on successive days of irradiation, travel, etc. The results cited
above suggest that aerosol formation rates are much higher in the urban plume
than on succeeding days after the plume has mixed, aged, and traveled. How-
ever, the data are much too scarce to be sure of what effect we have measured.
ACKNOWLEDGMENT
Most of the work reported here, including the preparation of this paper,
was supported by EPA Research Grant No. R 803851-03. We would also like to
acknowledge the help of Mr. Norman Valentine in the data reduction and of Mr.
Joseph Wolf in the design and construction of the UMML.
REFERENCES
Hering, S. V., S. K. Friedlander, J. J. Collins, and L. W. Richards. 1979.
Design and Evaluation of a New Low Pressure Impactor - II. Atmos.
Environ., in press.
Husar, R. B., N. V. Gillani, J. D. Husar, C. C. Paley, and P. N. Turcu. 1976.
Long-Range Transport of Pollutants Observed through Visibility Contour
Maps, Weather Maps, and Trajectory Analysis. Presented at the 3rd
Symposium on Atmospheric Turbulence, Diffusion and Air Quality, Raleigh,
NC.
Isaksen, I. S. A., E. Hesstvedt, and 0. Hov. 1978. A Chemical Model for
Urban Plumes: Test for Ozone and Particulate Sulfur Formation in St.
Louis Urban Plume. Atmos. Environ., 12:599-604.
Liu, B. Y. H., D. Y. H. Pui, K. T. Whitby, D. B. Kittelson, Y. Kousaka, and
R. L. McKenzie. 1978. The Aerosol Mobility Chromatograph: A New
Detector for Sulfuric Acid Aerosols. Atmos. Environ., 12:99-104.
McMurray, P. H., and B. Y. H. Liu. 1978. The Tandem Differential Mobility
Analyzer Applied to Studies on Particle Growth and Gas Phase Titration.
Final Report to DOE on Contract No. EY-76-S-02-1248.
McMurray, P. H., B. Y. H. Liu, and K. T. Whitby. 1978. Recent Progress in
Fine Particle Research at the University of Minnesota. Presented at the
GAF Conference, Vienna, Austria.
Roberts, P. T., and S. K. Friedlander. 1976. Analysis of Sulfur in Deposited
Aerosol Particles by Vaporization and Flame Photometric Detection. Atmos.
Environ., 10:403-408.
Whitby, K. T. 1979. Aerosol Formation in Urban Plumes. Presented at New York
Academy of Sciences Conference on Aerosols: Anthropogenic and Natural-
Sources.and Transport, New York, NY.
346
-------
Whitby, K. T. 1979a. University of Minnesota Particle Technology Laboratory
Mobile Laboratories. Technical Report, U.S. EPA Grant No. R803851.
White, W. H., J. A. Anderson, D. L Blumenthal, R. B. Husar, J. D. Husar, and
W. E. Wilson. 1976. Formation and Transport of Secondary Air Pollut-
ants: Ozone and Aerosols in the St. Louis Urban Plume. Science
194:187-189.
347
-------
O
o:
u.
8
10
'<= 6
= 2
SULFUR BY XRF , fjiqS m
-3
Figure 1. Comparison of total sulfur measured with XRF and with the LPI during
the STATE program in August, 1978.
348
-------
CO
100
80
0 60
x
U
_J
UJ
a:
20
0
i 1 \ 1 r
INCREASING HUMIDITY
I
O.05 0.06 0.07 0.08 0.09
PARTICLE DIAMETER ?/Arn^
Source: McMurray and Liu, 1978. ""
Figure 2. Particle diameters achieved when aerosols consisting
of mixtures of ^804 and (NH^SC^, internally
mixed in various proportions, are exposed to rela-
tive humidities shown on the ordinate. The initial
humidity in all cases was about 8 percent.
IOO -
I II I
DECREASING HUMIDITY
H2S04 '(NH4)2S04
0.06 0.07 0.08 0.09
....... PARTICLE DIAMETER , /Lun
Source: McMurray and Liu, 1978.
Figure 3. Particle diameters achieved when aerosols consisting
of mixtures of ^804 and (NH4)2SO4, internally
mixed in various proportions, are exposed to rela-
tive humidities shown on the ordinate. The initial
humidity in all cases was about 95 percent.
-------
60
40
10
o
IO
CO
Cn
o
>
•»
E
20
e PLUME DATA
O HAZE DATA
I
20 40 60
EAA+OPC.V, /im3/cm3
80
Figure 4. Comparison of total volume measured by the EAA and OPC with volumes calculated from LPI data. For LPI
volumes, the aerosol was assumed to be pure H2SO4 in equilibrium at ambient humidities.
-------
7/77>
CO
en
2
fU
60
50
IO
0
IO
1 40
Q.
O
o» 30
0
<
V.
< 20
10
0
0.
1 | 1 1 1 1
5 AUG
22'42-C
OUTLAW
HAZE D
- r
_i
i i i .
78
10:34
FIELD
ATA
1 ' ' K; ^ '
1 1 | III
/> /J y^/'y^^^y/'X
y^ 'XI r S * S f f f f r /*•*» 11 O^ \
|
?
\-r
1 I 1 1 , I ,2 1
03 0.1 1.0
—
— 1
1 1 1 1 1 1 1
1C
Dp,
Figure 5. Comparison of an aerosol volume distribution measured by the EAA and OPC
with one calculated from LPI sulfur distributions.
-------
00
en
100
X 50
o:
o
1 r
LABORATORY DATA
5-AUG-78
CLARKSVILLE ,TN.
i r
0.02
O
O O
i I i i I
O.I
Dp ,
0.4
Figure 6. Comparison of laboratory and field data for the growth of particles with increasing humidity.
These data were measured with a TDMA.
-------
1.0
0.5
Co
en
co
O.I -
TDM A DATA
of
5-AUG-78
CLARKSVILLE , TN
INITIAL DIAMETER
o .175 /xm
+ .114 /im
A .076 /Am
0 .175 jLLm
• .O52/LUT1
0.7
0.9
I.I
1.3
V2/ V2MAX
Figure 7. Number of particles (N) penetrating the second differential mobility analyzer (DMA2) normalized with respect
to the maximum number observed during a voltage scan (Nmax) versus the voltage on the DMA2 collector rod
(normalized with respect to the voltage at Nmax). The solid lines represent the theoretical curve for a monodis-
perse aerosol entering DMA2. This agreement between data and theory suggests the aerosol was internally mixed.
-------
TABLE 1. COMPARISON OF THE bscat> VOLUME FORMATION RATE (dV/dt), AND AEROSOL
SULFUR FORMATION RATE FOR THREE LOCATIONS
Location
St. Louis urban plume
Husar PEPE's estimate
Great Smoky Mountains
PPH2S04
*A<;/Ht - v H\//
bscat/
day
2.2
1
0.5
Ht = °'44 v
dV/dt, urn3
cm day
36
(16.4)t
8
H\//Ht m AA fnv
(dV/dt)/
bscat
16.4
16
70 novront Rt-H
AS/dt,*
M9 S/day
5.2
2.4
1.2
—o ^ uv/uo — —5
tCalculated using (dV/dt)/bscat = 16.
354
-------
APPENDIX A
UNIVERSITY OF MINNESOTA PARTICLE TECHNOLOGY LABORATORY
MOBILE LABORATORIES*
PURPOSE
The University of Minnesota Mobile Laboratory (UMML) and the University
of Minnesota Mobile Home (UMMH) were developed to make possible physical and
chemical aerosol and gas measurements of an advanced nature anywhere there are
roads passable to medium-sized trucks and motor homes. These vehicles are not
intended for long-term monitoring, but rather for intensive measurement.
The vehicles have been designed so that all the electrical power and
necessary support hardware are built in. All instruments are in racks that
can be easily removed because we have found that practically no two studies
are alike and that the complement of instruments is always somewhat different
for each study.
The emphasis has been on continuous real-time measuring instruments,
although a reasonable assortment of filter, virtual impactor, and cascade
impactor samplers may be used.
The data logger used is capable of sampling all electrical signals every
0.6 seconds. Nine-track magnetic tapes recorded by this data acquisition sys-
tem (DAS) can be reduced in the field on a POP 11/10 minicomputer that is
carried along. This off-line mode of operation has been found to be more
satisfactory than on-line data acquisition directly into the computer.
UNIVERSITY OF MINNESOTA MOBILE LABORATORY
Description
The UMML is a 1975 General Motors 15%-ft Step-Van constructed of aluminum.
The body is 16 ft from the driver's seat to the rear, 8 ft wide, and 80 in.
high. Into this shell have been built two Onan generators, one on each side.
*Not presented at the symposium.
355
-------
Two electrical systems have been installed, one delivering power to instruments
and the other delivering power to two 13,000 Btu air conditioners in the for-
ward part of the vehicle. An additional 13,000 Btu air conditioner is install-
ed towards the rear, powered by the vehicle engine. Provisions have also been
made for external power of up to 50 A at 220 V to be plugged in from the side.
The heavy-duty towing package has been installed so that trailers or, as has
often been done, an additional generator can be towed. When extra power is
required, a 15 kW generator may be towed.
A double roof has been installed to provide a sturdy surface on which to
walk and also a suitable mounting surface for instruments.
Figure A-l shows the left side interior of the UMML Two laboratory
benches with drawers for supplies are mounted above the generators and wheel
wells, one on each side. Instruments are mounted by bolting to the bench,
bolting to laboratory strips mounted vertically, or by strapping to logistic
strips mounted horizontally on the walls at several heights.
Electrical System
Two 6.5 kW Onan generators are built in, one on each side. The left-hand
generator supplies power for the instruments and the right-hand supplies power
for the air conditioners, pumps, and lights. An external connection on the
right side also can be plugged into line power or a towed generator. Operat-
ing on internal generators, a total of 13 kW is available, and while operating
with an external generator, a total of 21.5 kW of power is available.
Two 3.5-kW Sorenson line regulators have been provided to regulate power
to the instruments and data acquisition system. Enough transfer switches and
plugs have been provided so that almost any combination of power transfer can
be accommodated.
Pump Systems
Two identical pump compartments, one on each side at the rear of the vehicle,
have been provided. Each contains a 1-hp Cast vacuum pump and a 1-hp Rotron
blower. These pumps and blowers provide enough suction to operate most of the
filters and transfer systems under normal conditions. There is room in the
pump compartments for additional pumps as necessary. For example, a high-vacuum
pump has been installed to operate a low-pressure impactor for some studies.
356
-------
Sampling Systems
Two multiple-inlet sampling feed-throughs have been provided, one on each
side. The one on the left side is underneath the bag sampling system shown in
Figure A-2 and is used primarily for aerosol sampling. The sampling feed-
through on the right side is used both for gas sampling and for aerosol sampling
to a virtual impactor located below it on the laboratory bench. In addition,
some gas sampling inlets may be fed through the upper center part of the front
of the vehicle for more direct access to the instruments.
Sampling probes extend forward of the vehicle on both sides and are
arranged so that they can sample at either the roof level or from a level of
about 1 m above the ground. The roof-level sampling is used for most sampling,
and the lower-level sampling is used when it is desired to simulate what a
driver of a vehicle on a roadway would inhale through the intake of an auto-
mobile.
Bag Sampling System
Figure A-2 shows the present bag sampling system. A bag sampler is
necessary when making aerosol measurements under nonsteady conditions on
roadways because the electrical aerosol analyzer requires that the aerosol
concentration be stable through its operating cycle of about 2 min.
The bag sampler has an electronic control that automatically empties it,
fills it, and holds an aerosol in it for the required 2-min measurement cycle.
This automatic system can be either set to operate automatically with sampling
times from 2 min to 1 hr, or the bag sampling cycle can be initiated by pushing
a button.
Figure A-3 shows the measured transport efficiency of the system shown in
Figure A-2.
Instrument Complement
The instrument complement varies from study to study. Table A-l lists
the instruments that will be used for two field studies during the summer of
1979. Figure A-l shows the instruments on the left side as used during the
1976 Los Angeles Roadway Study. Since that figure was made, a Royco 245
sensor optimized for the 5- to 40-um size range has been mounted on a platform
at roof level in front of the windshield.
357
-------
Data Acquisition
The present data acquisition system is a modified Metrodata 640 data
logger recording on nine-track magnetic tape. The unit has been modified to
accept 0.1-, 1-, and 10-V signals. Normally, the data logger is operated at
its fastest scan rate of abut 0.6 s per scan in order to catch the flags from
the aerosol analyzing system. However, the scan rate can be slowed down
during periods in which monitoring over long periods must be done.
Eight or more channels of strip chart recording are also provided. There
are enough strip charts and visual indicators on all instruments, as well as
several digital volt meters aboard, so that if the data system logger is
inoperative, all instruments can be operated and data recorded, albeit at a
lower frequency.
Data Reduction
Because the only way to determine whether data recording is satisfactory
is to reduce the data on a computer, a Digital Equipment Corp. POP 11/10 with
32 K of memory and two floppy discs is ordinarily brought along, unless we are
operating close to our home base.
Meteorology Equipment
Wind speed, direction, and outside temperature are measured with a Meteo-
rology Research, Inc., meteorological package. The meteorology mast is arranged
so that it can be erected in about 1 min. Thus, wind direction and velocity
can be obtained if the vehicle is stationary for as little as 10 min.
Vehicle Position
Vehicle position is determined by a combination of distance measurement
by an onboard instrument, which records distance on the data logger, and
laboratory digitization of vehicle position from maps marked in the field with
location and time.
Crew and Operation
Normally, three or four people are required to operate the UMML in motion.
Two people can operate it satisfactorily when stationary and during long runs
overnight; it has been found that one person can operate all the instruments,
although at a reduced frequency.
358
-------
When the two vehicles are operating together in the field, normal crew is
five or six, with three or four in the UMML and two in the UMMH.
Detailed checklists for starting up and shutting down the labs have been
developed and found to work quite well.
UNIVERSITY OF MINNESOTA MOBILE HOME (UMMH)
Description
The UMMH is a Winnebago 28-ft motor inn with one 6.5-kW generator installed.
Two air conditioners, one forward and one aft, have been installed. The
electrical system has been rebuilt so that it can operate on either 220-V
internal generator or 110- or 220-V external power. Both air conditioners can
only be operated on external power. The UMMH has a refrigerator, stove,
toilet, shower, and considerable storage built in. In addition, a standard
four-drawer file cabinet has been installed for project records and supplies.
Other than the file cabinet, no other furniture has been built in. Experience
has shown that it is best to strap in the facilities needed for each study.
Instruments and facilities are secured by means of logistic strips placed at
several levels on each side.
Communications
Both the UMML and UMMH have a CB radio installed. In addition, the UMML
has a 720-channel aircraft radio. During some field studies, such as the
STATE study in Alabama and Tennessee, an additional ground communication radio
was installed in each vehicle. Thus, in the STATE program, it was possible to
communicate between vehicles and with aircraft with the base station. For
some studies, a standard radio telephone has also been installed.
MEASUREMENT PROJECTS PERFORMED WITH UMML/UMMH
Following is a brief summary of the measurement programs that have been
completed, and the two that are planned in 1979, with the UMML and UMMH. Also
given are a few comments about the major objectives and results of each study.
St. Louis, August 1976
In St. Louis in August 1976, the UMML made its first measurements of the
Labadie power plant plume as part of the MISTT program. Because this was the
first trip, only a modest amount of useful data was produced. Because the
359
-------
operating speed of the PDP/11-10 computer as a data acquisition system was too
slow, the use of the on-board computer was discontinued, and the Metrodata 640
data logger was used for data acquisition in all subsequent studies.
Los Angeles Roadway Study, October 1976
Under EPA sponsorship, the UMML was equipped with a variety of instruments
aimed at determining the amount of sulfuric acid from a catalyst-equipped car
that might be seen on major freeways. Although useful data on roadway aerosol
were obtained, the study showed that there were no significant concentrations
of H2S04.
State of Minnesota Regional Copper-Nickel Project, Summer 1977
In the summer of 1977, under the sponsorship of the State of Minnesota
Regional Copper-Nickel Project and EPA, measurements were made of the mine and
road dust concentrations found around several iron mines in northern Minnesota.
The most important conclusion was that dirt roads contributed significant
amounts of dust, perhaps even more dust than the mines. This study showed the
feasibility of measuring dust size particles in real time with the UMML.
Diesel Exhaust Studies, June 1978
Under General Motors sponsorship, the UMML was used to measure diesel
exhaust aerosols behind diesel cars. Size distributions and concentrations
were measured and found to agree reasonably well with laboratory measurements.
Sherco Project, June 1978
In June of 1978, the Sherburne County NSP Sherco power plant was studied
in collaboration with the University of Washington, EPA, EPRI, NSP, and Midwest
Research Institute. The purpose of this project was to characterize a clean
power plant equipped with scrubbers operating on low sulfur coal.
Pine Bend Refinery, July 1978
In July of 1978, a 1-day study of the Pine Bend refinery in Minnesota was
made. High concentrations of pollutants are observed in the shallow valley
where this refinery is located. Some very high sulfuric acid aerosol concentra-
tions were found.
360
-------
STATE Program. Summer 1978
The major study of 1978 was participation in the EPA-sponsored STATE
program in Tennessee and Alabama. Measurements on the Cumberland and Widow's
Creek power plants were made. Also, a large amount of data was collected on
the hazes of the east central United States, both in western Tennessee and the
Great Smoky Mountains.
Diesel Exhaust Studies, December 1978
In December 1978, additional diesel exhaust plume studies were done under
Coordinating Research Council sponsorship. The UMML was mounted on a flatbed
trailer that was towed behind a Cummins diesel-powered tractor. Aerosol samples
were obtained at various positions in the exhaust plume.
VISTTA Program, Summer 1979
In the summer of 1979, it is planned to participate in the EPA VISTTA
visibility program in the Four Corners area (Utah, Colorado, New Mexico,
Arizona) of the southwestern United States. The major objective will be
detailed physical and chemical characterization of the haze aerosols in that
area.
Bureau of Mines, Summer 1979
Following the VISTTA study, it is planned to make a dust study for the
U.S. Bureau of Mines at a coal mine near Craig, Colorado. The objective will
be to characterize the size distribution and concentration of dust generated
around this mine.
Other Projects, Summer 1979
Before and after the VISTTA and Bureau of Mines studies, additional
studies of diesel aerosols over roadways will be made.
361
-------
VEHICLE LEFT SIDE
Ladder
Under _
Compartment
Vacuum Transport High
Pump Blower Vacuum
Pump
Rosemount Temperature Sernon
Insulation Layer
Figure A-1. Interior schematic of the left side of the UMML showing the instrumentation
as used during the 1976 Los Angeles Roadway Study.
362
-------
-"====
- nqsm •-
s
2
f.
ft
J
/'
• •> An m
1
^1
*-
GO
cr>
CO
PRESSURE
GAGE
ROTARY
BLOWER
VALVE F
VALVE G
TUBE I
u
CO
10
it
MODE
BAG
FILL*
SAMPLE
DIRECT
OPEN VALVES
G.B
A,E
G,E
FLOW VELOCITY, cm /s
TUBE 1
1150
78
TUBE 2
7
92
TUBE 3
360
360
* TIME DURATION *!5s
Figure A-2. Schematic of the new bag sampler constructed in the spring of 1978. In addition to somewhat reduced losses,
this system permits direct sampling from straight above the UMML.
-------
100
CO
en
-pi
O BAG SAMPLING
D DIRECT SAMPLING
0.01
0.1 1.0
PARTICLE DIAMETER , MICROMETERS
Figure A-3. Sampling efficiencies of the UMML sampling system used in 1978. Efficiencies were measured experimentally
using monodisperse aerosol. "Direct Sampling" refers to drawing the sample through the horizontal tube and
directly down to the instruments at a flow of 531 Ipm. "Bag Sampling" refers to drawing a sample into the
bag through the horizontal tube at about 780 Ipm, and then drawing the sample from the bag at 6.8 Ipm. The
inside of the bag was coated with a thin soap film to reduce particle losses due to electric fields.
-------
TABLE A-l. UMML/UMMH
(available
INSTRUMENTATION
summer 1979)
Variable or Function
Position - maps
- speed and location
- distance traveled
Meteorology
Inside temperature
Dew point
Wind speed & direction (stationary)
and outside temperature
Visibility, bscat
Gas Chemistry
S02 (gaseous S)
NO, N02, NOV
03 X
CO
Aerosol Size Distribution
Aitken Nuclei
0.0056 - 1 |jm
0.056 - 6pm
5.6-40 [im
30 +
Aerosol Chemistry
X-ray fluorescence
Sulfur
X-ray fluorescence
Carbon analysis
Aerosol sulfur
Aerosol growth with humidity
Single particle XFL
Instrument
SPACEKOM, Inc., Auto Company
Rosemount Engineering Temperature Sensors
EG & G, Model 820 Hygrometer
MRI, Model 840-1 T sensor
Model 1074 WSPD & WDIR
Model 840 Aspirated Temperature Sensor
Modified MRI, Inc., Model 1561
Integrating Nephelometer
Meloy SA 285
TECO Series 14
Dasibi, Model 1003 - AH
Energetics Science, Inc., 2000 Series
General Electric GE-1 Condensation
Nuclei Counter
Environment/One
Rich 100 or TSI Model 3020
TSI Model 3030 EAA
Royco 220, MCA and Ratemeter
Royco 245, MCA and Ratemeter
Filter & Microscope
50 1/min dichotomous virtual impactor--
collection on filters—cut at 2.5 |jm
Low pressure impactor
Micro-orifice impactor—cuts at 0.1,
0.3, 1, and 2.5 pm—collection on
37-mm filter
47-mm filters for carbon analysis
Pulsed Meloy SA 285--continuous
TDMA—TSI Model 3020 CNC
Filter analysis at Penn State
(continued)
365
-------
TABLE A-l (continued)
Data
Data acquisition
Data recording
Data reduction
Strip charts
Communication
Ground
Air
Miscellaneous
Telephotometer
Sun photometer
Photography
Time lapse
Calibration
03, S02, NOX
Flow
OPC's
Solar radiation
Broad-band and ultraviolet
radiation
Metro Data Model 640 DL
Digi Data Corp. Model 1309-8-B2
9-track magnetic tape
DEC POP 11/10 Minicomputer - off-line
Two Chessel 3-channel, one Linear
3-channel, one Linear 2-channel
40-channel CB in UMML & UMMH
Genave Model 720 p 720 channel aircraft
radio in UMML
(Provided by EPA)
5-band Voltz sun photometer—modified
35-mm camera
Super 8 movie camera
Monitor Labs. Model 8500
gas calibrator
TSI Model 1352 and 1050 mass flow meters
Aerosolized PSL & Polystyrene
Eppley pyrheliometer
366
-------
COMPARISON OF AIR QUALITY WITHIN THE MIXED LAYER WITH THAT AT THE SURFACE
James J. B. Worth
Research Triangle Institute, Research Triangle Park, North Carolina
Abstract
Data collected during the Sulfate Regional Experiment were investigated
to determine how well data collected at surface stations compared to data
collected on board an instrumented aircraft flown within the mixed layer.
Data were compared for all flights and for flights grouped by time of day by
calculating correlation coefficients (r-values). The gaseous species, 03 and
S02, and measurements of particulate sulfate and total suspended particulate
were considered in this analysis. The S02 and sulfate data within the two
measurement regimes were found to have correlations significantly different
than zero for all time periods and the ozone data appeared to be correlated in
the afternoon.
Other analyses were made using averages of all the aircraft data, and
data from selected individual flights. These additional analyses were per-
formed to provide information useful in understanding the results of the cor-
relations. The additional analyses included:
1. Averaging the aircraft data over all flights within the mixed
layer, above the mixing height, and in 60-m (200-ft) increments.
2. Examination of aircraft data collected on individual flights through
the near-surface layer to an altitude of 15 m (50 ft) above the
ground.
3. Comparison of data collected on three vertical spirals separated by
no more than 80 km (50 mi).
INTRODUCTION
A primary objective of this study was to identify those conditions under
which surface air quality can be expected to represent the air quality in the
367
-------
mixed layer. This was done by comparing air quality data collected at the
surface with that collected by aircraft during the Sulfate Regional Experiment
(SURE). The results provide guidance for the development of study designs
that will make optimum use of valuable aircraft flight time in future studies.
The SURE program provided an excellent data base for this work because air
quality measurements made aboard aircraft on spiral flights above surface
stations were available at several locations, during the four seasons, and at
three times of the day.
Flight plans for the SURE aircraft called for two flights per day con-
sisting of three vertical spirals per flight. These flights were conducted
at sunrise and at midday during the first four measurement periods and at
midnight and midday during the remaining two measurement periods. A total of
approximately 400 spirals were flown by SURE aircraft. Spirals between the
surface and 3 km were flown upwind of, over, and downwind of a specified SURE
Class 1 station. The spacing between the spirals during the first four
intensives was 25 km (15 mi) and during the last two intensives was 80 km
(50 mi). A typical flight pattern is shown in Figure 1.
Measured parameters included meteorology, position, and air quality
data. Continuous air quality data included S02, 03, NO/NO . b__ . , and
X SCat
condensation nuclei concentration (CNC) data. Sulfur dioxide was also deter-
mined by analysis following 30 min of sampling through filter packs. Partic-
ulate samples were collected for 15-min and 30-min periods. The 30-min
samples were collected such that they provide resolution from the surface to
1.5 km and from 1.5 to 3.0 km. Particulate samples were subsequently analyzed
for $04, N03, and N04 in addition to particulate mass.
Data from selected flights were examined first to determine the agree-
ment between the air quality measured on the three spirals flown within 50 km
of one another. Second, the data collected on low-altitude extensions of the
typical spirals, which tied the aircraft data to the surface data, were exam-
ined. This analysis provided insight into the causes of poor agreement
between surface and aircraft data.
Statistical procedures were applied to the entire aircraft ground data
set. These analyses included averaging air quality data over all spirals and
calculating correlations between corresponding measurements at the surface
368
-------
and aboard the aircraft. The results of these procedures provided information
on the agreement between surface and aircraft data in general. The analysis
was not extended to investigate the relationship between surface and aircraft
air quality data between seasons at the same location or during the same sea-
son at different locations. The available data were not sufficient to provide
meaningful results for individual locations or seasons.
RESULTS AND DISCUSSION
Comparison of Data from Three Spirals at the Same Locations
The standard three-spiral flight plan followed during the SURE made an
important analysis possible. The measurements made during individual spirals
of the separate flights were compared by testing the mean concentrations of
gaseous parameters above and below the mixing height to determine if there
was any significant difference between the mean concentrations on the three
spirals.
Figures 2 and 3 present the data of three spirals overlaid on one another
for the morning and afternoon flights of 15 April 1978, respectively. These
flights took place at Research Triangle Park, NC, and were separated by
approximately 24 km (15 mi). These figures clearly show the agreement that
can be expected between spirals with this separation, not only with respect
to 03 and S02, but for the other parameters measured as well.
Throughout the first four periods of intensive measurements (intensives)
the agreement among the data from the three spirals at 24-km (15-mi) separa-
tions was consistent. As a result of this, it was decided to extend the dis-
tance between spirals to approximately 80 km (50 mi) during the remaining two
intensives. Table 1 lists the spiral means and results of tests of signifi-
cance for 03, S02, and NO among the spirals separated by 80 km (50 mi)
/\
during the July and October 1978 intensives. During these intensives at
Scranton, PA, and Zanesville, OH, there was no significant difference in the
mean concentrations of 03, S02, or NO below 1.5 km (nominal mixing height)
s\
at the 80-km (50-mi) separation. There was a significant difference for NO
/\
above the mixing height during both intensives and for 03 above the mixing
height during the October intensive at Zanesville, OH. These results suggest
that, depending on the needs of a particular measurement program, individual
spirals flown more frequently during the day may be a more efficient use of
369
-------
valuable aircraft flight time than the flight plans followed in this study.
Figures 4 and 5 represent the data of the three spirals overlaid on one
another from the noon and midnight flights of 12 July 1978 at Scranton, PA,
respectively. Once again, in both of these figures the data for all parame-
ters agree quite well. Note in Figure 5 that while the mixing height varied
considerably over the three spirals in the midnight flight, the concentrations
of 03 above and below the mixing height on the three spirals were in good
agreement.
Low-Approach Flight Data
Vertical spirals were extended from the typical spiral base altitude
(0.45 km [1,500 ft] MSL) down to approximately 15 m (50 ft) above the ground
near a surface station to examine the vertical distribution of the measured
pollutants near the surface. Figures 6 through 9 show selected examples of
these data along with the corresponding surface station measurements. These
data have been helpful in understanding the reasons for the poor correlations
between the surface and aircraft measurements.
Figure 6 represents the low-approach and surface station measurements
made during the morning flight on 12 April 1978 at Research Triangle Park, NC.
For this flight, the extension of the aircraft spiral was completed at RDU
airport approximately 4.8 km (8 mi) from the surface station. On the morning
flight of 12 April 1978, there were steep gradients 'of NO, NO , and 03 near
j\
the surface. These gradients were a direct result of the presence of a noc-
turnal radiation inversion. The nocturnal inversion served to trap NO emis-
*\
sions through the night in a shallow layer near the surface. The rapid reac-
tion between NO and 03 caused a decrease in the 03 concentration and an
.increase in the N02 concentration (represented by the difference between NO
f\
and NO). Figure 7 represents data from a similar low-approach comparison
flight at Research Triangle Park, NC, during the afternoon of 24 April 1978.
By the time of the afternoon flight, the surface inversion had dissipated and
good mixing had been established. The lowest part of the vertical profile
agreed quite well with the surface station measurements during this flight.
On this flight the ozone concentration up to 0.59 km (2,000 ft) MSL agreed
well with the surface measurements. These results suggest that there is an
apparent relationship between surface ozone and the mean ozone concentration
within the mixed layer during the afternoon.
370
-------
Local sources sometimes have a serious impact on the comparison between
surface and aircraft data. The effect of a local source is illustrated by
the low-approach and surface station data from the afternoon flight of 13 Octo-
ber 1978 at Zanesville, OH, shown in Figure 8. There is a local source of
S02 in the area and, as seen in the figure, the S02 at the surface was much
greater than the concentration measured aboard the aircraft. These data also
illustrate the effect of sampling aboard an aircraft on the edge of a large
urban-type plume. The spiral path may have caused the aircraft to periodi-
cally enter and leave an area of high NO and NO concentration characteristic
/\
of an urban plume.
Figure 9 represents the low-approach and surface station data collected
on the night flight of 23 October 1978. The influence of a shallow surface
inversion is evident in the steep gradient of ozone very near the surface.
The inversion layer is formed shortly.after sunset and continues to deepen
and become progressively more stable through the night. The destruction of
reactive species such as ozone is influenced not only by chemical reaction
with other pollutants but also by uptake and reaction at the surface. The
data in Figure 9 show that while the NO concentration has not increased
significantly at the surface, the ozone concentration has already decreased
considerably below the mean ozone concentration above the inversion.
Mean Concentrations
The means and ranges of 03 and S02 concentrations at the surface, below
the mixing height, and above the mixing height for all flights considered in
this analysis are presented in Tables 2 and 3. The means and ranges for all
flights grouped by the separate time periods are also listed in Tables 2
and 3. The mean ozone concentrations increased with increasing altitude over
all flights. These observations were consistent even when the data were
averaged over the flights within the separate time periods, although during
the midday flights the mean ozone concentrations at the surface, below, and
above the mixing height were nearly equal.
The behavior of S02 was quite different from that of ozone. Both the
mean and maximum S02 concentrations generally decreased with altitude over
all flights as well as over the flights within the separate time periods.
371
-------
Mean ozone concentrations above the mixing height were consistently
greater than the mean concentration either below the mixing height or at the
surface, although the difference was small during the noon flights. The mean
ozone concentration above the mixing height was nearly the same in each of
the time periods. There was a significant difference, however, between the
ozone concentration at the surface during the midnight and early morning
flights compared to the mean surface ozone concentration during the noon
flights. The mean ozone concentration within the mixed layer, although lower
than that above the mixing height, was also quite consistent across the
separate time periods. These trends suggest that the effects of transport,
synthesis, and destruction of ozone are more in balance above the surface on
the diurnal time scale than they are at the surface. The impact that the
ozone above the mixing height has on surface air quality is as yet not fully
understood. However, as atmospheric conditions change, some of the ozone
above the mixing height may be mixed to the near-surface layer and affect
surface air quality.
The mean concentrations of S02 at the surface, below, and above the
mixing height were similar a'coss the separate time periods, although the
maximum concentrations differed considerably. The consistency of the means
over the time periods may suggest that the primary S02 sources are continuous
throughout the day. Large stationary sources of S02, such as coal-fired
power plants, may therefore be the dominant influence on the mean surface
concentrations, with the large ranges observed resulting from the presence or
absence of the direct influence of relatively local sources.
Table 4 lists the means and ranges of the sulfate concentration measured
at the surface, below 1.5 km (5,000 ft), and above 1.5 km (5,000 ft) over all
flights and for the morning and afternoon flights. The data from midnight
flights were too few to establish any meaningful conclusions and were not
considered. The data in Table 4 indicate that, on the average, the sulfate
concentrations at the surface and below 1.5 km (5,000 ft) agreed quite well.
The agreement between the 3-hr average sulfate concentration at the surface
and the mean sulfate concentration below 1.5 km (5,000 ft), averaged over all
three spirals, was within 10 percent on both the morning and midday flights.
The sulfate concentrations measured above 1.5 km (5,000 ft) were consistently
372
-------
lower than those measured both below 1.5 km (5,000 ft) and at the surface.
The sharp decrease in sulfate concentration above 1.5 km (5,000 ft) confirms
that atmospheric sulfate arises primarily from surface-related sources.
Mean Vertical Spirals
All of the available spiral data were averaged in 60-m (200-ft) incre-
ments, and the results for each individual parameter were plotted as mean
vertical spirals. There is one plot for data below 1.5 km (5,000 ft) and one
for data above 1.5 km (5,000 ft) for each parameter. These plots appear in
Figures 10 through 17 for S02, NO , and 03, respectively. The solid dots
/\
represent plus and minus one standard deviation from the mean in each of the
60-m (200-ft) increments. To provide good visual resolution, the concen-
tration data for individual species were scaled differently in the two alti-
tude regions.
Figures 10 and 11 show the plots of mean S02 concentrations below and
above 1.5 km (5,000 ft) MSL, respectively. The maximum mean S02 concentration
was approximately 0.015 ppm and occurred at 0.4 km (1,400 ft), approximately
180-240 m (600-800 ft) above the average surface elevation. At the lowest
elevation, the mean concentration was approximately 0.009 ppm, and the mean
surface concentration was 0.014 ppm. In the low-altitude portion of the mean
spiral, the standard deviation for S02 was large, nearly 100 percent of the
mean concentration, indicating that there was a large variability in S02
concentration measured near the surface. These plots represent data averaged
over spirals flown at several different locations, and much of the variability
is suspected of resulting from local sources and variations in surface terrain.
At altitudes above 0.4 km (1,400 ft), both the mean S02 concentration and the
standard deviation decreased continually. The occurrence of the maximum at
0.4 km (1,400 ft) suggests that stationary ground sources with tall stacks
have a significant impact on the initial vertical distribution of S02. The
rapidly decreasing mean S02 concentrations all the way to 3 km (10,000 ft)
result from dilution and chemical reaction as the S02 is mixed away from the
source area.
The mean vertical spiral for NO shown in Figures 12 and 13 is similar
/\
to that for S02. The maximum value of approximately 0.013 ppm occurred at
0.35 km (1,200 ft). There was a secondary maximum of approximately 0.012 ppm
373
-------
in the very lowest altitude layer and the mean NO concentration at the sur-
/\
face was 0.010 ppm. The maximum mean concentration at 0.35 km (1,200 ft) is
probably a result of the same type of emissions as were suggested for S02.
The secondary maximum in the lowest altitude increment shows the importance
of surface sources, primarily automobiles, on the vertical distribution of
NO . The standard deviation for NO , like that for S02, was large very near
/N X
the surface and decreased with altitude, especially below 1.5 km (5,000 ft).
The variability of local source strengths and surface terrain are, again,
suspected of contributing to the large variability in NO concentrations
J\
observed near the surface. The decreasing concentrations with altitude are
also a result of dilution and chemical reaction.
Unlike S02 and NO , the mean vertical concentrations of ozone, which are
y\
shown in Figures 14 and 15 increased with altitude up to 1.5 km (5,000 ft).
The standard deviation remained nearly constant over the entire altitude
range. The lack of significant amounts of NO and S02 at altitudes above
A
1.5 km (5,000 ft) indicates that these surface-related primary emissions are
diluted and chemically altered as they are mixed in the atmosphere. The mean
profile of ozone, however, further suggests that there is a more stable
balance among synthesis, destruction, and transport at altitudes above 1.5 km
(5,000 ft) than at the surface. The standard deviations of the means for
ozone were consistent over the entire altitude range, further substantiating
»
the stability of high ozone concentrations above 1.5 km (5,000 ft).
The mean values of b . decreased, in general, from the lowest portions
of the aircraft spirals to 1.5 km (5,000 ft) as shown in Figure 16. The
standard deviation from the mean was similar over that altitude range. This
is tc be expected because the surface itself can serve as a source of parti-
cles that are included in the measurement of b .. Above 1.5 km (5,000 ft),
however, the mean b . values, shown in Figure 17, decreased more slowly,
and above Z.I km (7,000 ft) they remained nearly constant. The standard
deviation started to vary consistently above 2.1 km (7,000 ft). There is
often a concentration of b . just below an inversion layer and the large
and variable standard deviations shown in Figure 17 above 2.1 km (7,000 ft)
were a result of variation in the mixing depth over the spirals.
374
-------
Correlations Between Surface and Aircraft Data
Correlations were calculated between aircraft data and surface station
data for comparison. Table 5 lists the correlations obtained for the data of
all flights. A statistical test was performed to determine if the correla-
tions were significantly different from zero; those correlations in Table 5
that were significant are marked by an asterisk. Selected scatter diagrams
of the data used in calculating these correlations are shown in Figures 18
through 23. Although some of the correlations were determined to be signifi-
cant, inspection of the scatter diagrams reveals, in some cases, that the
correlations were apparently fortuitous and not actually meaningful.
The S02, sulfate, and ozone concentrations within the mixed layer and
S02 above the mixing height were found to be correlated with the correspond-
ing surface data. Figure 18 is the scatter diagram for S02 within the mixing
layer versus S02 at the surface for all flights. Although a large fraction
of the data are clustered around the origin, there does appear to be the
potential for significant correlation between the two sets of data. There
are two outliers for which the surface S02 was approximately 0.07 ppm and the
average S02 within the mixed layer was below 0.015 ppm. Undoubtedly the
correlation would be improved considerably if these two outliers were removed.
Even with them included, the correlation of 0.56 was significant. The scatter
diagram of sulfate below 1.5 km (5,000 ft) versus sulfate at the surface,
shown in Figure 19, with a correlation of 0.61 also shows a potential for a
significant correlation. In addition, a linear regression equation fit to
these data would have a slope very close to unity. There were no outliers in
this data set, indicating that, in general, the surface sulfate concentration
is similar to the sulfate concentration in the mixed layer. This result
further substantiates the agreement between surface sulfate and sulfate
concentrations below 1.5 km (5,000 ft) inferred from the table of mean concen-
trations.
The correlations between surface and aircraft data for flights within
the separate time periods are listed in Table 6. Once again the asterisk
marks those correlations that are significantly different from zero.
The scatter diagrams for S02 and sulfate within the mixed layer versus
S02 and sulfate at the surface for morning flights, both of which have a cor-
375
-------
relation of 0.54, are shown in Figures 20 and 21. In both cases the data
appear to justify the significance of the correlation. The correlation of
0.69 for S02 above the mixing height versus S02 at the surface for morning
flights was determined to be significant. The scatter diagram of these data,
however, shown in Figure 22, indicates that a single outlier influenced the
correlation significantly, and the remainder of the data did not support the
significance of the correlation.
The scatter diagram for S02 within the mixed layer versus S02 at the
surface for afternoon flights, shown in Figure 23, strongly supports the sig-
nificance of the 0.80 correlation. For these data the regression equation
would have a slope near unity as was the case for sulfate below 1.5 km
(5,000 ft) over all flights. At the present time, it appears that a potential
for significant correlation exists between S02 and sulfate at the surface and
within the mixed layer at all times. The 03 concentration within the mixed
layer is also likely to be correlated with surface ozone concentrations in
the afternoon.
CONCLUSIONS
V
The mean concentrations of 03 increased with altitude over all flights
from the surface to above the mixing height. The mean concentrations of S02
decreased with height over all flights from the surface to above the mixing
height, and the mean sulfate concentrations were approximately equal at the
surface and below 1.5 km (5,000 ft) and decreased by approximately 50 percent
above 1.5 km (5,000 ft).
Due to the mean vertical distribution of ozone, large amounts of ozone
can ba isolated from rapid destructive influences near the surface, and may
be transported large distances downwind. The vertical distribution of ozone
was found to be dependent on time of day. The mean concentrations of ozone
at the surface, within the mixed layer, and above the mixing height during
the afternoon flights, were within 10 percent of one another.
The mean concentrations of S02 and sulfate at the surface below the mix-
ing height and above the mixing height, in general, were not dependent on the
time of day.
The mean concentration of 03 in 60-m (200-ft) increments increased with
height and the standard deviation remained approximately the same with alti-
376
-------
tude. The mean concentrations of S02 and NO had maxima at 0.40 km (1,400 ft)
/\
and 0.35 km (1,200 ft), respectively, and the standard deviations decreased
continually with height.
Correlations of S02 and sulfate at the surface and within the mixed
layer were found to be significantly different from zero. Ozone concentra-
tions at the surface and within the mixed layer showed an apparent correlation
for the afternoon flights but not during the morning flights.
Data collection aboard the aircraft down to altitudes within 15 m (50 ft)
above the ground provides information useful in understanding the reasons for
poor correlations between the surface air quality and the mean air quality
within the mixed layer.
Evidence suggests that there may be no significant difference between
the mean concentration of ozone or S02 below the mixing height on spirals
separated by 24 km (15 mi) or 80 km (50 mi).
More efficient use of valuable aircraft flight time may be possible if
single spirals are flown more frequently during the day.
377
-------
to
>xi
00
10,000 FT
DOWNWIND
REFERENCE
(SPIRAL #3)
CLASS!
STATION SITE
(SPIRAL #2)
UPWIND REFERENCE
(SPIRAL #1)
Figure 1. Typical three-spiral SURE aircraft flight plan.
-------
CO
~j
-------
CO
00
o
11—I
10.
7_
ft-,
tL.
H
d 3_
2_
1.
-SO -25
.0 .03
-20 -15 -10 -5 0
.06 .09 .12 PPM 03,S02
HO ,NOX 0 .05 .10
J 10 15 20 25 80
0 2 I *
.15 PPH NO/NOv I 3
0 25 50 78 100
35
t
1
128
• • ••
40 »C TEMP.OEM PT
10 TURBULENCE
I 10-VM BWflT
150 lO'/CC CN
Figure 3. Concentration profiles of all three spirals of the afternoon flight of 15 April 1978 overlaid on one another.
-------
CO
00
BrSCflT
>0.61
.0.60
-30
.0
-25
.03
i
-20
.06
NO
i
-15
.09
.NOx
i r
-10-5
.12 PPM 09.
0 .05
S
10
15
.10 .19 PPH NO/NOx
.0.00
20 25 ' 30 55 10 «C TEMP.DEH PT
2 1 68 10 TURBULENCE
02*6 10-VHflscflT
Figure 4. Concentration profiles of all three spirals of the afternoon flight of 12 July 1978 overlaid on one another.
-------
4.85
LO
00
FXJ
-5 0
.12 PPM08.IOa
.0 .05 .10
10 TURBULENCE
» 10-VMBjcflT
Figure 5. Concentration profiles of all three spirals of the midnight flight of 12 July 1978 overlaid on one another.
-------
OO
00
CO
2.0
_18
"5
tie
o
V*
"§1.4
03
o1.2
H
ujlO
t -8
i_
ti
< .6
l
1 1 1 1 1 II
-SO2 OjNONOx DP TEMP bSCAT -
, |
Hx,%
- >!"
\
}
*
-
V
1
^
\
\
\
\
\
\
\
\
\
> 1 \ \
li V \
1 >! \ t
-
/
- /
... , , -/
• ' 1
i \ 1
> V /
; XN '
i *
*• \
.. — r- T . i i », i i
_
—
—
—
"
•—
—
1 1
' 1 ' 1 1 • 1 1 i 1
.61
In
fZK (1)
«OO j~.
0)
.49*
O
C/5
.43-0
c
03
.37 |
•c
.30 r
LU
O
.24 g
1—
.18 <
-20 -15 -10 -5 0^^ 5 10 15 20 25 30
0 .03 .06 .09 .12 ^\. 0^ 2 4 6
I .0 .01 .02 .03] \.
4>PM 03802^1^ NO NOX LTEMP nO"4M bsCAT
DEW PT
Figure 6. Low approach and surface station data collected during the morning flight on 12 April 1978 at
Research Triangle Park, NC.
-------
oo
CD
o1.2
< .6|
±
NO NOX O3
±
±
-15 -10 -5 (
.03 .06 .09 .12
.0 .01 .02
'PM 03SO2LPPM NO NOX
10
TEMP
15
0
20
2
25
4
.61
In
.55^
a>
.49 f
o
c/>
43"o
.-ro c
03
(f)
^30-
Q
.24 P
.18 <
30
6
bSCAT
DEW PT
Figure 7. Low approach and surface station data collected during the afternoon flight
on 4 April 1978 at Research Triangle Park, NC.
-------
00
cx>
Ul
2.0
.0)
CD
c
CO
4
-**
o1.2
LLltO
Q
t -8
b
< .6
20
0
1
-15
.02
DP1 TEMP1
I
I
1
I
-10
.04
0
-5 0
.06 .08
.02 .04
NO NOX
10
15
0
TEMP
DEW PT
I
20
2
b !SC AT
I
25
4
.61
o>
.49
o
CO
43 T3
• c
CO
*>
o
•24t
b
.18 <
30
6
"IO~4M bSCAT
Figure 8. Low approach and surface station data collected during the afternoon flight on 13 October 1978 at Zanesville, OH.
-------
2.0
^1.8
0)
o>
!ii 16
"o
(0 '
en
§1.2
h-
!ljj1.0
Q
15
t -8
t] '
< .6
l I II II l I I
SO2 O3 NO NOX D
I L "^
\ \ h
A y V\
\ I \
\ K
I //
i j r
P TEMP bs
I
I
\
\
\
1
1
-/ r r^ ^ ^
AT ._
_
__
-. ' ~
—
—
_ . —
— —
i i ii i i i i i
.61
.55 5
0)
.49^
0
•43I
0)
•37§
JZ
•30m
Q
.24^
^
.18 <
-20 -15 -10 -5 0^ 5 10 15 20 25 30
0 .02 .04 .06 .08 ^\. 0246
I 0 .02 .04 .06 ] \.
4>PM OaSC^^PM NO NOX LTEMP ^10"4 M bSCAT
DEW PT
Figure 9. Low approach and surface station data collected during the midnight flight on 23 October 1978 at Zanesville, OH.
-------
CO
00
CO
•
O
o
O
o
r-
LL1
O
D
r-
5-
4-
3-
o
-.006 -.003 .000 .003 '.006 .009 .012 .015 ' .018 .021 .024 .027
PPM
Figure 10. Mean SO2 concentrations in 200-ft increments below 1.5 km (5,000 ft) over all SURE aircraft spirals.
-------
CO
00
00
_J
C/)
10
I 9
^
CO
d 8
o
o
o
7
LU 6
h-
•
.003
.000
PPM
.003
.006
Figure 11. Mean SO2 concentrations in 200-ft increments above 1.5 km (5,000 ft) over all SURE aircraft spirals.
-------
CO
00
CO
6
o
o
o
t—
HI
Q
t
_J
<
5-
4-
3-
2-
1-
\
.000 .002 .004 .006 .008 .010 .012
PPM
.014 . .016 .018 .020
Figure 12. Mean NOX concentrations in 200-ft increments below 1.5 km (5,000 ft) over all SURE aircraft spirals.
-------
u>
u>
o
CO
0
o
o
o
t—
LLJ
Q
h;
r-
<
10
9
8
7-
6-
5-
•
•
•
e
-.002
.000
.002
PPM
.004
.006
Figure 13. Mean NOX concentrations in 200-ft increments above 1.5 km (5,000 ft) over all SURE aircraft spirals.
-------
CO
CO
•
o
o
o
o
r-
LU
O
1-
.015 ;020 .025 .030 .035 .040 .045 ,050 \055 .060 .065 .070 .075 .080
PPM
Figure 14. Mean 03 concentrations in 200-ft increments below 1.5 km (5,000 ft) over all SURE aircraft spirals.
-------
CO
CO
•
o
o
o
o
r-
LU
O
D
10
9
^
8
7
6
5-1
.035 .040 .045 .050 .055
PPM
.060
.065 .070
.075
.080
Figure 15. Mean 03 concentrations in 200-ft increments above 1.5 km (5,000 ft) over all SURE aircraft spirals.
-------
CO
V£>
CO
CO
•
o
2-
o
o
o i
111
Q
D
r-
•
i
•
<
•
Oil
i.o
1:4
2:2
216
'scat 10"4/m
Figure 16. Mean bscat values in 200-ft increments below 1.5 km (5,000 ft)
over all SURE aircraft spirals.
-------
CO
ID
10
co 8
•
o
7
O
o
UJ
Q
-0.2
0.2 0.6 1.0
"scat 10"4/
1.4
1.8
2.2
m
Figure 17. Mean bscat values in-200-ft increments below 1.5 km (5,000 ft)
over all SURE aircraft spirals.
-------
01
0.040
S2
UJ
I
2
X
0.030
O 0.020
UJ
DO
UJ
0
X
O
O
cc
0..010
D
CO 0.000
0.00
0.01 0.02 0.03 0.04 0.05
SULFUR DIOXIDE AT SURFACE (PPM)
0.06
0.07
0.08
Figure 18. Scatter diagram of sulfur dioxide concentrations at the surface versus below the mixing height for all flights.
-------
CO
to
O
o
27 r
24
21
18
15
12
If)
•
3 9
Ul
CO
LU
5 6
LL
D
(/) 3
0 -
0
• •
• •
8
10
12
SULFATE AT SURFACE (/*g/m3)
14
16
Figure 19. Scatter diagram of sulfate concentrations at the surface versus below 1.5 km (5,000 ft) MSL for all flights.
-------
co
ID
0.040
g
UJ
0.030
0.020
UJ
m
LU
9
O
5
DC
o.oio
CO 0.000
• •• •
0.00
0.02
0.04
0.06
0.08
SULFUR DIOXIDE AT SURFACE (PPM)
Figure 20. Scatter diagram of sulfur dioxide concentrations at the surface versus below the mixing
height for all morning flights.
-------
24 r
CO
l£>
00
CO
t 18
o
o
q
10
in
•
r-
O
UJ
CD
UJ
12
0
0
8
10
12
14
16
SULFATE AT SURFACE
Figure 21. Scatter diagram of sulfate concentrations at the surface versus below 1.5 km
(5,000 ft) for all morning flights.
-------
CO
UD
I
r-
o
0.040
0.030
Uj 0.020
>
o
00
<
111
x 9-010
g
o
cc
0.000
0.00
0.02
0.04
0.06
0.08
SULFUR DIOXIDE AT SURFACE (PPM)
Figure 22. Scatter diagram of sulfur dioxide concentrations at the surface versus above the mixing height for all flights.
-------
O
O
«£ U.UM-U
Q.
Q_
r-
(D
!£ 0.030
CD
X
^
«2
g 0.020
0
__J
LU
CO
LU
Q
x ,0.010
O
Q
DC
LL
|
D 0.000
•
•
• •
•
•
•
•
- • •
• •
•
•
• •
•
."*"•
-?••••• , . ,
CO
0.00
0.02 0.04
SULFUR DIOXIDE AT SURFACE (PPM)
0.06
Figure 23. Scatter diagram of sulfur dioxide concentrations at the surface versus below the
mixing height for afternoon flights.
-------
TABLE 1. MEAN S02, NO , AND 03 CONCENTRATIONS ON THREE SPIRALS DURING
JUL? AND OCTOBER 1978 INTENSIVES
July 1978
Scranton, PA
S02
N0x
03
October 1978
Duncan Falls, OH
S02
N0x
03
Spiral
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
1
2
3
Below mixi
Sample mean
(ppm)
.0011
.0012
.0019
.004
.002
.003
.068
.077
.082
.0032
.0030
.0044
.013
.010
.010
.043
.044
.040
ng height
Level of
significance
NS
NS
NS
NS
NS
NS
Above mixing height
Sample mean Level of
(ppm) significance
.0002
.0006 NS
.0002
.002
.001 0.05
.001
.079
.077 NS
.085
.0008
.0006 NS
.0006
.005
.003 0.05
.004
.054
.052 0.05
.049
NS = Not significant.
-------
TABLE 2. VERTICAL DISTRIBUTION OF OZONE DURING SURE
INTENSIVE MEASUREMENT PERIODS
Mean [03], ppm Range, ppm
All flights
Surface
Below
Above
mixi
mixi
ng
ng
height
height
0.
0.
0.
034
047
054
0.
0.
0.
000 -
004 -
018 -
0.
0.
0.
084
113
101
Sunrise flights
Surface 0.020 0.000 - 0.044
Below mixing height 0.042 0.010 - 0.077
Above mixing height 0.052 0.026 - 0.101
Noon flights
Surface 0.047 0.001 - 0.084
Below mixing height 0.052 0.004 - 0.113
Above mixing height 0.054 0.018 - 0.092
Midnight flights
Surface 0.022 0.005 - 0.039
Below mixing height 0.049 0.023 - 0.079
Above mixing height 0.060 0.025 - 0.089
402
-------
TABLE 3. VERTICAL DISTRIBUTION OF S02 DURING SURE
INTENSIVE MEASUREMENT PERIODS
Mean [S02] , ppm
All
flights
Surface
Below mixi
Above mixi
ng
ng
height
height
0.
0.
0.
014
007
002
Range, ppm
0.
0.
0.
000
000
000
- 0.
- 0.
- o.
186
041
035
Sunrise flights
Noon
Surface
Below mixi
Above mixi
flights
Surface
Below mixi
Above mixi
ng
ng
ng
ng
height
height
height
height
0.
0.
0.
0.
0.
0.
013
008
002
015
008
001
0.
0.
0.
0.
0.
0.
001
000
000
000
000
000
- 0.
- 0.
- 0.
- 0.
- 0.
- 0.
072
034
035
186
041
010
Midnight flights
Surface
Below mixi
Above mixi
ng
ng
height
height
0.
0.
0.
013
003
001
0.
0.
0.
000
000
000
- 0.
- 0.
- 0.
075
006
004
TABLE 4. VERTICAL DISTRIBUTION OF SULFATE DURING SURE
INTENSIVE MEASUREMENT PERIODS
[S04], ug/m3
Range, ug/m3
All flights
Surface
Below 1.5 km (5,000 ft) MSL
Above 1.5 km (5,000 ft) MSL
Sunrise flights
Surface
Below 1.5 km (5,000 ft) MSL
Above 1.5 km (5,000 ft) MSL
Noon flights
Surface
Below 1.5 km (5,000 ft) MSL
Above 1.5 km (5,000 ft) MSL
7.456
7.301
3.615
6.934
6.206
3.100
7.913
8.305
4.096
1.200
0.170
0.000
2.100
0.170
0.000
1.200
2.060
0.000
32.100
27.760
10.860
14.300
17.180
8.420
32.100
27.760
10.860
403
-------
TABLE 5. CORRELATIONS BETWEEN SURFACE STATION MEASUREMENTS AND
MEAN AIRCRAFT MEASUREMENTS ABOVE AND BELOW THE MIXING HEIGHT
OVER ALL FLIGHTS
Air
Ground
SO,
SO,
TSP Sample size
Below mixing height
S02 . 56*
03 .4
S04
TSP
75
1* 78
. 61* 69
-.14 20
Above mixing height
S02
03
S04
TSP
.45*
.10
.19
-.60
75
78
67
6
Correlations significantly different from zero.
404
-------
TABLE 6. CORRELATIONS BETWEEN SURFACE STATION MEASUREMENTS AND MEAN
AIRCRAFT MEASUREMENTS ABOVE AND BELOW THE MIXING HEIGHT
FOR FLIGHTS GROUPED BY TIME OF DAY
Below mixing
height
a.m.
p.m.
Midnight
Above mixing
height
a.m.
p.m.
Midnight
Ground
Air S02 03 S04 TSP
S02 . 54*
03 -.07
S04 . 54*
TSP -.73
S02 . 80*
03 .49*
S04 .65*
TSP .40
S02 .71
03 -93*
S02 .69*
03 -.20
S04 . 11
S02 . 06
03 .26
S04 , . 24
S02 -.25
03 .64
Sample size
30
32
32
7
38
39
31
13
7
7
30
32
37
38
39
36
7
7
Correlations significantly different from zero.
405
-------
PASSIVE REMOTE SENSING OF S02
Millan M. Millan
Atmospheric Environment Service, Toronto, Canada
INTRODUCTION
Over the last decade or so it has become increasingly evident that
atmospheric pollutants can be transported much farther and remain airborne
for longer periods than had previously been expected. In turn, the need to
study and characterize the behavior of the polluted air parcels has resulted
in a requirement for the development of increasingly sophisticated measure-
ment techniques. Among these are remote sensors that can analyze airborne
pollution from ground, airborne, and satellite-based platforms.
The spectroscopic techniques used for this purpose can be broadly
divided into the active, which use a controlled radiation source, and the
passive, which use naturally available radiation. So far, two main instru-
mental families have emerged to the level of extended and routine operation:
lidars and correlation spectrometers. The former are inherently active
instruments, and the latter have been utilized mainly in a passive mode of
operation. Here we will examine the application of passive correlation
spectroscopic techniques. In particular, we will concentrate on the applica-
tion of the commercially available COSPEC (Barringer Research Ltd.) with
emphasis on the work carried out in the Canadian Department of the Environ-
ment.
For a more extensive review of both lidar and correlation spectrometers
as they apply to the detection of S02 in the atmosphere, the reader is re-
ferred to the review paper by Hamilton et al. (1978).
INSTRUMENTAL SYSTEM AND OPERATIONAL CHARACTERISTICS
The correlation spectrometer (COSPEC) is an instrument designed to
measure the quantity of a specific target gas (S02 or N02 in the commercial
406
-------
version) along its line of sight to a suitable source of radiation. For
passive remote sensing, the radiation source is the scattered solar light from
the sky. The instrument consists of a telescope to collect the radiation, a
grating spectrometer that disperses and displays the spectrum of the incoming
radiation onto a disc in its exit plane (Figure 1), and a photodetector with
associated signal processing electronics. Engraved on the disc are four masks
or arrays of exit slits, which have been designed to coincide alternatively
with the absorption maxima and minima of the target gas (MilIan and Hoff,
1977a; 1978). By rotating the disc, the photodetector (a photomultiplier
tube) receives a modulated signal that is processed to yield an output propor-
tional to the line-integral of the concentration of the target gas along the
line of sight.
Two calibration cells containing known amounts of S02 (or N02) can be
inserted in the optical path to provide a suitable calibration of the output.
From the point of view of data reduction and interpretation, the simplest
mode of operation is to use the sensor looking vertically upward to the zenith
sky (Millan, 1979). The sensor is usually mounted on a vehicle that travels
underneath the target gas cloud. Figure 2 shows the conceptual relationship
between the output and the plume parameters for the case of a Gaussian plume
when a sensor is used to traverse under a plume.
The operational characteristics of any passive remote sensing instrument
are, to a large extent, dominated by the peculiarities of the background. In
the case of the COSPEC measuring S02, Figure 3 shows the kind of variation in
the spectral radiance of the zenith sky as a function of the solar elevation.
The intensity of the radiation and penetration of the shorter wavelengths
increases with solar elevation, and, accordingly, follows a well-marked
diurnal cycle. The results of this diurnal variation are nominally: (a) a
zero, no-gas signal, or baseline diurnal drift; and (b) a diurnal sensitivity
change (Millan and Hoff, 1978).
The present COSPEC design allows for the cancellation or minimization of
the baseline drift (see below). The diurnal sensitivity variation remains,
however, and must be overcome by frequent instrumental calibrations. Figure 4
shows the diurnal baseline behaviour in a well-compensated sensor.
RESEARCH IN THE ATMOSPHERIC ENVIRONMENT SERVICE OF CANADA (AES)
The use of this type of remote sensing technique in the Canadian Depart-
407
-------
ment of the Environment was considered under the aegis of the following broad
objectives: (1) to conduct fundamental and applied research on the applica-
tion of remote sensing techniques to determine air pollution concentrations
and atmospheric processes related to pollution dispersal and transport in the
atmosphere, and (2) to optimize data handling methodologies for the applica-
tion of field observations to the development of diffusion models from point
and area sources.
The data collection and processing methodologies have been reported by
the author and his colleagues (Mi 11 an, 1978; Hoff et al., 1978; Mi 11 an et al. ,
1978) and will not be discussed in any detail here. For the purpose of this
presentation, it should be indicated that individual plume profiles obtained
during a preselected time period (h or 1 hr) are projected onto a plane perpen-
dicular to the average plume direction during the same period. These profiles
are averaged directly, as they occur, to yield an Eulerian average. They can
also be recentered with respect to their individual centers of gravity (COG)
and averaged to yield a fixed-distance Lagrangian average. This latter process
nominally removes from the average the meandering of the plume centerline
during the time period considered. The two averaging processes are illustrated
in Figure 5.
Present areas of research that may be of direct applicability to a PEPE
type program are those concerned with the measurement of plume diffusion from
tall stacks in the regional scale (up to ~500 km), and with the measurement
of transboundary fluxes of S02; i.e., the profiling of large and low concentra-
tion S02 clouds. These applications require sensors that are very stable and
very sensitive and that have an extended operational time. To this effect, some
of the objectives of our present work are:
1. To extend the time of operation of the COSPEC to very low, or zero,
solar elevation angles.
2. To minimize the diurnal baseline drift.
3. To investigate methods of establishing an absolute zero for S02 and
N02 burden measurements.
The first two objectives have now been successfully completed (Millan
and Hoff, 1976; 1977b) and Figure 4 shows the result of the baseline stabili-
zation method developed. Our first experience with the regional transport of
408
-------
the plume from a tall stack took place during the development of this methodol-
ogy. As Figures 6 and 7 illustrate, the departure of the baseline (Figure 6)
toward S02 readings at ~1600 hr (LSI) can be attributed to the plume from the
INCO superstack (381 m) in Sudbury, Ontario. This was established by calculat-
ing the back trajectories for the air parcel over Toronto at the time of the
observation (MilIan and Chung, 1977). These are shown in Figure 7. After
this serendipitous observation, a program was initiated to use the baseline-
stabilized COSPEC's together with appropriate trajectory forecasting to study
the dispersion of the plume from this source on the regional scale and under
similar meteorological conditions. Figure 8 shows six plume profiles obtained
on May 10, 1977, at approximately 300 km from Sudbury on Highway 9 in south-
western Ontario between Orangeville and Palmerston. Two COSPEC vehicles were
used (nearly) simultaneously for the first four profiles, and one for the last
two. The plume remains quite cohesive and meandering until approximately 1400
hr (EOT). After this time it starts spreading laterally very quickly. This
is caused, presumably, by the convection activity reaching the plume level.
This program of measurements is continuing at present.
PLUME DIFFUSION AT THE LOCAL SCALE (up to -50 km)
Characterization of plume behaviour in this scale is intended to support
the development and verification of plume diffusion models. For this work,
the COSPEC vehicle(s) is also outfitted with an S02 point monitor to delineate
the plume impingement area, or ground plume, and several units are normally
used simultaneously at various distances from the source. Figure 9, for
example, shows two half-hour Eulerian averages of seven plume profiles each.
These profiles were obtained in a lakeshore environment at approximatley 14 km
from the plant under conditions of strong and well-defined wind. Both graphs
include the averages of the vertically integrated as well as the ground S02
plumes.
Our experimental evidence indicates that under conditions of strong and
well-defined winds (i.e., near neutral conditions): (1) the averages are
well-behaved, i.e., reasonably smooth; (2) the ground plume is usually displaced
to the left of the verticallly integrated plume; and (3) this is always associ-
ated with the usual turning (veering) of the wind with height.
The situation can be quite different for the plume of the same plant
409
-------
under a well-developed fumigation due to its interaction with the thermally
induced boundary layer (TIBL). In Figure lOa, the wind backed with height
throughout the thickness of the boundary layer. For this case, the center of
gravity of the ground plume is to the right of the overall plume. As seen in
Figure lOb, 2 hr later when the wind had reverted to a veering profile, the
COG of the ground plume is again to the left of the overall plume. Some
thousand profiles with these characteristics strongly indicate that wind
directional shear effects play a very important role in plume diffusion, both
in the enhancement of plume spread and in the location of the ground plume vs.
the plume aloft.
Another example of a sheared plume propagating downwind is provided by
the profiles shown in Figures 11 and 12. These are compositions of fixed-
distance Lagrangian averages (h hr) of plume profiles obtained with three
COSPEC's operating simultaneously at increasing distances from the stack. The
time lag between the averages shown was selected to represent the time required
for a representative air parcel to travel from one distance to the next.
AIRBORNE APPLICATIONS
Initially, one can envisage two modes of operation for an airborne
COSPEC or similar instrument: upward looking and downward looking. In the
first mode, the operational characteristics and data reduction procedures are
not very different from those for a ground-based sensor. The coverage is much
faster, and traverses are no longer dependent on the availability of a suitable
road network. Information on the gas burden from the ground to aircraft
altitude, however, is lost. This method of operation has been used success-
fully to quantify the transboundary flux of S02 between Germany and the Nether-
lands (Van Egmond et al., 1978).
It is important to realize that in any type of remote sensing of atmos-
pheric components, one is confronted with a radiation transfer problem. This
indicates that, in general, the data must be processed (i.e., "inverted") in
order to retrieve the desired information. Under certain conditions, however,
the radiation transfer problem is sufficiently simple that an instrumental.
calibration is all that is required to retrieve the signal. This is mostly
the case for the upward-looking geometry of observation of a sensor like the
COSPEC, operating in the near ultraviolet (S02) or in the blue-visible (N02).
410
-------
Up to this point it has been considered that the COSPEC output is propor-
tional to the line-integral of the target gas concentration along the line-of-
sight (to the zenith sky). This assumes that the S02 lies in a clear particle-
free atmosphere and is illuminated entirely from behind. In practice, the gas
will be accompanied by particles that scatter radiation into the line of sight
from other angles. If the sulfur dioxide is distributed uniformly, this can
only increase the effective pathlength and increase the measurement of the
vertical burden. It can be shown, however, that under most atmospheric
conditions the increase is not more than 4 to 10 percent of the true burden
(Mi 11 an, 1979). The same applies to the aircraft-based upward-looking applica-
tion.
For the downward mode of operation, the data reduction is not as simple.
A fraction of the available radiation is reflected from the surface after
traversing at various angles through the atmospheric layer below the aircraft,
and the rest of the radiation is scattered upward from the particles in the
field-of-view of the sensor. Retrieval of the desired information now re-
quires the development of a suitable radiation transfer model and a data
inversion procedure, which, up to the present, has not been seriously pursued.
FUTURE DEVELOPMENTS
The development of an improved or a new generation of sensors with a
diurnal baseline drift of less than 10-15 ppm-m, and/or an absolute zeroing
technique, can be expected within 1 or 2 years. These would be particularly
useful for the measurement of transboundary fluxes with more accuracy than is
easily achievable at present. Software techniques, such as the radiation
transfer model and data inversion procedures for downward-looking airborne
application, can be developed at present for use by the current generation of
sensors. The development of this work, however, appears to be utimately
related to program requirements and may not be initiated until all the possibil-
ities of the ground-based and aircraft-based upward-looking modes of operation
are exhausted.
411
-------
REFERENCES
Hamilton, P. M., R. H. Varey, and M. M. Millan. 1978. Remote sensing of
sulphur dioxide. Atmos. Environ. 12:127-133.
Hoff, R. M., A. J. Gallant, and M. M. Millan. 1978. Data processing proce-
dures for the Barringer Correlation Spectrometer (COSPEC). 1978.
Technical Memoranda, TEC 860. Atmospheric Environment Service, Downsview,
Ontario, Canada.
Millan, M. M., and R. M. Hoff. 1976. The COSPEC remote sensor, II Elec-
tronic set-up procedures. Report ARQT-6-76, Atmospheric Environment
Service, Downsview, Ontario, Canada.
Millan, M. M., and R. M. Hoff. 1977a. Dispersive correlation spectroscopy:
A study of mask optimization procedures. Appl. Opt. 16:1609-1618.
Millan, M. M. , and R. M. Hoff. 1977b. How to minimize the baseline drift in
a COSPEC remote sensor. Atmos. Environ. 11:857-860.
Millan, M. M., and Y-S, Chung. 1977. Detection of a plume 400 km from the
source. Atmos. Environ. 11:939-944.
Millan, M. M. 1978. Remote sensing of S02, a data processing methodology.
Proceedings, 4th Joint Conference on Sensing of Environmental Pollutants,
American Chemists Society.
Millan, M. M. and R. M. Hoff. 1978. Remote sensing of air pollutants by
correlation spectroscopy--instrumental response characteristics. Atmos.
Environ. 12:853-864.
Millan, M. M., A. J. Gallant, and J. Markes. 1978. Nanticoke Environmental
Study: COSPEC/SIGN-X project. Report ARQT-6-78, Atmospheric Environment
Service, Downsview, Ontario, Canada.
Millan, M. M. 1979. Remote sensing of air pollutants: A study of some
atmospheric scattering effects. Atmos. Environ, (in press).
Van Egmond, N. D., 0. Tissing, D. Onderdelinden, and C. Bartels. 1978.
Quantitative evaluation of mesoscale air pollution transport. Atmos.
Environ. 12:2279-2287.
412
-------
INCOMING SPECTRUM
DISC
DISPERSED
SPECTRUM
4 W—•MASK
MASK
EXJT
PLANE
T Telescope
ES Entrance Slit '
M Mirrors
G Grating
D Disc
FL Fabry Lens
CL Calibration Cells
OF Optical Filter
PM Photomultiplier Tube
MASK # 4
MASKS # 2 and # 3
A
nm
300
305 310
DISPERSED S02SPECTRUM
315
300 ppm-m ABSORPTION ON A SPECTRALLY FLAT BACKGROUND
SPECTRAL ENTRANCE SLIT O.Snm WIDE
Figure 1. Schematic of the COSPEC optical layout. The masks are shown in their designed
positions with respect to the dispersed spectrum (Millan and Hoff, 1978).
413
-------
uv. radiation from
"ROAD
TRAVEUED
COSPEC output
iwte«ra.te4
concentration
output
Ground level
diffusive, term
Figure 2. Plume profiling concept using a COSPEC remote sensor to measure the overall
or vertically integrated S02 plume and an SO2 point monitor to define the
ground plume. The conceptual relationship between the COSPEC line-integrated
measurement (burden) and the plume parameters is shown for an idealized
Gaussian plume. 414
-------
CJl
~ -2
O T
1 i
o T
< fc
ff (N
E -3
I0-9
0.5
ZENITH SKY
CLEAR
h o , solar elevation
cL = 300 ppm—m
SO2
300
310
320
nm
Figure 3. SO2 transmission spectrum in the spectral region used for passive SO2 measurements, and three
typical zenith sky backgrounds in the same region (Millan and Hoff, 1978).
-------
01
-tL-7.17
396 ppm-m
7:30
8:00
10:00
OPERATIONAL RANGE 07171915=11:58*
CORRECTED RANGE 0730-1900= 11:30 h
NON-CROSSING SOLUTION
TIMES 7:30-13:15; 13:15-19:00
kt«1.12
Calibrations
395 ppm-m
BASELINE
13:15
12:00 ti =to 14:00
SOLAR NOON AT 13:18 EOT
29 AUGUST 1976
TIME CONSTANT 1. SEC
16:00
18:00
19:00
Figure 4. Example of a COSPEC diurnal baseline drift in a compensated instrument. Times ±t|_ indicate the
ends of the operational day; ±t2 are the ends of the compensated time period for minimization of,
baseline drift. The AGC curve is inversely proportional to the amount of radiation available (Millan
and Hoff, 1977b).
-------
i i i i i i |
E1ERIRN TORGE
core
RVERHGE
CD5PEC
Figure 5. Example of the Eulerian and Lagrangian averages of seven COSPEC plume profiles.
These are always presented as seen from the stack looking downwind. The Eulerian
average is the same used in Figure 9a.
417
-------
CALIBRATIONS 395ppm-m
00
»L=1812
SOUR ELEVATION
7°3l'
-•»
0711
0800 ' 1000
OPERATIONAL RANGE °>0619 to 1812
CORRECTED RANGE «0711 to 1725
MEDIAN
1200
LOCAL STANDARD TIME
1400
1600
1800
Figure 6. COSPEC diurnal baseline run of August 30, 1976, at the AES building in the northern edge of Toronto, Canada.
The departure toward SO2 readings centered at approximately 1400 hr (LST) indicates the passage of an SO2
cloud overhead (Millan and Chung, 1977).
-------
17
^23-29 i
\ !
SAULT STE MARIE
SUOBURY
20 r \
Figure 7. Air parcel back trajectories in the boundary layer terminating 1400 LST, 30 August 1976, at Toronto.
Va and Vgo represent trajectories at the levels of ~300 and 600 m, respectively, above ground (Millan
and Chung, 1977).
-------
,0.2
0.2
\ai5:lm/fls!'77
13!S:i0/0S/77
IH00M0/0S/77
IH37:i0/0S/77
1 1 1 H-
-375T00 -22500 -7500
7S00
22S00
37500
Figure 8. Six plume profiles over Highway 9, in southern Ontario on May 10, 1977. Forecast
of the approximate plume position for most efficient deployment of the COSPEC's
was provided by Y- S Chung.
420
-------
0.3 •
0.2 •
0.1 •
/-\
z:
i
a.
CL
CS
CS
0
a
in
0.2
B.l •
0
0.15-
0.1 •
0.05-
s c
/~\
2:
CL
CL
a
-2500 -1500 -500 500
ELJLER1RN RVERflGE
CD5PEC
1500
2500
Figure 9. Eulerian averages of the plume from a power plant at 13.4 km from the source. The
1/2-hr average used seven individual profiles, as shown in Figure 5. Solid lines and left
scales are for the vertically integrated SO2 average plumes, weaker lines and right
scales are for the SIGN-X ground level SO2 average plume. Vertical bars indicate the
position of the centers of gravity of each average, c for COSPEC, s for SIGN-X. Sy(c),
Sy(s) are the standard deviations of the distributions represented by the vertically
integrated and ground plumes, respectively. Time shown in brackets are the beginnings
of the first and last profiles used in the average.
421
-------
0.3 •
0.2 •
0.1 •
I
z
a
a.
ca
*
vx
n
ui
0.2 •
0.1 •
c s
-60B0
1—
-3EB0
-1200 1200
ELJLERIRN RVERREE
CD5PEC
3B00
0.3 •
/"\
2:
a.
a.
a
ui
B000
Figure 10. The same as for Figure 9 at 10.5 km from the source under fumigation conditions.
These are hourly averages of seven plume profiles. Note the irregularity of the
ground level SO2 average under the unstable (fumigation) conditions.
422
-------
0.1*
0.1 •
0.0*
0
™0
a
in
£0.is;T
a.
C30.J •
eg
eg
-0.0*T
0.1*-
0.1 •
- 35". 'OS fc«n
-f
+
-I-
51,= //3lm
2= lfc-"2
DE/BG/7B (H43- Uio)
5' _ 7/4 m
0.0S-
•
0 •
1
i I I 1 I I i I I .1
-B750 -£2*0 -1750 CDG 1750 5250
LHERRNEIRN RVERREE
CD5PEC
B750
Figure 11. Composite of three half-hour Lagrangian averages of a plume at three distances
downwind from the source. The time lag between successive averages was selected
to represent the time required to transport a representative air parcel from one
location to the next. 423
-------
-B750 -5250 -1750 CDS 1750 5250
LRGRRNE1RN RVERREE
CD5PEC
B750
Figure 12. The same as for Figure 11. In both these figures, the skewness of the average
profile is transported downwind.
424
-------
SUMMARY OF THE NASA REMOTE SENSING PROGRAM
S. H. Melfi
Office of Space and Terrestrial Applications
National Aeronautics and Space Administration, Washington, D.C.
INTRODUCTION
Over the past few years, concern with the quality of the atmosphere has
matured to one of the many important but often competing national goals.
The role of anthropogenic activities in affecting the quality of the atmos-
phere remains a complex process because complete understanding of the physics
and chemistry of the natural atmosphere requires understanding the photo-
chemistry, dynamics, and radiation budget of the atmosphere and their inter-
active processes. Improved understanding of atmospheric processes requires
a combination of atmospheric computer modeling; global, regional, and local-
ized measurements; and laboratory studies. Remote sensors capable of obtain-
ing wide spatial coverage and measuring temporal variability of gaseous and
aerosol species are valuable tools for performing measurements in the global
and regional troposphere. The National Aeronautics and Space Administration
(NASA), in response to the national concern for understanding the physics
and chemistry of the atmosphere, has developed unique satellite, aircraft,
balloon, and ground-based remote sensing instrumentation for measuring
atmospheric gaseous and aerosol species, meteorological parameters, and
earth resource parameters.
NASA has been engaged in a technology development and measurement
program to develop remote sensors for studying the global and regional
troposphere. Airborne remote sensing technology has matured to the point
that it has the potential to complement traditional measurements in Environ-
mental Protection Agency (EPA) regional field studies. Unique capabilities
of airborne remote sensors include wide spatial coverage, measurements of
temporal variability, multiple species detection (i.e., 03, S02, aerosols,
CO,. . .), and vertical ranging. Remote sensors will supplement in situ
sensors both in characterizing pollution sources and in providing the synoptic
425
-------
coverage needed to address local and particularly regional problems. Ulti-
mately remote sensors from space may provide measurements of the photochemistry
and transport of the global troposphere and provide valuable scientific
understanding of the impact of anthropogenic activity on the natural tropos-
phere.
In this paper, a summary of the remote sensing technology in the NASA
Environmental Quality Program is presented in three categories including:
1. Earth observation satellite systems that have potential application
to improved prediction of Persistent Elevated Pollution Episodes
(PEPE's).
2. Passive remote sensing technology that can be used for detection
of trace species in the troposphere on a regional scale that will
provide complementary measurements in EPA regional studies.
3. Active remote sensing technology that can be used for measuring
range-resolved and total burden of gaseous and aerosol species in
the troposphere on a regional scale that could be used to perform
complementary measurements in the 1980 EPA regional studies.
It must be realized that in studies related to the global and regional
troposphere, measurements of meteorological parameters to understand the
photochemistry and dynamics of the troposphere are important to obtain a
complete understanding of the underlying physics and chemistry of the global
and regional troposphere. This summary will not include the instrumentation
and technology developed by NASA for obtaining meteorological parameters
such as pressure, temperature, winds, and cloud cover, although identifica-
tion of these "state" variables with remote sensing technology has been
developed by NASA for application from satellite, aircraft, balloon, and
ground-based platforms.
EARTH OBSERVATION SATELLITE SYSTEMS
The occurrence of high atmospheric sulfate aerosols in the United
States has become a problem of national concern and is being given national
attention in aerometric measurement programs such as those conducted by the
EPA, the Department of Energy (DOE), and the Electric Power Research Insti-
tute (EPRI). Strong evidence has emerged in the last several years to
support the need for regional, if not subcontinental, consideration of the
426
-------
effects of sulfur pollution. Of particular interest are the regional charac-
teristics of atmospheric sulfate occurrences that are now thought to be of
more direct epidemiological consequences than gaseous S02. While a network
of ground-based pollution sensors is needed to provide accurate measure-
ments of pollutant concentrations, the localized measurements are limited to
tracking and evaluating episodes of regional dimensions. Airborne remote
sensing techniques have the potential for extending the regional coverage
requirements but are difficult to implement in a program to identify regional
pollution episodes in the formative stages. Earth observation satellites
can potentially complement the ground-based network and airborne remote
sensors by providing complete areal coverage of regional pollutant concen-
trations, identifying regional pollution episodes in the formative stages,
and tracking the episodes as they move across various regions and locales.
Therefore, the combination of satellite observations, aircraft flights, and
a network of in situ sensors should provide the necessary information for
the verification of models that will be used for the design of environmental
air quality control strategies.
NASA has developed a number of earth observation satellite systems for
various meteorological and earth/water resource applications. Table 1 is a
summary of Earth Observation Satellite Systems that have been developed for
specific applications such as the generation of high-resolution earth images
for meteorology, the measurement of sea surface and cloud top temperatures,
land use and water resource studies, and agricultural surveys. These instru-
ments are classified generically as visible/infrared spectroradiometers and
operate in spectral bands ranging from the visible (0.45 urn) through the
infrared (12.5 urn) range of the electromagnetic spectrum, in selected spec-
tral bands. Spatial resolution of the instruments ranges from 30 m to
13.6 km. In Table 1, the satellite system used as the instrument platform,
the orbits, and the launch schedule are indicated for each sensor as well as
the application and objective originally intended for the instrumentation.
Additional information on the specifications for the satellite observation
instruments summarized in Table 1 are found in references 1, 2, and 3.
A review of the current status of satellite imagery for the detection
of regional haze, with potential application to the PEPE program has been
presented in proceedings of this conference.4 5 Under sponsorship of the
427
-------
NASA Environmental Quality Program, two research efforts are currently in
progress to study the quantitative interpretation of selected satellite
imagery to measure and detect aerosol loadings in regional haze over land,6
and to measure tropospheric aerosol loadings over inland bodies of water.7
The objective of the first study6 is to perform visual and digital
analyses of satellite visible and infrared data for selected high-sulfate
episodes over the northeastern United States, and to compare results of
satellite data to ground truth information from a unique and comprehensive
aerometric data base available in Environmental Research and Technology's
data archive. Data from the NOAA/VHRR, DMSP, GOES and Landsat satellite
systems are being evaluated.
Analysis of satellite data for three selected high-sulfate episodes over
the northeastern United States has been completed. The dates of these
high-sulfate episodes are: (1) 8-10 July 1974, (2) 26 June-3 July 1975, and
(3) 3-5 August 1977. In addition to analysis of satellite data for the above
three cases, selected NOAA/VHRR and GOES imagery has been acquired and analyzed
for a suspected high-sulfate episode during the period between 19 and 24 July
1978. Figures 1 and 2 show typical examples of satellite imagery obtained
with the SMS satellite system on 9 and 10 July 1974 at 2100 GMT and identify
an area of elevated pollution (haze blob) associated with a high-sulfate
episode over the northeastern United States. Comparison of the two images
demonstrates extension of a high-sulfate episode east of the northeastern
coast over a 24-hr period. Detailed analysis of the satellite data has
shown that in three of four cases analyzed, areas of higher reflectance are
observed to correlate well with the location of highest sulfate measurements.
The synoptic weather pattern over the northeast during the period from 19-24
July 1978 is similar to other verified high-sulfate episodes. Analysis of
the preliminary 24-hr average sulfate values for eight of the nine Class I
SURE stations indicates that a regional sulfate episode did occur over the
northeast during this time period. Also, this very limited data sample dis-
plays a transport of high-sulfate levels from the Ohio Valley region east-
ward to off the mid-Atlantic coast, with maximum levels in the 30- to 40-ug/m3
range. The location of the maximum sulfate levels shows a distinct correla-
tion with the location of the patterns of increased brightness levels observed
in the satellite imagery and the regions of reduced surface visibility.
428
-------
The objective of the second study7 is to provide the capability of
quantitatively monitoring tropospheric aerosols over land masses by evaluat-
ing the relationship of upwelling near-infrared radiance over inland bodies
of water to the atmospheric aerosol optical thickness. The technique is
similar to one previously developed in the Landsat program for observations
over oceans,8 and will utilize raw data tapes from Landsat 2 MSS7 for five
inland sites for which ground truth data were obtained in the previous
Landsat study. From initial results of brightness contrasts as seen from
space, regional pollution episodes in the form of haze blobs have the poten-
tial to be detected even at the early formative stage. These results are
promising for future planning for satellite monitoring of pollution episodes.
However, these results must be tempered with the realization that other
meteorological conditions (most frequently clouds) and surface conditions
(different ground albedos) will obscure and limit the quantitative evalua-
tion of pollution episodes from satellite imagery.
PASSIVE REMOTE SENSORS
This section summarizes the current status of passive remote sensing
instrumentation for measurement of trace gaseous pollutants in the tropos-
phere. The feasibility of utilizing passive remote sensors from aircraft
and space platforms for measuring trace gaseous pollutants was established
in a study performed in 1969.9 Since that time, instruments and associated
software systems have undergone laboratory testing and aircraft flight tests
leading to the NASA approval for a passive CO sensing instrument to fly on
the second Shuttle Orbital Flight Test (OFT-2).
Passive remote sensors represent that class of remote sensors that
utilizes radiation existing naturally in the system whose properties are
being measured. For tropospheric measurements, two sources of radiation are
available: reflected solar radiation and the thermal radiation emitted by
the earth-atmosphere system. In general, at wavelengths shorter than 2.5 or
3 |jm, the primary radiation is reflected solar energy, while at wavelengths
longer than 3.5 or 4.5 urn, the primary radiation is emitted by the earth-
atmosphere system. Within the region from 2.5 to 4.5 (jm, both sources
contribute significantly and must be considered. For ground-based systems
measuring tropospheric gaseous species, a third source of radiation is
429
-------
direct solar radiation through the measurement of absorption of solar radia-
tion by the intervening atmospheric absorption lines of pollutants of interest.
Although this technique is limited in spatial coverage, measurements of this
type can be extremely important for uniformly mixed gases in which study of
transport is not a major scientific objective, or for gases whose concentra-
tions are extremely low and whose vibration-rotation lines are weak. Tech-
niques of this type, from ground-based mobile platforms, could play an im-
portant role in establishing variability of gases whose sources or sinks
depend upon the nature of the terrain (i.e., NH3). This may be important as '
a sentinel measurement to detect changes in the atmospheric absorption
spectra that can later be analyzed retrospectively for trace constituent
trends.
In establishing the feasibility of a remote sensor to detect tropos-
pheric pollutant gases, several basic parameters must be established prior
to instrument design. These general parameters include: (1) instrument
spectral bandpass, trading off interference effects of gases with overlapping
bands with total signal; (2) required noise level to achieve adequate signal-
to-noise ratios for the expected gas concentrations and atmospheric condi-
tions; and (3) for some instruments, high-resolution filter characteristics
to set the signal level and height in the atmosphere at which the gas con-
centration is being measured. In general, the major atmospheric pollutant
gases have line widths at atmospheric pressure on the order of 0.1 cm , and
because of the overlap of the pollutant gas spectra with spectra of naturally
occurring constituents, it is necessary that passive sensors have an effec-
tive resolution that is less than 0.1 cm . Further, high spectral resolu-
tion also provides the sensor with the ability to provide sharper vertical
distribution of gaseous pollutants in the troposphere. However, higher
spectral resolution must be evaluated against the loss of incoming signal
*
and the fundamental limitation of the upwelling thermal radiation from the
atmosphere to provide vertical resolution. In general, spectral bandpasses
less than 0.01 cm do not provide any significant improvement in providing
the ability to vertically discriminate tropospheric gases.
The energy received at a wavelength and altitude is a composite of the
energy emitted and absorbed by the atmosphere at all altitudes between the
source and the sensor. Because of the pressure broadening of the absorption
430
-------
lines, the energy received in the wings of the line should represent the
effect of the pollutant gas located at greater pressures (i.e., lower alti-
tude), while the energy received near the line center should be more repre-
sentative of the effects of the pollutant gas that is located at lower
pressure (i.e., high altitudes). This gives rise to the possibility that by
selectively measuring the radiation at various distances from the line
center, one could, in principle, generate vertical profiles of the gas
concentration, This is analogous to the inference of temperature profiles
by remote sensing methods using thermal infrared wavelengths where measure-
ments in various parts of the absorption band are used to effect the alti-
tude discrimination. For measurement of tropospheric pollutant gases, the
bands are generally very weak and current instrumentation is not sufficiently
sensitive to allow significant height discrimination from aircraft altitudes.
This situation is alleviated when the sensor is utilized from a space platform,
and vertical discrimination is possible although it is still difficult to
isolate the air mass less than a few kilometers above the ground (i.e.,
biosphere).
A summary of the current passive remote sensors developed under the
NASA remote sensing program is given in Table 2. The instruments have been
categorized generically into four classes of instruments: (1) Gas Filter
Correlation (GFC) techniques; (2) Interferometry; (3) Infrared Heterodyne
Radiometry, and (4) UV/VIS/IR Spectroradiometers. Characteristics of the
instruments including detectable species, spectral coverage, field of view,
and vertical resolution are shown on the table. Three viewing modes are
delineated including the thermal IR (>4.5 urn), solar reflected (2.0-4.5
urn), and direct solar (3-15 |jm). In general, instruments in the thermal IR
and solar reflected viewing modes have immediate application in aircraft and
space platforms while instruments in the direct solar viewing mode have
immediate application to ground-based platforms. Technology developments in
cryogenic cooling of sensor systems (i.e., viewing optics) and in high
quantum efficiency wideband photomixers will allow application of these
sensors to aircraft and satellite platforms in the nadir viewing mode.
For immediate application to the EPA 1980 regional field studies,
however, the aircraft MAPS (measuring air pollution from satellites) Gas
Filter Correlation instrument can be utilized to measure the dispersion of
431
-------
urban plumes from an aircraft platform. Aircraft MAPS has undergone exten-
sive testing and evaluation, and is the prototype instrument for OFT MAPS,
which is scheduled to fly on the OFT-2 Shuttle flight to measure interhemis-
pheric mixing of CO. Figure 3 represents some recent results obtained with
the aircraft MAPS instruments to identify dispersion of the urban plume from
the Chicago area, and demonstrates the ability of passive remote sensing
technology to measure total burden of CO in the atmosphere and to use atmos-
pheric CO as a tracer of the urban plume.
ACTIVE REMOTE SENSORS
NASA has been engaged in a variety of technological programs in active
remote sensing to develop, evaluate, and apply lidar techniques to perform
remote measurements of gaseous and aerosol species in the lower troposphere
from airborne platforms in order to study the spatial and temporal variabil-
ity of these species in the regional and global troposphere. Programs at
NASA have evolved from initial studies with fixed-wavelength ruby lidar
systems10 to the use of tunable and narrow bandwidth lidar systems for
remote measurements of atmospheric gases and aerosol characteristics.11
Large telescope lidar systems have been applied to the measurement of strat-
ospheric aerosols and an airborne version of the instrument is currently
being used to provide correlative measurements for the Stratospheric Aerosol
Measurement (SAM) and the Stratospheric Aerosol and Gas Experiment (SAGE II)
satellite experiments. Investigations have been conducted using Raman
scattering techniques for remote measurements of water vapor and S02 but
this technique is limited to nighttime operation, high gas concentration,
and short measuring distances. The differential absorption lidar (DIAL)
technique can overcome these limitations and has emerged as the primary
lidar technique for providing range-resolved and total column measurements
of trace gases. The NASA-developed or -sponsored DIAL systems are in various
stages of development and cover the spectral region from the ultraviolet
through the infrared. In the evolutionary lidar program at NASA, ground-
based mobile lidar systems have been applied to measurements of plume rise
and dispersion studies, model validation studies, and optical extinction
analysis through measurement of atmospheric aerosols and sulfur dioxide. As
an example, experiments were conducted by the NASA Langley Research Center
432
-------
and Old Dominion University at the Virginia Electric and Power Company
(VEPCO) plant at Yorktown, Virginia, in July 1976 and at the Morgantown
Generating Plant of the Potomac Electric Power Company, Morgantown, Maryland,
in September and October of 1976 and June 1978. Detailed discussion of the
instrumentation and experimental results through 1977 are given in reference 11.
Recent results from the 1978 studies are shown in Figure 4, in which the S02
column content (ppm-meters) measured at different times was obtained as a
function of the horizontal coordinate across the plume. The plume investi-
gated was a high stack vertical plume with the horizontal lidar scans 275 m
above the stack exit, and through comparison with correlative measurements
including the JPL, Multispectral Observation of Pollutant Species (MOPS)
sensor, feasibility of DIAL S02 measurements and system sensitivity was
established. Long-path measurements in the IR portion of the spectrum
(9-12 urn) using cooperative reflectors have also been used by the Jet Propul-
sion Laboratory to measure 03 and NO in heavily congested travel areas in
Southern California, and have provided the technology for the design and
implementation of the Laser Absorption Spectrometer remote IR lidar system
for measurement of total column 03.
Measurements from mobile ground-based platforms with prototype engi-
neering models have provided NASA with the data base to develop airborne
remote sensors for cooperative efforts with EPA in regional field studies to
complement EPA remote sensors and in situ sensors. Three airborne active
remote sensors available for cooperative efforts in the 1980 regional pro-
gram are:
1. Airborne UV-VIS-IR Differential Absorption Lidar (DIAL)
2. High Spectral Resolution Lidar (HSRL)
3. Laser Absorption Spectrometer (LAS)
A fourth system, Airborne IR Lidar System, is currently being assembled into
an airborne system and has the potential for participation in the 1980 field
studies with accelerated development. "
A summary chart of the current active remote sensors under the NASA
program is given in Table 3. Instrument parameters including detectable
species, spectral coverage, type of transmitting laser, receiver telescope
diameter, viewing mode, and platform are specified for each sensor. The
last two columns summarize current status and field application programs for
433
-------
each instrument. A summary of the four airborne active remote sensors that
are potential candidates for participation in 1980 EPA field studies follows.
Airborne UV-VIS-IR DIAL System
The airborne DIAL system was developed by NASA to simultaneously measure
the vertical and horizontal distribution of 03 and aerosols or S02 and
aerosols with the potential, with future modifications, to measure the three
species simultaneously. The present system contains two NdrYAG pump lasers
and two high-conversion-efficiency dye lasers for conducting measurements of
S02, or 03 in the ultraviolet, and aerosols in the visible spectral range.
The airborne DIAL system also has the capability to measure N02 in the
0.40-um range and H20 in the 0.70-um range. Simultaneous measurements of
the concentration and distribution of S02, 03, and aerosols and their temporal
variations are important to improve current understanding of the atmospheric
chemistry and transport of these species. Complete understanding of regional
chemistry and transport requires the rapid acquisition of concentration and
distribution of pollutants on a regional scale, requirements that are satis-
fied by the Airborne DIAL System. In addition, regional measurements of N02
can provide information on transformation, transport, and budgets in the
atmosphere with respect to photochemical oxidant problem, and with comple-
mentary measurements of total hydrocarbons (Airborne IR Lidar System) can
provide the fundamental data for improved understanding of the atmospheric
reactions involved in the photochemical oxidant production.
High Spectral Resolution Lidar (HSRL)
Measurements of the atmospheric spatial distribution of aerosol optical
extinction coefficients are needed to characterize the transport and develop-
ment of regional scale hazy air masses. Such information can be used alone
or in connection with satellite and conventional information to assess the
impact of pollutant sources on visibility and air quality. An accurate
measurement of the optical extinction coefficient is required to protect
visibility in Class I Federal Areas from manmade impairment related to air
pollution.
The HSRL has been developed to measure the vertical and horizontal
distribution of the atmospheric optical extinction coefficient, the aerosol-
to-molecular scattering ratio, and the aerosol backscattering phase function
434
-------
from aircraft platforms. The HSRL is eye safe, can be operated under day-
light conditions over centers of population, and has the potential for
providing correlative measurements for the interpretation and analysis of
elevated pollution episodes monitored by satellite imagery. The HSRL could
complement the interpretation of the EPA two-wavelength lidar for the charac-
terization of atmospheric contaminant mass loading, size distribution, and
composition.
An engineering model of the HSRL has been developed by the University
of Wisconsin under NASA sponsorship for aircraft flight testing aboard the
NASA CV-990 in the October 1979 time frame. Details of the instrument and
experiment concept are given in reference 11. Recent results obtained with
the HSRL are shown in Figures 5 and 6, in which the aerosol-to-molecular
scattering ratio and the aerosol optical extinction coefficient as a func-
tion of vertical range (km) obtained from HSRL data are shown. These data
were generated from typical daytime HSRL signals obtained from a "clear"
atmosphere at Madison,. Wisconsin, on June 10, 1978, at 0900 CDT. Airborne
data of this kind can be used to map the temporal and spatial distribution
of the aerosol optical extinction coefficient in regional scale hazy air
masses, and can be used to characterize the regional transport of aerosols
and regional variation of mixed layer depth.
Laser Absorption Spectrometer (LAS)
The LAS is a remote sensing instrument developed by the Jet Propulsion
Laboratory under NASA sponsorship and designed to interface with a twin-
engine Beechcraft aircraft. The LAS was flown in the NASA CV-990 as part of
the space simulation mission in 1977. The LAS utilizes the differential
molecular absorption technique with reflection from ground targets to obtain
the total column content of 03 from the aircraft altitude to the ground.
Two 1-W waveguide C02 lasers operating at two closely spaced wavelengths are
used for laser transmitters; one wavelength overlaps an absorption line of
the molecule of interest, and the second operates at a nearby "clear" wave-
length to discriminate interfering effects from speckle ground reflection
and atmospheric interferants (i.e., H20, C02). Heterodyne detection is used
to monitor the back-scattered laser radiation with two high quantum effi-
ciency mercury-cadmium-telluride photomixers. The present LAS instrument
435
-------
has monitored ozone column abundances and will be used in future programs to
measure atmospheric ammonia. The LAS is a prototype instrument for a poten-
tial Space Shuttle System to monitor global distributions of ozone, ammonia,
water vapor, ethylene, methyl chloride, and other species with absorption
spectra in the 9- to 12-um spectral region.
In July 1978, JPL participated in the Southeastern Virginia Urban Plume
Study to investigate the formation of 03 in parcels of air that originated
from urban areas during morning hours (Figure 7). Figure 8 shows four
measurements taken on four different aircraft passes over the same path at
different hours of the day. The wind direction was initially from the
southwest, so that the Norfolk area plume was swept up toward Wallops,
Virginia. On these plots, the diamonds are measurements performed with a
Dasibi Ozone Monitor on board the Beechcraft and the circles are the LAS
measurements. A plume of polluted air can clearly be identified at the
region near position F1 for the passes at 11:40, 3:30, and 5:45. For regional
field studies, the LAS instrument can be utilized to perform similar measure-
ments of total 03 or NH3 with the capability to obtain vertical profiles by
changing aircraft altitudes over different areas.
Airborne IR Lidar System
Measurements of hydrocarbon (HC) concentrations and spatial distribu-
tions are needed on a regional scale to determine the impact of anthropogenic
and natural sources on the production of photochemical oxidants. Such
information can be used in evaluating the relationship between HC emissions
and levels of ozone in the atmosphere. A measurement of HC transport and
ozone production would permit the assessment of current control strategies
and provide data for planning future strategies. The combustion of coal and
oil provides a significant anthropogenic source of nitrous oxide (N20).
This gas is the primary sink for ozone in the stratosphere and is third only
to C02 and H20 in its impact on the Earth's temperature due to the greenhouse
effect. Since there are no known sinks for N20 in the troposphere, measure-
ments need to be made initially on a regional scale and ultimately on a
global scale to determine the fluxes of N20 being emitted into the atmosphere
from anthropogenic and natural sources (e.g., lightning).
436
-------
Several key components of the IR DIAL system, including the IR laser
constructed by Westinghouse and the IR detection system, have been developed
under the previous Energy-Related Remote and In Situ Instrument Development
Program (EPA Project 80 BLEU). The IR laser has the capability to be tuned
from 1.4 to 3.9 urn. This permits measurements to be made of HC near 3.4 urn
and N20 near 3.85 urn. The IR laser is eye safe, and it can be operated in
the daytime. In conjunction with the measurement of ozone in the troposphere
by the Airborne DIAL System, the relationship between the distribution and
transport of HC and ozone production could be studied in support of EPA's
regional modeling effort. In addition, the IR DIAL will have a high sensitiv-
ity for the measurement of HC1, which has been detected in coal field power
plants and in vinyl chloride flares.
SUMMARY
NASA has been engaged in the development of remote sensing instrumenta-
tion for performing scientific measurements of the quality of the earth's
air, land, and water resources. This technology has matured to the point
that satellite measurements with application to earth and ocean resources
are being performed routinely. Additional measurements are taken to deter-
mine meteorological parameters in the stratosphere and troposphere, and to
study the photochemistry, transport, and radiation budget of the stratosphere.
An important gaseous species in determining the quality of the natural and
polluted troposphere, carbon monoxide will be measured in an early Space
Shuttle flight to determine its interhemispheric mixing ratio, important in
understanding the impact of anthropogenic activity on the sources and sinks
of global CO.
This paper has summarized the current NASA Remote Sensor Technology for
studying the global and regional troposphere from space, aircraft, balloon,
and ground-based platforms in the NASA Environmental Quality Program.
As part of its program to develop remote sensors for utilization from
space, NASA has developed a series of passive and active remote sensors that
have undergone extensive testing from aircraft platforms. Four of these
sensors, including aircraft MAPS, UV/VIS DIAL, HSRL, and LAS, can be imple-
mented to supplement in situ and remote sensors (developed by EPA) in the
1980 regional field test program. A fifth lidar system, the Airborne IR
437
-------
Lidar System, is currently being assembled into an aircraft system and with
accelerated development has the potential for participation in the field
test program.
Similarly, earth observation satellite systems have demonstrated the
ability to detect regions of elevated pollution and have the potential to
supplement current in situ and aircraft systems in identifying elevated
pollution episodes in the formative stages.
REFERENCES
1. Goddard Space Flight Center. LANDSAT Data Users Handbook. Document
no. 76SDS 4258 September 2, 1976.
2. Hussey, John W. The TIROS-N Polar Orbiting Environmental Satellite
System. National Oceanic and Atmospheric Administration, U.S. Depart-
ment of Commerce, Washington, D.C., October 1977.
3. Remote Sensing Handbook. Marshall Space Flight Center, June 30, 1977.
4. Husar, Rudolf B. Identification of PEPE's by Visibility Isopleths.
Proceedings of the Seminar/Workshop on Persistent Elevated Pollution
Episodes, March 19-23, 1979, Environmental Sciences Research Laboratory,
Research Triangle Park, N.C.
5. Lyons, Walter A. Satellite Measurements of PEPE's. Proceedings of the
Seminar/Workshop on Persistent Elevated Pollution Episodes, March 19-23,
1979, Environmental Sciences Research Laboratory, Research Triangle
Park, N.C.
6. Barnes, James C., and W. Ayers. Study to Evaluate the Capabilities of
Satellite Imagery for Monitoring Air Pollution Episodes. Environmental
Research and Technology, Inc., and NASA, Langley Research Center,
Hampton, Virginia.
7. Griggs, M., and W. Ayers. Satellite Measurements for Tropospheric
Aerosols. Science Applications, Inc., and NASA, Langley Research
Center, Hampton, Virginia.
8. Measurement of Atmospheric Aerosol Optical Thickness Using ERTS-1 Data.
J. Air Poll. Contr. Assoc., 25:622, 1975. Comment on Relative Atmos-
pheric Aerosol Content From ERTS Observations. J. Geophys. Res. ,
82:4972, 1977.
9. Ludwig, C. B., et al. Study of Air Pollutant Detection By Remote
Sensors. NASA CR-1380, July 1969.
10. Lawrence, J. D., et al. Optical Radar Studies of the Atmosphere.
Proceedings of the Fifth Symposium on Remote Sensing of the Environment,
Ann Arbor, Michigan, April 16-19, 1968.
438
-------
11. Browell, E. V. Lidar Remote Sensing of Tropospheric Pollutants and
Trace Gases-Programs of NASA Langley Research Center. Proceedings
of the 4th Joint Conference on Sensing of Environmental Pollutants,
American Chemical Society, 1978.
439
-------
Figure 1. Satellite imagery with SMS satellite system,
July 9, 1974, @ 2100 GMT.
440
-------
00 191:74 Ol-tt-2 0100 0998 C2 MIA
Figure 2. Satellite imagery with SMS satellite system,
July 10, 1974, @ 2100 GMT.
441
-------
ro
W; ft£3^f^P^ras ^ -$^r" X• •»
Figure 3. Dispersion of urban plume from Chicago area using
the A/C MAPS instrument
-------
S02LIDAR MEASUREMENTS ON 6/22/78
2000
co
e
9
0>
I
Q.
a.
UJ
8
O
o
1500
1000
500
TIME OF SCANS
22.53
22:54
22:56
22:57
VERTICAL PLUME WITH HORIZONTAL
LIDAR SCANS 275 m ABOVE STACK
EXIT
200 150 100 50 0 -50 -100
HORIZONTAL COORDINATE ACROSS PLUME (METERS)
(REFERENCED TO CENTER OF STACK)
-150
-200
Figure 4. Sulfur dioxide column content measured at different times
as a function of horizontal coordinate across plume.
-------
HSRL AEROSOL DATA
2.5r
- 2.0h
g
5
ac
rr
LU
5
O
1.5
I.Q
I I
I I I
•T 0.12
0.10
0.08
.
LJ
O
0 0.06
o
0.04
x 0.02
LJ
02468
VERTICAL RANGE (KM)
O.OQ
I I I I I
0246
VERTICAL RANGE (KM)
J
8
Figure 5. Aerosol-to-molecular scattering ratio Figure 6. Aerosol optical extinction coefficient
as a function of vertical range (km) (HSRL). as a function of vertical range (km) (HSRL).
444
-------
cn
0 20 KM
Figure 7. Virginia urban plume study geography.
-------
I 1-1
LAS OZONE MEASUREMENTS, LEG E-F
BEECHCRAFT ALTITUDE-3500ft.
JULY 21, 1978
150
2 ioo
a.
r>
~ 50
O
h- 0
• •
^^ i&
^j
0 S ^
o^oo S
— —
< E, p, 11:40 AM E, p, 5:45 PM
o 150
* ioo
^Cv
u 50
o
N n
— —
§
_ o &<>0_
— —
OOOOgDQOo^-
_ o o_
O LAS
O DAS IB! IN SITU DATA
Figure 8. LAS measurements of total burden ozone in the
southeastern Virginia plume study.
-------
TABLE 1. EARTH OBSERVATIONS SATELLITE SYSTEMS
EARTH OBSERVATIONS SATELLITE SYSTEMS
CATEGORY AM) INSTRUMENTS
o USSR
VISIBLE SPIN SCAR
RADIOMETER
0 VAS
(VISSR ATMOSPHERIC SOUNDER)
o AVHRR
(ADVANCED VERY HIGH RESOLUTION
RADIOMETER)
0 NSS
(MULTI-SPECTRAL SCANNER)
0 REV
(RETURN BEAN VIDILON CAMERA)
0 NSS
(S-BAND NULTI -SPECTRAL
SCANNER)
o RBV
(RETURN BEAN VIDICON CANERA)
0 THEMATIC MAPPER
SPECTRAL COVERAGE
(MICRONS)
IR IMAGES (10.S - 12.6)
VIS IMAGES .55 - .7$
VIS IMAGES .55 - .75
VIS INAGES .55 - .75
VIS INAGES .55 - .75
IR 1 VIS IMAGING IDENTICAL
TO VISSR WITH ADDITIONAL
SOUNDING CHANNELS AT
IR 3.73 - 8.0)
IR 8.0 - 14.7)
VIS INAGES 0.55 - 0.70)
NEAR IR INAGES 0.72$ - 0.90)
IR INAGES 3.5S - 3.93)
IR INAGES 10.5 - 11. S)
VIS O.S - 0.6
IR 0.6 - 0.7
IR 0.7 - 0.8
IR 0.8 - 1.1
8LUE-GRTEN (.475 - .$75)
ORANGE-RED (.580 - .680)
NEAR IR (.690 - .830)
SANE AS LANDSAT 1 1 2 WITH
ADDITIONAL NSS CHANNEL
IR (10.4 - 12.6)
VIS .IS - .«
VIS .52 - .60
VIS .63 - .69
IR .79 - .901
15 ibfi : 1*3
SPATIAL RESOLUTION
(KM)
8
4
2
1 and 2
1 and 2
16 KN
1.1
1.1
1.1
1.1
.08
.08
.237
.03
.03
.03
.03
.03
.120
SATELLITE
SYSTEN
SNS/SOES 1-3
GOES D/E/F
(NOAA REIMBURSABLE)
TIROS N
A THRU G
LANDSAT 1 1 2
LAKDSAT 3
LANDSAT-D
ORBIT
GEOSYNCHRONOUS
GEOSYNCHRONOUS
SUN SYNCHRONOUS
SUN SYNCHRONOUS
910 KM
SUN SYNCHRONOUS
917 XH
SUN SYNCHRONOUS
LAUNCH
SCHEDULE
SNS-1 SNS-2 SNS-3
5/74 2/75 10/7$
GOES-2 BOES-3
6/77 6/78
D E F
7/80 2/81 7/83
TIROS N
10/78
NOAA
A B* C«
4/79 79 80
D« E* F» G»
81 82 83 84
•AS NEEKD
1 2
7/72-1/78 1/75
V78
3Ti QUARTERS
81
APPLICATIONS/OBJECTIVE
0 CLOUD INAGES
o SEA SURFACE TEMPERATURE
0 HIGH RESOLUTION INAGES
o ATMOSPHERIC SOUNDING
0 SEA SURFACE TEMPERATURE
0 CLOUD TOP TEMPERATURES
0 METEOROLOGICAL HIGH RESOLUTION
EARTH INAGES
0 CLOUD INAGES
0 SEA SURFACE TEMPERATURE
0 HIGH RESOLUTION IMAGES
SOLAR REFLECTED
0 LAND USE
0 HATER RESOURCES
o AGRICULTURAL SURVEY
SOLAR REaECTED
0 LAND USE
0 MATER RESOURCES
0 AGRICULTURAL SURVEY
SOLAR REFLECTED RADIATION
o LAND USE
0 MATER RESOURCES
0 AGRICULTURAL SURVEY
-------
TABLE 2. PASSIVE REMOTE SENSORS
PASSIVE DENOTE SENSORS
CATEGORY AND INSTRUMENTS
GAS FILTER CORRELATION (6fC)
• A/C MAPS
• On WPS (SHUTTLE)
• PHI
• OCR
INTERFEROMETRV
• COPE
• CIMTS
• HSI
INFRARED HETERODYNE
RAoiOMETRY
• IHR
• INS
SPECTRORAD10METERS
• SBUV - TOMS
OF POLLUTANTS (HOPS)
SPECIES
CO. CH.
Entwnteble to Other Eases
CO
Exttndrtl* to Other G»et
CH4. CO. NjO. W3
CHi. NHa. CO. N»0.
NCI* N02. 502. (Stic)
CO. 0)4
•,. H^N^CO. CH,
SPECTRALLT SCANNING
03. NH3
SPECTRAL COVERAGE
(CONTINUOUS)
12 SELECTED* CHANNELS
SO,. 03. NO,
SPECTRAL COVERAGE
MICRONS
4.52 - 4.8
4.52 - 4.8
4. 52 - 4.8
8.00 - 11.0
4.0 - 9.0
2.35
2.0 - 2.4
4.0 - 9.0
1.0 - 6.0 \m
9 - 12 im
(DISCRETE STEPS MITH
C02 LASER)
7.5 - 13.0 I*
(0.25 - .34)
(O.X - .CO)
FIELD OF VIEW
(DEGREES)
7.5°
4.5°
4.5°
2-70
7»
2-7°
20
0.25 mr
0.25 ••
f
MULTI-SPECTRAL
IMAGING
VERTICAL
RESOLUTION
(KM)
<5
-------
TABLE 3. ACTIVE REMOTE SENSORS
ACTIVE REMOTE SENSORS
CATEGORY AM INSTRUCT
FIIEO MAVELENGTH
1 LAKE TELESCOPE LIOAR STSTCN
• VC STRATOSPHERIC AEROSOL IIOM
STSTEM
• END-BASED RUBY LIDAR STSTEM
• LASER ABSORPTION SPECTROMETER
(LAS)
• HIGH SPECTRAL RESOLUTION LIOAR
(HSRL>
TUNABLE MAVaENGTH
• U» DIAL
• NEAR 1R DIAL
• MID IR OIAL
t A/C OIAL STSTEM
* DIFFERENTIAL AISORPTIOR POLLUTION
SENSOR
• SHUTTLE HOAR
SPECIES
STRATOSPHERIC
AEROSOL'S
STRATOSPHERIC
AEROSOL'S
TMPOSPHERIC
AEROSOL'S
0,. NH,
AEROSOL EXTINCTION
c
*&,&*•
HjO. PRESSURE.
TEMPERATURE PROFILES
CO, NCi. OU, CO;.
•¥>• "jo-"^. HZS
(NPKJ
AEROUL'i! TEWERJTURE.
PRESSURE
NjO. CO. NH,
MULTIPLE SPECIES^
aOUOS. AEROSOL'S.
NET PARAMETERS
SPECTRAL RANGE
(MICRONS)
.693
.693
.693
9-12
(OISCRETELV TUNABLE)
.4- .45
.28-. 31
.70-.7S
1.4-3.7
.28-. 9
.45-. 60
7.S-13.0
.30-13.0
TRANSNiniNB
LASER
RUBY
RUBT
RUBT
CO. MAVEGUIDE
' LASER
»2 - PUMPED
OTE
Nd:VA6
PUMPED DTE
RUBT
PUMPED DTE
OPTICAL
PARAMETRIC
OSCILLATOR
Nd:YA6
PUMPED
DTE
TUNABLE
DIODE
LASERS
IVOLUTiONART
HOAR
STSTEM
RECEIVER
TELESCOPE OIAN.
(CM)
122
36
31
IS
36
31
SI
36
36
6
12S
VIEUING
NODE
ATMOS.
BACKSCATTER
ATMOS.
BACKSCATTER
ATMDS.
BACKSCATTER
REFLECTED END
(NDLECULAR
ABSORPTION)
ATNOS.
BACKSCATTER
DIFFERENTIAL
ABSORPTION FROM
ATMOS. BACKSCATTER
DIFFERENTIAL
ABSORPTION FROM
ATMOS. BACKSCATTER
DIAL.
GND REFLECTION
DIFFERENTIAL
ABSORPTION FROM
ATNDS. BACKSCATTER
LASER
ABSORPTION IN
20 NETER PATH
ATMds7lACKSCATTER~
FLUORESCENCE
DIAL
PLATFORM
MOBILE
GND
VC
MOBILE
GNO
VC
VC
MOBILE
GNO.
GNO/VERTICAL
PROFILES
VC
VC
VC
SPACELAB
CURRENT
STATUS
OPERATIONAL
OPERATIONAL
OPERATIONAL
OPERATIONAL
SYSTEM ASSEMBLY
FOR 8/79 RGT.
(CV-990)
OPERATIONAL
OPERATIONAL
SYSTEM ASSENBLT
FOR 8/79 a ST.
STSTEM ASSEMBLY
FOR 8/79 FIST.
OPERATIONAL
- -PHASE—
STUDY
COMPLETE
Fiao
APPLICATIONS
SATELLITES
(SAN II. SAGE)
• VOLCANIC ERUPTIONS
• GNO TRUTH FOR
SATELLITtS
(SAN II. SAGE)
• POMER PLANT
PLUME STUDIES
• TOTAL BURDEN
• REGIONAL AEROSOL
EXTINCTION PROFILES
• REGIONAL/URBAN
PLUME STUDIES
• TROPOSPHERIC
METEOROLOGICAL
PARAMETERS
• REGIONAL (HC)
t TROP. SPECIES
SURVEY
• REGIONAL STUDY OF
0,. SO,. AEROSOL'S.
TROP. AjO PROFILES
• REGIONAL STUDY
OF TROPOSPHERIC
BUDGETS t DYNAMICS
-TaoBAT
TROPOSPHERE
STRATOSPHERE
NESOSPHERE
THERMOSPHERC
-------
VALIDITY OF BALLOONS FOR AIR PARCEL TRACKING
Frank B. Tatom*
Engineering Analysis, Inc., Huntsville, Alabama
George H. Fichtl
NASA/George C. Marshall Space Flight Center,
Marshall Space Flight Center, Alabama
Abstract
Constant-volume balloons (CVB's), tetroons, and similar types of bal-
loons are commonly used for measuring atmospheric winds.1 These balloons
are assumed to follow the mean wind and thus represent a logical choice for
tracing air parcel trajectories during "persistent elevated pollution epi-
sodes" (PEPE's). Various aerodynamic and thermodynamic forces, however, may
interact with the balloon and cause it to deviate from the mean wind path.
The discussion that follows is divided into three separate areas:
Examples of tetroon behavior.
Analytical studies of CVB motion.
Displacement of CVB's due to the vertical component of wind.
This material is designed to provide a better insight into the question of
the validity of balloons for air parcel tracking.
EXAMPLES OF TETROON BEHAVIOR
A large amount of tetroon tracking data has been collected covering a
wide range of altitudes and conditions.2 Two examples of such data are
provided in Figures 1 and 2. These radar tracks of tetroons were collected
by NASA in August 1967 at Cape Kennedy.
The first tetroon track (Figure 1) corresponds to a morning launch
commencing at 0800 EST. As shown in the altitude history plot, the tetroon
ascended to an altitude of roughly 425 m and generally remained at that
^Speaker.
450
-------
altitude for a major portion of the track. Vertical oscillations appear to
have a period of about 20 min with an amplitude of 25 m. The direction and
speed of the balloon remained relatively constant with respect to time.
The second tetroon track (Figure 2) corresponds to a late morning
launch (1100 EST). The altitude history is characterized by sizeable verti-
cal oscillations with amplitudes as great as 200 m and time periods on the
order of 16 min. The corresponding direction history is more erratic than
for the first case, but the mean direction of drift is roughly the same.
The speed history for the second case is also more erratic, with a magnitude
that is approximately twice that observed in the earlier launch.
These two examples illustrate the general observation that atmospheric
conditions are generally more stable in the early morning than later in the
day. The amplitude and period of the observed oscillations are also of
interest as subsequently discussed.
ANALYTICAL STUDIES OF CVB MOTION
As part of an earlier study2 of CVB's, the computer program BALLOON
was written to provide a means of predicting the response of a CVB to a
three-dimensional transient flow field.
Basic Governing Equations
The motion of a CVB is governed by the conservation of momentum
equation, which in dimensionless form can be written as
v (u. - v.) + J
cno (u, - v.) |u -
jo. i o j. ^ / ' ' 4- \ o j. ~ / LIX, i i •
. 9M --,
(v . Q )_L . 2(1- o) g
, .
— dt'
-18 V-S- / —~ (i = 1,2,3)
451
-------
This equation is numerically integrated by means of a fourth-order Runge-
Kutta technique to obtain the response of the CVB to the flow field. A
periodic flow field model was used to represent the atmosphere. The model,
which is nonhomogeneous, transient, and three-dimensional with mean transla-
tional motion, has the form
u. = Oj 5n + sin [ k (xj_ + x2 + j<3) . %i + 9. ] (1 = 1,2,3) (2)
By selecting the 3 values of 6. spaced 120° apart, the model satisfies the
continuity equation.
Natural Period of Oscillation
Based on first-order theory,3 the natural period of oscillation of an
air parcel is
TB.V.
The reciprocal of this expression is generally referred to as the Brunt-
Vaisala frequency. The corresponding period of oscillation for a CVB,4 is
TCVB
The adiabatic lapse rate, r , has a value of 9.76° K/km while the constant
density lapse rate, r , has a value of 34.2° K/km. Examination of equations
(3) and (4) indicates that the natural period of oscillation of the CVB will
always be less than that of the air parcel. For example, with an equilibrium
temperature, T , of 300° K and a lapse rate, r, of 6° K/km,
TB v = 566.76 sec or 9.45 min
452
-------
while
= 206.95 sec or 3.45 min
The periods of oscillation of the tetroons discussed previously appear more
representative of the natural frequency of an air parcel than of a balloon.
Taylor's Hypothesis
In the course of the earlier investigation, the relative magnitude of
three velocities was observed to be of special significance. These three
velocities were the mean wind, ult the mean balloon velocity, vls and the
phase velocity, c, where
c = i/ic (5)
With respect to these three velocities, two versions of Taylor's Hypothesis
were identified. The first, which is Taylor's hypothesis for the fluid
(THF), is satisfied when the mean wind velocity equals the phase velocity or
^* -*•
ul = c
The second version, which is Taylor's hypothesis for the balloon (THB), is
satisfied where the mean balloon velocity equals the phase velocity or
^ ***
vl =c
Numerical Prediction
A series of balloon trajectories was generated with the program BALLOON.
Table I presents typical input values. The outputs included the position
and velocity of the balloon as a function of time and also the velocity of
the wind at the same position and time.
The results were examined in terms of both the periodic motion and the
mean motion of the balloon. With regard to the mean motion it was observed
that the balloon displayed mean transverse motion in the horizontal plane
453
-------
when THB was not satisfied. This feature is illustrated in Figures 3, 4,
and 5. In addition to transverse motion, the mean Xi-velocity component of
the balloon did always match the mean wind velocity Uj. When the Xi-component
of the mean wind velocity greatly exceeded the phase velocity (Ui»c), the
balloon mean Xx-velocity component generally was slightly less than the wind
velocity component (vj < Ui), as shown in Figure 6. However, as c increased
relative to ul5 \/l also increased until, as THF was approached, vx first
equaled and then exceeded Uj. Subsequently, when THF was satisfied, Vj
exceeded both ul and c. For a value of c slightly greater (~5 percent) than
UL THB was satisfied, with c becoming equal to vx. Further increases in c
resulted in a reduction in Vj until, for cases where c was very large compared
with ulf vx became essentially equal to Uj.
VERTICAL DISPLACEMENT OF CVB's DUE TO THE VERTICAL COMPONENT OF WIND
In the presence of a steady vertical velocity, the CVB will be displaced
from its normal equilibrium altitude. The new equilibrium altitude is
reached when two forces, drag and buoyancy, are balanced. In dimensionless
form the relationship between the vertical velocity and the displacement is
Hi
W = [(1-yZ) Y - 1]1/2 (Z^O) (6)
Hi
W = -[1 - (l-YZ) Y ]1/2 (Z < 0) (7)
where
W =
Z =
R T
o
Y =
T
454
-------
For values of y ranging from 0 to 0.287, the variation of W with Z is present-
ed in Figure 7. The dimensionless form of this plot, while useful because
of its general nature, does not permit immediate appreciation of the magni-
tude of typical dimensional parameters. For the case of a lapse rate of 6°
K/km and a balloon with a 1-m radius and a drag coefficient of 0.75, the
corresponding dimensional plot is shown in Figure 8. This figure clearly
indicates that a steady vertical velocity component of 2 m/s could cause a
displacement of ~1,200 m. If the conditions associated with the second
tetroon track (Figure 2) were similar to those in Figure 8, the vertical
velocity component involved would be on the order of 1 m/s.
SUMMARY
The preceding discussion provides a description of certain effects
produced by aerodynamic and thermodynamic forces that might cause a tetroon
or CVB to deviate from the motion of the mean wind. Based on such material
there is reason to believe that under certain conditions the balloon velocity
may deviate from the mean velocity in both magnitude and direction. In
addition, the balloon may be displaced a significant distance from its
equilibrium altitude due to vertical currents. None of the material presented,
however, is based on a flow field with the random characteristics of atmos-
pheric turbulence. Thus, these conclusions must be applied with caution.
To fully resolve the issues would involve coupling the BALLOON program with
a three-dimensional simulation of atmospheric turbulence.
LIST OF SYMBOLS
English Symbols
Symbol Definition
A amplitude of periodic flow field model
A area of balloon cross section
c phase velocity, uu/k
c dimensionless phase velocity, iu/k
CD£ balloon drag coefficient (103
-------
vi
V
w
X30
xj
IN*
zJ
Greek Symbols
Symbol
Y
r
r
9i
V
rv
V
P
rv
P
a
T
U)
ui
Definition
wave number
dimensionless wave number, kD
specific gas constant
time
dimensionless time, At/D
fluid temperature at equilibrium altitude
fluid velocity
dimensionless fluid velocity, u./A
balloon velocity
dimensionless balloon velocity, v./A
volume of balloon
dimensionless vertical veloci
equilibrium altitude
Cartesian coordinate
dimensionless Cartesian coordinate, x./D
J
dimensionless vertical displacement, g(x3-x30)/(RT )
Definition
dimensionless lapse rate, FR/g
atmospheric lapse rate
adiabatic lapse rate
constant-density lapse rate
Kronecker delta function
phase angle of flow field model
kinematic viscosity
dimensionless kinematic viscosity, v/(AD)
fluid density
dimensionless fluid density, p/a
balloon density
period of oscillation
frequency of flow field model
dimensionless frequency of flow field model, uuD/A
456
-------
REFERENCES
1. Tatom, F. B., and G. H. Fichtl. The Response of a Constant-Volume Balloon
to Periodic Three-Dimensional Flow. Seventh Conference on Aerospace and
Aeronautical Meteorology and Symposium on Remote Sensing from Satellites.
November 16-19, Melbourne, Florida.
2. Tatom, F. B., and R. L. King. Determination of Constant-Volume Balloon
Capabilities for Aeronautical Research. NASA CR 2805. George C. Marshall
Space Flight Center Space Sciences Laboratory, Marshall Space Flight
Center, Alabama, May 1976.
3. Hess, S. L. Introduction to Theoretical Meteorology. Holt, Rinehart
and Winston, New York, 1959.
4. Angell, J. K., and D. H. Pack. Analysis of Low-Level Constant Volume
Balloon (Tetroon) Flights from Wallops Island. Journal of The Atmos-
pheric Sciences, 19:87-98, 1962.
457
-------
en
00
80O
700
600
« 500
400
300
200
IOO
T = 20 min
-I
TIME OF LAUNCH
0800 EST
10 20 30 40 50 60 70
Tim* (m i n)
80
90
100 110
120
Figure 1a. Tetroon altitude history (morning launch).
-------
Ul
CO
360
* 270
>80
o
a*
i 090
360
20
12
•o
01
01
Q.
tn
TIME OF LAUNCH
0800 EST
10
20
30
40
80
SO 60 70
Time ( m i n )
Figure 1b. Tetroon direction and speed history (morning launch).
90
IOO
no
I2C
-------
1 O UU
1400
1200
*« 1000
IA
>*
E
*M*
« 800
•o
3
•-
«-
cn < 600
o
400
200
O
.""•
* •
.-•'• ' *
•
•
•
•
• •. .-
• • •
• •.
•f.
»•
* • •
•
•.••
-
•
.•
•
..—
h"1
i
i
*••
•
•
6 min —
.1.
•
/"
• .**•
.•'
•
•
TIME OF LAUNCH
1100 EST
10
20
30
40
80
50 60 70
Time (min)
Figure 2a. Tetroon altitude history (late morning launch).
90
100
110
120
-------
360
ti
k.
01
o*
c
o
270
180
u
01
.i: 090
O
360
20
16
u
w
-------
1.0
345
DIMENSIONLESSTIME.7
Figure 3. Transverse balloon velocity for case T-j (v < c).
462
-------
-0.1
OIMENS1ONLESS TIME, t
Figure 4. Transverse balloon velocity for case "\2 (v =
463
-------
1.0
-0.1
6
12345
DIMENSIONLESS TIM6.T
Figure 5. Transverse balloon velocity for case 73 (v > c).
464
-------
1.0
THE
THF
_ / >±r^-^.=-
c-u.
1.0
c/u,
Figure 6. General variation of mean balloon velocity with phase velocity
based on numerical results.
465
-------
0.
to
4/1
Ul
o
00
-0.4 -0.3 -0.2 -0.1 0 0.1 0.2 0-3 0.4
DIMENSIONLESS VERTICAL VELOCITY
Figure 7. Dimensionless variation of vertical displacement with vertical velocity.
466
-------
1500-
1000-
500-
Q.
oo
i—*
Q
I
-------
TABLE 1. TYPICAL DIMENSIONAL VALUES
Physical parameter Value
TQ 218° K
R 2.870224 x 106 cm2/s2 °K
T 0.0° K/cm
g 980.6 cm/s
H 1.40646 x ID'4 g/cm s
A 100.0 cm/s
D 130.0 cm
uj 1000.0 cm/s
a 3.61494 x 10~4 g/cm3
pQ 3.61494 x 10-4 g/cm3
k 6.28 x IQ-6 to 6.28 x 10'3 cm'1*
u> 6.28 x ID'3 to 6.28 x 10"1 sec'1*
*Such ranges of k and tu correspond to spatial wavelengths from 10 m to
10 km and time periods of 1 s to 1,000 s.
468
-------
NAVIGATIONAL AIDS FOR AIR PARCEL TRACKING
James G. Haidt
Research Triangle Institute, Research Triangle Park, North Carolina
Abstract
The purpose of this paper is to discuss the application of radio navigation
systems to the tracking of tetroons used as air parcel markers. It is concluded
that an acceptable tracking system could be designed that used either Omega or
Loran-C but that several problems, both technical and logistical, remain to be
faced.
RETRANSMISSION CONCEPT
The various means by which one might follow the progress of a tetroon, as
it is carried along in a parcel of air, divide themselves naturally into two
categories according to whether or not the tetroon payload includes electronic
gear devoted specifically to the tracking function. For tetroons that include
no such electronic gear, tracking techniques rely on sighting the tetroon either
visually or with a primary radar; techniques of this sort are not discussed
further here. Tracking systems in the complementary "electronic" category,
those which are of interest here, are conveniently distinguished according
to the type of electronic payload carried by the tetroon. Figure 1
illustrates three generic classes of electronic instrumentation: beacon,
transponder, and receiver/retransmitter (or simply retransmitter).
In the beacon class of tracking systems, the tetroon carries a transmitter
that periodically broadcasts a modulated signal. A receiver monitoring these
broadcasts can then determine the position of the transmitter along with its
identity. The beacon may also be equipped to transmit data in its broadcasts.
The primary example of a beacon tracking system is that which employs satellites
469
-------
(e.g., Tiros-N) as the monitoring receivers. Other monitoring platforms can
be used in principle, subject to the constraints that they achieve line-of-
sight communication with the transmitter and that they know time and their
position and velocity accurately.
The transponder class of systems carries a receiver/transmitter pair that
responds to electronic interrogation from a radar. In this case, the position
of the transponder is determined in the standard radar way (antenna-pointing
direction and time delay in the transponder reply to interrogation). Position
determination of the transponder obviously requires that the position of the
radar be known. As in a beacon system, the transponder may relay data (e.g.,
identification code) in its reply.
In the retransmitter class of systems a distinctly different approach is
taken to positioning the tetroon. Here the electronic payload includes a
navigation receiver, by means of which the tetroon effectively determines its
own position, and a transmitter, which relays this information to a monitor
receiver. It follows, therefore, that the monitor need not know its own
position and may, in fact, be moving, an important consideration in assessing
the ease with which tetroons may be tracked. Along with position data, the
navigation retransmitter may pass along other data in a natural way.
The retransmitter principle of tracking is depicted in greater detail in
Figure 2a. The standard radio navigation receiver, as shown, consists of
three main elements that have been labelled demodulator, processor, and tracker
for convenience. The demodulator element is meant to include all of the radio
frequency (RF) gear of the receiver, i.e., antenna, RF amplifier, etc. It is
responsible for accepting the signals appropriate to the navigation system in
use (e.g., Omega) and extracting from these signals the raw data (e.g.,, phase
differences) from which position can be determined. The actual conversion
of raw data into position data is performed by the second element of the
receiver, the processor. The periodic position updates so produced are fed
into a tracker where they are assembled into a smooth trajectory.
Adaptation of the standard navigation receiver to the retransmitter
concept involves breaking it as shown in Figure 2b and inserting a transmitter/
receiver pair in the gap. In the new configuration, the demodulator and
470
-------
transmitter comprise the tetroon payload while the receiver, processor, and
tracker become part of the monitor receiver. Clearly the original navigation
receiver could be divided at some other point besides that shown in Figure 2b
to achieve the same end, but minimization of the electronics carried by the
tetroon dictate the arrangement shown.
CANDIDATE RADIO NAVIGATION SYSTEMS
Tracking Requirements
It is not the intent of this paper to lay out an experimental design from
which a list of requirements concerning a tetroon tracking system would be
forthcoming. To provide a basis for discussion, however, it is useful to have
at least a list of desirable attributes that the ideal tracking system might
possess. Such a listing is advanced in Figure 3.
The parameters displayed in Figure 3 must be viewed, to reiterate, more
as being desirable than necessary. Indeed, it is highly unlikely that any of
the tracking systems mentioned earlier can realize all of them in practice.
For example, an hourly position update rate will not be possible with the
satellite set currently in orbit or projected to be operational in the near
future. Likewise, neither a radar nor a retransmitter system is capable of a
faithful hourly update rate over extended periods of time without sizeable
expenditures of either manpower or money.
These and similar caveats should be borne in mind in the following dis-
cussion, the point being that some compromises will need to be made.
Navigational Systems
Table 1 lists three candidate radio navigational systems that could be
used in a retransmitter tracking system. It should be noted at the outset
that both Omega and Loran-C are currently operational; the last, GPS, is in
the process of being implemented and tested but will probably not be available
for routine operation until near the middle of the next decade. It is included
in the lineup for completeness and to set forth a future candidate that may
prove desirable. (Indeed, GPS is the only system of the three that will give
three-dimensional position.)
471
-------
Table 1 rates the various systems according to the criteria of accuracy,
coverage, weight/volume, and cost. The ratings pertain to the baseline param-
eters set forth in Figure 3.
With respect to tracking accuracy, Omega is the poorest performing system
but is, nevertheless, marginally capable of meeting a specification of 1-mile
position uncertainty. Although the system can perform to this accuracy during
daytime, at night 2-mile errors are more realistic when RF propagation at very
low frequencies (10-14 kHz) tends to become less predictable.* The other systems,
Loran-C and GPS, should always provide position estimate accuracy to 1 mile.
Omega presently provides full coverage of the United States through use of
the stations in Hawaii, North Dakota, and Norway. However, in the region west
of the Rocky Mountains it is usually more desirable to substitute the station in
Japan for that in Norway. The station in Liberia is also a possibility, but the
inferiority of east-to-west propagation makes its use marginal. GPS., likewise,
will provide full coverage of the United States. Loran-C, originally designed
to service marine traffic in the coastal confluence region, leaves the region
between the Rocky Mountains and the Mississippi River largely uncovered (see
Figure 4). Plans are being formulated to rectify this situation, but no timetable
for implementation has been forthcoming.
With regard to weight/volume, both Omega and Loran-C appear capable of
meeting a 750-g payload limit. Evidence of this assertion is provided by
equipment currently manufactured (e.g., by Beukers Laboratories, Inc.). Thus,
it is reasonable to assume that in either case a receiver/retransmitter combina-
tion weighing less than 200 g (perhaps half this value) could be assembled. If
power were provided by high energy-density lithium cells, an additional 100 g of
payload woujd suffice to power the electronics for 3 to 6 days depending on the
transmitter duty cycle employed. Although these numbers are admittedly estimates,
they do indicate that a 750-g payload limit does not present a fundamental
difficulty to either Omega or Loran-C. The weight of a GPS receiver/retransmitter
is projected to be roughly twice that of Omega, implying that it, too, will be
capable of meeting the weight allowance.
The cost of either an Omega or Loran-C receiver/retransmitter is thought to
be in the neighborhood of $100; that of GPS, perhaps two to three times greater.
*It should be noted that differential Omega, a distinct possibility in
tetroon tracking, reduces position uncertainty to one-half mile, day or night.
472
-------
Again, there appears to be no difficulty in meeting a $1,000 ceiling, even
with the cost of tetroons.
IMPLEMENTATION
The preceding discussion indicates that any of the radio navigation
systems considered would be feasible for use in what might be called the "air
segment" of a tetroon tracking system; i.e., that portion of the overall
system carried aloft by the tetroons. That is to say, one should experience
no fundamental problem in: (1) designing a receiver that is capable of giving
position data continuously if need be and to the desired accuracy; and (2)
designing a companion retransmitter, the total package weighing 200 to 300 g
and costing $100 to $200.
Problems do arise, however, when one attempts to envision how one might
monitor the movements of multiple tetroons simultaneously over extended periods
of time and make tracking data available to, say, a central facility on a near
real-time basis. In this regard, two questions must be answered:
How does one handle several transmitters operating simultaneously?
How does one position oneself to listen to these transmitters when
the tetroons carrying them may see their paths stretching over
hundreds of miles and diverging by like distances?
The first question has a straightforward answer: encode the transmissions
so as to create a code-division multiplex scheme. Such an approach is used
routinely in military systems and would require little additional hardware to
be added to the tetroon payload.
The second question is not so susceptible to solution. Indeed, the only
answer would appear to be one or more aircraft monitors in the air virtually
continuously, receiving the tetroon broadcasts and relaying them to a central
control point. Such an approach is obviously quite demanding on men and
equipment.
Were one to relax either the position update rate or the data turnaround
time requirement, however, several interesting possibilities arise that exploit
the unique characteristics of a retransmitter tracking system; to wit, that
tetroon position is an air-derived measurement (distinct from a radar tracking
system in which tetroon position is a ground-derived measurement). From this
property of retransmitter systems, two important observations emerge:
473
-------
One can easily relocate the tetroons after extended periods of no
surveillance. Indeed, on the basis of very crude position predic-
tions, one need only send an aircraft up to 10,000 ft, for example,
in the general vicinity of the tetroons to be able to receive all
their broadcasts emanating from within a range of several hundred
miles.
Position data generated by the tetroon can be recorded for later
broadcast on command to the observation aircraft considered in the
preceding paragraph. In this way, one could recreate the complete
trajectory a tetroon has followed by monitoring it twice a day, for
example.
It is recognized, of course, that the last concept requires still more airborne
equipment, but with solid-state memories it should be feasible.
In summary, the concept of using the retransmission technique for multiday
tracking of air parcels appears feasible within a set of requirements somewhat
less demanding than those of Figure 3 concerning positions update rate and data
turnaround rate. Furthermore, a starting point for the design of the required
equipment is commercially available.
474
-------
TETROON
RECEIVER
BEACON SYSTEM
TETROON
RADAR
TRANSPONDER
SYSTEM
RADIO
NAVIGATION
TRANSMITTER
TETROON
RECEIVER
RETRANSMITTER
SYSTEM
Figure 1. Classes of tracking systems.
475
-------
NAVAID
RF SIGNALS
NAVAID
RF SIGNALS
DEMODULATOR
PROCESSOR
TRACKER
RAW NAVIGATIONAL
DATA
(PHASE DIFFERENCES, ETC.)
PROCESSED NAVIGATIONAL
DATA
(POSITION, VELOCITY)
DEMODULATOR
RETRANSMITTER
PROCESSOR
TRACKER
BASE
RECEIVER
DISPLAY,
RECORDER, ETC.
DISPLAY,
RECORDER, ETC,
Figure 2. Retransmission System Principle. Figure 2a (top) shows standard navigation receiver.
Figure 2b (bottom) shows conversion to retransmission system.
-------
EXPENDABLE SENSORS (MAXIMUM COST: $1,000)
MINIMAL AVIATION HAZARD (MAXIMUM WEIGHT: 750 g)
MULTIDAY DURATION (MINIMUM BATTERY LIFE: 72 HR)
LOW ALTITUDE (3,000-4,000 FT)
MULTIPLE SENSORS (MINIMUM 3-5)
POSITION DATA ONLY
MODERATE TRACK UPDATE RATE (HOURLY IF POSSIBLE)
MODERATE POSITION ACCURACY (1 MILE IF POSSIBLE)
RAPID DATA TURNAROUND (REAL-TIME IF POSSIBLE)
Figure 3. Idealized mission requirements.
477
-------
•t*
^J
00
Figure 4. Lor an-C coverage of the United States.
-------
TABLE 1. CANDIDATE RADIO NAVIGATION SYSTEMS
Omega
Loran-C
Navstar-GPS
Accuracy
Moderate
Moderate to
high
High
Weight/
volume
Low
Low
Moderate
(projected)
Coverage
High
Moderate
Very high
Cost
Low
Low to
moderate
Moderate
(projected)
-------
SATELLITE POSITIONING OF TRACERS
James J. B. Worth
Research Triangle Institute
Research Triangle Park, North Carolina
The availability of the Random Access Measurement System (RAMS) on the
later satellites of the NIMBUS series, and the successful use of this system
for tracking balloons in air currents and buoys in the ocean, suggested that
the RAMS might be used for tagging air parcels in which pollutant transforma-
tions were taking place. The RAMS consists of the radio transmitter for
which the position is to be determined and the receiver-storage-transmitter
in the satellite. Position data are retrieved by transmission from the
satellite to a ground station with each orbit (minimum 12 hr, with possibil-
ity of every 6 hr).
The transmitter being tracked transmits a coded radio signal periodi-
cally (1 s of each minute). This signal is received by the satellite if the
satellite's position in its orbit gives it a line-of-sight view of the
transmitter. From the coded signal, the transmitter can be identified, and
from repeated signals from the transmitter (at 1-min intervals) during one
pass of the satellite, the position can be determined by the Doppler shift
of the signal frequency. The period and the inclination of the satellite's
orbit determine the number of position determinations that can be made in a
24-hr period. With the usual orbit of NIMBUS-type satellites, a minimum of
two positions per day is obtained. Up to 100 transmitters can be handled by
a single satellite.
The transmitter, or beacon, shown in Figure 1 is a prototype designed
and built to test the circuitry (Figures 2 through 7) rather than to demon-
strate a model feasible for use on low-altitude balToons or tetroons in
crowded air space. With development we anticipate that this prototype that
now weighs 822 g (with batteries) and measures 22 x 8 x 8 cm can be reduced
480
-------
in weight to less than 280 g and in size to approximately 20 x 4 x 3 cm.
With these reductions the marker beacon could easily be carried by a single
1-m3 tetroon and would be acceptable to the FAA as a flight package. The
batteries would give this package a transmitting life of approximately
72 hr—a reasonable fraction of the lifetime of a PEPE.
For use in investigating the transformation of pollutants in PEPE's, we
envision a cluster of four to six tetroon-supported beacons being released
at a point within the developing system. The release location would be
selected with the anticipation that the trajectories followed by the tetroons
would be within the PEPE for a period of several days. Since superpressure
balloons do not always hold their planned altitude, wind shear can be expected
to disperse the tetroons of the cluster during the several-day period. With
successive positions of the tetroons at 12-hr intervals, reasonable estima-
tion of the path of a parcel of air within the PEPE can be made and approxi-
mations of the dispersion within the PEPE can be obtained from the separation
of the tetroons.
481
-------
CO
(VS
•
• • j_i..ni j-j^
I /'/'/'j'|'/'/'l'/'j'/Ui|'/f l
i .-. •- . J i..'ii^iirr J/«.-..• «4».^!;> !
i|i|i|i|i|MilM«p«14i|MMM'p^W
WtSlCOII M»
RUtSM
Figure 1. Random access measurement system beacon.
-------
BATTER.V
PACK
XTA.L .
RF SOURCE
00
CO
' POWER .
SUPPL.V
PHASE.
SHIFTE.R
DIGITAL.
EN CO PER
POWER
AMPLIFIER
PONDER OUTPUT BOARD
Figure 2. System block diagram.
-------
-Pi
00
.£»
TIME IN
StCOHOS
POVtLK UP
POWIR OUT
RESET COUNTERS
DATA SHlrr
(u)
DATA LOA C U)
PATA CLOCK.
C IOOIIO
57
58
59
-#-
• r
CYCLE RESETS 4 T T
SECONDS
-tt-
~~
n mfn nnnnnnnnnnnnnnnnnnnnnnn
t it
n
a
n nnnn
BIT SYNC PATTERN
FIUKE 3YNC FATTtRN
1C AND DATA ...
CrrPKAL DAM SHOM/N)
I
Figure 3. Timing diagram.
-------
00
TP\ - OSCILLATOK ZQ KHZ
TPi- DATA CLOCK SO"/. DUTY CYCLE
7PJ3 - \/ct PIN H (
TEST POINTS TP5- DATA 5HIFT E.MABLE
TP6— DATA LOAD
TP7- POWtR OUT
TP8— POWER L/P
TP9— RESET COUNTERS
Figure 4. Timing generation circuitry.
-------
TP10- -4,0"
TEST POIWTS FP11- + tO*
TP l£ - O*
00
81* SYNC p*rri
rpmir STNC PATTC
JUMPCfti W«
w|
** 2
T
BIT 1D • <
y.i
KN
.J
0
b
k*L
r
*
',
\
«
«
«
11
L
rJ
1
UI4
r
/5
„
UI5
I
Lilt,
9
UI7
1
a
\
-| '
P
i
TPi*
Figure 5. Data encoding circuitry.
-------
c»
-MOV
CW\/V
-btf
r
-I
L4
Dio DH o
-N—t4—i>l—if
RF POWER.AMP
Tbl
PHASE E.NCODE.R
Figure 6. Phase encoder and rf power amplifier circuits.
-------
-
B2
I
JL
00
00
B5
1
POWER UP
(H)
T
-O + 10 VOUT
c
Figure 7. Regulated DC power supply.
-------
RADAR TRACKING
C. Raymond Dickson
National Oceanic and Atmospheric Administration, Idaho Falls, Idaho
BRIEF HISTORY OF TETROON TRACKING
Tetroon tracking was designed to solve atmospheric transport problems.
In the 1950's we were following tetroons with double theodolites to obtain
trajectories. This supported the radioactive plume studies at the National
Reactor Testing Station (now INERL) at Idaho Falls, ID. In the late 1950's
we began doing diffusion studies that went out to many kilometers. A means
of following a tetroon to determine the centerline of the plume had to be
found. Because of the little success we had had with double theodolities,
we tried to work with an AN/APS-3 radar taken from a Navy torpedo bomber.
We first tried to skin track by putting x-band corner reflectors on the
tetroon. This did not work very well because of the low power of this type
of radar (this was a 1948 vintage radar).
We realized that it was necessary to develop an inexpensive radar
transponder that could be followed to a range of 15 to 20 mi. In 1961 we
obtained an M-33 radar, and with the help of Mr. Earl Pound we developed a
system with the capability of automatically tracking transponders to a range
of 60 mi. The M-33 radar had a 1° beam width with a range of 100,000 yards.
The characteristics of this system were reported at a 1961 radar conference
in Kansas City.
TYPES OF RADARS USED FOR TETROON TRACKING
We have used many types of radars. One, the SCR-584, has an x- and
s-band and a 4° conical scan. We were able to track transponders with it,
but its accuracy was not the best. Another type is the BSPS-1 Bendix Coast
Guard, which has a 1.6° conical scan. We were able to follow both tetroons
and ocean buoys with it. We tracked tetroons with the AN/MPS-19 s-band radar,
which belonged to NASA and had a scan of 3°. Still another type is the
489
-------
AMPQ-12 s-band with a 2.4° conical scan. We have also followed balloons
without transponders from an FPS-16 at Wallops Island; and again with trans-
ponders on the WSR-57, a National Weather Service (NWS) radar. The NWS
radars were not designed to track tetroons and do not give a very accurate
height indication because of the large amount of scanning required to find
the height from the signal. We used three different WSR-57 radars to follow
transponders out to 200 mi. We were able to successfully track tetroons
with the Plessey WF3 automatic tracking radar used on the Isle of Wight in
the British Isles. Other types of radars we have used are the FAA 2-MW
radars in the 1,340-mHz band, used in the Air Traffic Control Radar Beacon
System, and the radar developed for the Air Force B-52 bombing runs, manu-
factured by Reeves.
Many types of radars are on the market. The versatility of tracking
tetroons with and without transponders can be seen. The M-33 radar is the
one we have used, developed, and modified. We now have five semimobile
M-33's in operation. These have a narrow beam width and we can track the
transponders automatically.
When using a transponder on a tetroon it is advantageous to put the
radar as high as possible in order to follow the tetroon near the ground and
below radar height. Using this method the only signal you will see is the
electronic signal from the transponder. During skin tracking, the radar is
put as low as possible so that it looks upward and ground clutter is mini-
mized.
At present day prices the tetroon-transponder system, based on the
study we completed in Tennessee in September 1978, costs approximately $125
&
per flight. In the past 20 years we have released nearly 2,000 tetroons,
many with transponders. We have completed studies in many different parts
of the country using radar to follow trajectories. These include studies in
Cincinnati, Ohio, Columbus, Ohio, Wallops Island, Virginia, Salt Lake City,
Utah, Atlantic City, New Jersey, New York City, the Isle of Wight in Great
Britain, Oklahoma City, Oklahoma, Haswell, Colorado, Knoxville, Tennessee,
Clarksville, Tennessee, four diffusion tests in Los Angeles, California; the
Southwest Energy Study in Southern Utah; and many tests in Idaho Falls, Idaho.
Figure 1 shows the effect of the earth's curvature on the line-of-sight
distance required for radar tracking. The earth's curvature determines the
490
-------
distance any tetroon at a given height can be tracked by ground-based radar.
Tetroons at low altitude above ground (1,500-3,000 ft) can be tracked only
short distances. For example, if a balloon is flown at 1,500 ft, the best
range you can get is 52 statute miles because of the earth's curvature,
provided there is no interference from terrain features. If the balloon is
3,000 ft in the air, it can be followed for 74 statute miles.
Another problem we have encountered in following tetroons and using
them as tracers is the oscillation. The balloons follow a constant density
surface rather than a constant height. Figures 2, 3, and 4 show the up and
down movements of a balloon. In the morning with a class E or F stability
condition (very stable), the balloon has very little oscillation—the height
of constant density surfaces changes very little and the balloon remains at
a constant height. As heating takes place, thermal mixing occurs and oscil-
lations result. By midafternoon the oscillations may carry the tetroon
through the complete mixing layer. Figures 5 and 6 show nighttime and early
morning oscillation of tetroons in the west coast marine environment. If
you compare these with smoke coming from a stack, you will see looping type
plumes and coning type plumes. The motion of the balloons compares with
that of the plumes and because the pollutants will follow the same kind of
motion, the tetroons should be very good tracers.
Under stagnation conditions we have followed tetroons for as long as
72 hr. During the STATE Project we followed some for 16^ hr (out to 330 km).
An interesting thing about transponders is that because they can each be
tuned to a different frequency, several can be released simultaneously and
each can be tracked separately.
Figure 7 is a computer listing of radar data. We plot the map and the
trajectory and compute the statistics from the radar data. A printout from
the system gives us an x, y, and z position and in varying terrain gives the
altitude above ground. Range and bearing from the radar site are also
given. We measure u, v, and w, rotate it, and get the along stream, cross
stream, and w components, then run the correlations and auto-correlations of
all of these components. This is a standard printout of the radar system.
Last year we began development of a portable radar to use on the STATE
Project. Although it wasn't completed, we took it to Tennessee and it
worked satisfactorily. It is 12 ft long and has computer-controlled auto-
491
-------
matic levelers and a self-erecting system for the antenna structure. The
antenna structure can be removed in a few minutes and stored while the radar
is being moved. The radar is equipped with its own power source, air condi-
tioning, heating, and a complete communications system for communicating
with the control base. There are some problems with a mobile system when
you are using more than one radar and leap-frogging for many miles; locations
must be picked out in advance, and permission obtained from landowners to
set up radars on their property.
492
-------
-p.
l£>
CO
MEAN
PLUME HEIGHT
1500-3000 ft.
MIXING DEPTH
SOOOl't.
TARGET HEIGHT
ABOVE ANTENNA
(FEET)
MAX. RANGE-
(STATUTE
MILES)
EFFECTIVE EARTH
RAPIUS
EQUALS 4/3 ACTUAL
(STANDARD
ATMOSPHERE )
Figure "\. Earth curvature limitation to radar range.
-------
15
1H
13
13
IS
10
OXC-71 FLIGHT 14 SEPT 29, 1971 08:51-11:07 COT
9
o
X
N
8
7
I
5
m
09:11
09:55
09:26 09:H1
TIME OF DAY (COT?
Figure 2. Altitude oscillations of a tetroon in a stable (early morning) air mass.
10:11
10:c6
-------
COLO 72 FLIGHT 14A RADAR*! AUO 7.1972 06:C3 - 10:15 KCT
-p»
en
23
22
21
20
19
r
- is
o
o
1 7
'
16
15
14 -
13
i
12
i
- ; i
i j
i i
j j
07:
I
'
58 08:
1 j
1
|
i
23 03:
Jj
28 08:
••
.. •••
•
••
•—
. .
43 08
•r
— '
p—
1
I
—
r—
rr
1
.
1 i
'
i
58 09:
TIME OF DAY (MDT)
Figure 3. Altitude oscillations of a tetroon as convection begins.
-------
13 r
LAX-69 FLT M6 9/20/69 1839-1613 HARBOR LAKE
1310
1440
1340. 1410
TIME OF DAY (POT).
Figure 4. Altitude oscillations of a tetroon during well-developed convection.
15 in
-------
18
16
IH
LAX-69 FLT 97 9/30/69 8145-0024 CITY OF COMM.
.-•
10
*0
N
M
18
10
14
28
N
10
PATH LENGTH THDS/METER
15
30
33 MO
PATH LENGTH THDG/METER
50
55
60
Figure 5. A tetroon flight in a nighttime west coast maritime air mass showing a minimum of altitude change.
-------
18
m
2e
LAX-69 FLT 58 9/83/89 0818-1211. RANCHO CIENOA
PATH LENGTH THOS/K£T£R
in
oo
' 18
:•: m
x12
If 10
N
H
8
35 >
,••.''•'
60
PATH 4.ENGTH THDS/KETER
Figure 6. A tetroon flight in a west coast air mass during the morning hours. Vertical oscillations extend through the entire
layer below the maritime invasion.
-------
LAX-S3
10s IO<2/69 OS33-0933 CITY OF COWI.
IN
j» sec
92122
95??2
3^5^*3
9£*»2-
9^S32
9£6.2
253.0
251.9
2*t3.0
2x3.2
237.1
229.3
N >
MOT A8V PANCE
TERRAIN IN
ttTESS TABOS NS
13*. 7 l3-»73 60
155.1 ISC-il 60
173.6 15523 60
?03.3 15371 60
2-*. 3 13223 60
273. s 11070 60
283.3 1X961 60
29s. 2 1*363 60
3IS.2 1*168 60
33?. 0 isess 60
3x7.3 IS3I3 60
318.3 ISSOS 60
280.3 1X306 60
219.0 |S23£ 60
153.2 1*; 31 60
56. 7 . 1X117 60
59.1 ISC33 S3
56. 8 13953 63
60.7 I3>ca 60
61.9 13832 60
30.1 133C3 69
33.5 13706 60
S3.S 13596 60
73.1 13x35 60
105.2 13359 60
I2U'.0 !32!9 60
ISO. 5 130-1 60
133.9 leo^ 68
123.1 12634 60
K2.9 12515 60
109. 3 123*9 60
MEAN
78 VARIANCE
STANDARD DEVIATION
HOT
IN
«£TM
280
296
321
360
HCQ
H*«Q
H39
1,57
**9d
509
329
501
S5S
SOS
3Vs
291
253
252
2S3
272
26X
260
271
309
350
383
352
337
354
350
339
369.7
9027.3
93.0
COCPO IN
RAQA3"
X
yss
3199
5137
5103
30X0
XS65
*382
S903
s72 3
"S55
*593
*530
SS63
*396
S360
S353
S132
X3x8
S3I9
S299
S275
S2SI
•••238
S223
;.2l7
S237
* 1 32
SISS
S12S
S069
3339
METER
(0. 01
T
-I3S3I
-13328
-1 VIS
-1309s
-12971
-12863
-12779
-12711
-12638
-12531
-12S5S
-12369
-12301
-12250
-IP200
-12136
-12552
-IISSS
-11953
-11916
-11839
-11771
-llf.77
-1157s
-11x35
-11315
-11155
-10993
- IC8x I
-10590
-10535
niLES
FROM
fiAOAH
8.95
8.69
8.81
8.73
8.65
8.37
8.50
8. XX
3.38
8.32
8.25
8.18
8.13
8.09
8.03
8.01
7.97
7.93
V.90
7.87
7.83
7.73
7.72
7.65
7.53
7.50
" -»i
7.31
7.31
7.11
7.01
CHANCE :s MKTER/SEC STPJAM
cow«:Nrs
HI SO KINO
( V X.T.H I in r.ETERS'ScC OIR. i°££0
U V
-O.BS
-0.77
-0.69
-0.83
- .06
- .25
- .33
- .32
- .33
- .13
- .OS
- .05
- .12
- .07
-0.6S
-0.11
-0.02
-0.07
-O.S8
-0.33
-O.XI
-0.23
-0.38
-0.23
-0.10
.61
.71
.89
.02
.05
.77
.S2
. IX
.21
.XS
.53
.S3
.IS
.85
.82
.07
.23
.10
.6s
.70
.95
.S7
.57
.71
.93
-o.ia 2.32
-0>0 2.52
-0.30 2.77
-0.53 2.62
-0.93 2.52
-1.32 2.26
-0.83 2.01
O.IS 0.33
O.SO 0.62
li LCNO
0.33 I. 81
0.27 1.87
C.X2 2.01
0.65 2.21
0.75 2.30
0.57 2.11
0.31 . .8S
0.14 .55
0.37 .63
0.3S .'i3
0.30 .86
-O.SS ,7s
-0.62 .S3
-0.97 .19
-1.02 .01
-0.90 .03
-0.55 .is
0 . 06 .Ox
0.10 .77
0.07 .77
-0.13 .03
-0.07 .sx
0.18 .60
0.62 .67
0.71 .87
0.5X 2.21
0.16 2.X9
-0.09 2 67
-3.31 2.67
-0.30 2.68
-0.19 2.59
O.Ox 2.17
0.22 0.3X
O.S7 C.5P
CROSS
0.18
O.C6
-0.03
a. 05
U. 19
O.xO
0.7*
C.79
0.76
O.Md
0.36
oleo
0.67
0.23
-0.31
-U.*5
-0.36
0.20
0.0*
(I.C?
-0.35
-o!x2
-o.r.v
-0/I2
-0 C>9
-0.73
-0.39
-0.10
0.3C
-0.00
0.20
O.S5
(CIO.: !W»i
15? .C?.
lf.6 ' 07
ICC .01
ISCi .21
«!>J .53
IsS .16
I3S .90
131 .7s
132 .83
Is? .8s
|S7 .09
iss .79
1 36 . SO.
129 .37
IS? .Cs
1 7s .0™
179 .£J
177 .10
1*3 .C1
155 .'•'
157 .(.I
171 - .«i«i
160 -f.2
172 .VI
IV V .03.
PS 2.3J
171 c.i6
I7x 2."'-^
166 2.70
I uG 2.08
150 fl.OI
157.3 2.21
IS6.0 O.ls
12.1 O.Ml
OISPLACD1CHT YEI.OCITT - 2.17 flEAH TERRAIN HOT - ISS.6 hE*N HOT ABV TERRAIN 215.2
UV
O.OIS
VRVN
-0/089
UW
O..OI7
VRM
0.050
VH
O.C6I
VNH
-0.039
RUV
0.058
RvRVN
-0.3S6
RUU
0.092
RVRU
0.185
RVU
0.213
RVNM
-0.188
Figure 7. Computer printout of radar data.
499
-------
LIGHT-WEIGHT, LOW-COST RADAR, TRANSPONDER SYSTEMS
Earl F. Pound
Utah State University, Logan, Utah
I have worked on transponder systems for weather applications since
June of 1960. Most of this work has been for Ray Dickson's group at the
Idaho Engineering Laboratory, Idaho Falls, Idaho. All of this work has been
on one type of system using a relatively high-power pulsed radar transmitter
at X or S bands to beam a signal to the transponder with a reply at 403 MHz.
This category of transponder leaves the complexity on the ground, allowing
for a light-weight, long-battery-life, low-cost transponder. The relatively
high pulse power of the transmitter (50 kW to 2 MW) allows for a simple
crystal video receiver in the transponder. The high frequency (X or S band)
and the highly directional antenna used to transmit the radar signal, along
with control of the transmitter power, allow for angular tracking of the
transponder. The slant range is determined by the round-trip delay of the
reply.
The transponder receiver consists of an X or S band antenna, a d.c.-
biased microwave detector, and a high-gain video amplifier. This portion of
the transponder must operate continuously, always ready to receive a signal,
which produces most of the battery drain. The use of high-gain bandwidth
transistors in this amplifier allows for relatively low battery drain. The
rest of the transponder system works in a pulsed mode, drawing power for
1/1000 or less of the time that the transponder is being interrogated;
hence, it contributes only modestly to the battery drain while providing a
1-W pulsed reply signal. This pulsed part of the transponder consists of a
blocking oscillator to establish a fixed reply pulse width, a pulsed power
amplifier, and a self-excited pulsed transmitter. The system operates about
96 hr at room temperature on two 9-V transistor radio batteries, with a
total weight of 150 g. Alkaline or mercury type batteries would almost
double this time with only a slight increase in total weight. Lithium type
500
-------
batteries would give at least 3 weeks of life even at reduced temperatures.
I have made a low drain model for the Wildlife Service that has a battery
life of 4 to 5 months and no increase in size or weight.
A conical scan-type radar is used with this transponder for automatic
tracking; that is, the transmitted beam from the radar nutates in a small
circle around the average direction of the antenna. The amplitude of the
target reply compared to the nutating of the antenna gives the error signal
for automatic tracking. The amplitude of the transponder reply is constant;
hence, at first glance a person would say that this type of automatic track-
ing would not work. If the transmitted power is controlled either manually
or automatically so that the transponder interrogation is turned off and on
by the nutating of the antenna, the average effect would be the same as for
a passive target and an automatic track would be achieved. If the trans-
mitted power from the radar is higher than needed, an automatic track would
still be achieved but the apparent direction of the transponder would be
found somewhere on a circle around the true track. This power control is
not overly critical for practical applications.
A reflective target gives an error signal for automatic tracking that
is proportional to the misalignment of the radar antenna and the target;
that is, it is a linear servo system. The transponder, as a target, gives
an on-off error signal and thus causes the automatic tracking servo to
become nonlinear. Because the transponder is operating close to maximum
sensitivity when it is turning on and off, the signal level is close to the
background noise level. The noise added to the signal causes a somewhat
random triggering of the transponder reply, thereby restoring some of the
lost linearity by the proportion of pulses missed.
The transponder will also work with a PPI type radar display with the
target appearing as a line on the screen. The true angular position of
target is the center of this line.
The transponder will not work with a mono-pulse type radar unless it is
made especially for the transponder. A mono-pulse system derives its error
signal for automatic tracking from the received signal. The transponder
described here replies at 403 MHz; hence, it requires an additional antenna
and a receiver to replace the original receiver. This addition would defeat
the original tracking system rendering it useless. Personnel at White Sands
501
-------
Missile Range have modified a Nike Ajax Radar System to receive in a mono-
pulse mode at 1,680 MHz. This system works with a cross band transponder,
with X band for the up leg and L band for the down leg. The main reason for
cross band operation is that no light-weight, cost-effective X or S band
transmitters are available at present.
Another well-known transponder is that used by radiosondes with the
GMD 2 or GMD 4 system. This uses a doppler tone ranging system at approx-
imately 82 kHz and, in addition, the GMD 4 has pulsed course ranging in the
event the track is not continuous. This transponder system has some advan-
tages and disadvantages compared to the pulsed system. The main advantage
of the GMD system is the lower initial cost, which is partially offset by
the availability of surplus radars such as the M33. The main disadvantage
of the GMD system is that the transponder transmitter must operate continu-
ously. This means that even the best low-drain all-solid-state GMD trans-
ponder would take about 100 times as much battery power as the pulsed type.
Consideration has been given to using the existing FAA radar network
with transponders attached to tetroons for weather applications. This is
probably not cost-effective because of major limitations of the FAA system
for this application.
The FAA transponder radar transmits on a frequency of 1,030 MHz with a
peak pulse power of from 25 to 3,000 W depending upon the desired range of
the station. Even at 3,000 W, which would involve very few stations, this
is about 1/100 of the peak radiated power of the M33 radar. A simple crys-
tal video receiver is therefore impractical for the FAA system. A super-
regenerative receiver (the type used in the GMD transponder) is also imprac-
tical because the quenching frequency could not be made high enough to
accurately recover the short pulses transmitted by the FAA radar. The only
practical transponder receiver would be of the super-heterodyne type, which
would be at least an order of magnitude more complex than the crystal-video
type.
The FAA transponder system receives on a frequency of 1,090 MHz. The
associated transmit-receive switches and low relative performance of the
receiver, coupled with the automatic sensitivity time control, require a
relatively powerful transponder reply transmitter. The minimum power used
for aircraft transponders is 100 W peak power. The balloon-borne trans-
502
-------
ponder must reply with at least 10 W peak to give even marginal performance
with this system. This reply is at 1,090 MHz, where a self-excited oscilla-
tor is not practical; hence, a transmitter with two stages or more would be
required. This means that the transmitter must be about an order of magni-
tude more complicated than for the 403 MHz transmitter described earlier.
The use of a reflector could be cost-effective when used with the FAA
or other radar systems using a PPI type display. The skin track portion of
the FAA system transmits a peak pulse power of 2 MW on a frequency of about
1,340 MHz. A corner reflector built into or hung below a tetroon would have
good range with this skin track system. The identification problem could be
solved by mounting an array of diodes on the corner reflector so that passing
a current through the diodes would change the reflectance. This would cause
the radar display to scintillate, the code giving identification, tempera-
ture, barometric pressure, etc.
Other sensors such as temperature, barometric pressure, humidity, etc.,
although practical, have not been added to the transponder. Each would
require more battery power and add weight and cost to the transponder. The
low production levels for transponders has not allowed these additional
sensors to be cost-effective compared to conventional radiosondes. The most
practical sensor reply technique would be to add additional pulses to the
radar reply pulse using pulse separation to return the sensor signal, or to
turn off the radar transmitter and allow the transponder to free run at a
lower frequency multiplexed to give sensor signals.
503
-------
MINIATURIZED LASER DOPPLER VELOCIMETER*
David H. Dickson
U.S. Army Atmospheric Sciences Laboratory,
White Sands Missile Range, New Mexico
Abstract
This laser doppler velocimeter (LDV) was designed for a unique applica-
tion. The operational concept is a Helicopter Remote Wind Sensor (HRWS) to
augment a fire control system for attack helicopters. The HRWS will measure
wind speed and direction in the aircraft turbulent environment and average
winds for several hundred meters.
The HRWS was fabricated for the Atmospheric Sciences Laboratory by
Raytheon Company. This system has been successfully operated on a UH-1 air-
craft and a comparative ground test. This paper will describe the system,
data results, and potential applications.
SYSTEM DESCRIPTION
A C02 laser heterodyne system has been constructed and mounted in a pod
compatible with an AH-1 or UH-1 helicopter or fixed wing aircraft equipped
with standard wing shackles. A block diagram of the system is shown in
Figure 1. It can be subdivided into three subsystems: the transmitter, the
scanners, and the processor. The transmitter consists of a C02 laser and
appropriate beam-combining optics for heterodyne operation. Two scanners
are employed; one scans in range and the other produces a conical scan. The
beam pattern is shown conceptually in Figure 2. The processor includes the
detector and its associated biasing network, a frequency tracker and counter.
A 5-W laser transmits a CW beam into an interferometer, which propagates
most of the energy into a telescope and the rest to the detector for use as
a local oscillator. The telescope expands the beam to 66 mm and provides
the ability to change the location of the focus by translating the
*This paper was prepared for, but not presented, at the Seminar/Workshop.
504
-------
telescope secondary. A rotating wedge is placed at the telescope output to
generate the conical scan. A complete rotation of the cone takes place
approximately every 50 ms, while the location of the focus varies from 1 m
to 33 m every second.
Energy scattered from naturally occurring atmospheric aerosols is
collected by the same telescope, passed through the interferometer onto a
mercury cadmium telluride detector. The detector output is processed by a
frequency tracker and digital counter producing data proportional to the
signal frequency in both analogue and digital format. A new piece of datum
is available every 2.8°, approximately every 400 us. Output data consist of
range, angle, and frequency information. For the flight tests, these data
were recorded, although for future applications, a microprocessor will be
integrated into the system.
DOPPLER THEORY
Operation of a laser doppler velocimeter may be understood with the aid
of Figure 3. A laser beam is transmitted through an interferometer that
passes most of the energy to the target and diverts a small amount to the
detector. The transmitted beam is passed through a telescope to the target,
which, in the present case, consists of aerosols naturally suspended in the
atmosphere. Some of the transmitted energy is backscattered by the target
particles and enters the detector by the path shown. This returning energy
has been shifted in frequency by an amount proportional to the component of
particle velocity parallel to the direction of propagation V^, according to
the Doppler principle. Figure 4 shows the geometry of the laser system and
target velocity vector relationship. Only VX1 contributes to the Doppler
shift.
There are two beams falling on the detector at two different frequen-
cies as shown in Figure 5: a reference beam at the optical frequency f
with a power of a few milliwatts and the smaller power signal beam at the
new frequency f + f . where f . is the Doppler shift frequency. When these
beams are superimposed, the combination contains energy at the difference of
these frequencies. The difference is the Doppler frequency and is related
to the target velocity.
505
-------
CONICAL SCANNING
A laser Doppler system measures a single velocity component during any
measurement interval. The conical scanning technique, developed for micro-
wave radars for measuring the velocity components of a laminar atmosphere,1
was selected for this application. Data are taken at several points around
the periphery of the cone. As can be seen in Figure 6, the angle between
the propagation direction and the velocity vector changes as the beam sweeps
around the cone. Therefore, a different velocity component is measured at
each point, and the velocity vector can be obtained.
If the cone axis is in the z-direction, it can be easily shown that
_ o t t t
f, = -r (V, cos -^ + V sin S cos tut + V sin « sin wt)
Q A Z £ X £ y f.
where:
V , V , V = velocity components
x y z
c
2 = half-angle
uu = conical scan frequency.
Thus, the frequency consists of a DC term and a sinusoidal component.
The DC represents the velocity component parallel to the cone axis, while
the magnitude and phase of the time varying part permit the calculation of
the other two components. Deviation from laminar flow will yield a more
complex time variation, but three-dimensional data can be obtained as long
as the data approximate a single frequency sinusoidal function.
COMPONENT DESCRIPTION
Calculations indicated that a 5-W laser would provide sufficient power
to reliably measure helicopter flow fields, so a previously developed Raytheon
Company 5-W air-cooled laser with low weight and size and an extremely high
degree of ruggedness and reliability was selected. The features of this
laser include an all metal/ceramic laser tube and a stable lightweight laser
frame.
The heterodyne optics consist of a combination of beam splitters and
polarizing elements that generate the local oscillator and the transmitter
beams, and recombine the collected signal with the local oscillator. Meas-
506
-------
urements of the detector indicate that local oscillator beams of approxi-
mately 1 to 2 mW provide the maximum signal-to-noise ratio. The reflectivity
of the beam splitters was selected to generate a local oscillator beam of
this value; the remainder of the energy is transmitted into a beam expander.
Afer collection of the scattered energy by the beam expander, the heterodyne
optics combine the local oscillator with the signal and direct this super-
position toward the detector.
The beam expander produces an output beam diameter of 66 mm. It is
capable of focusing the beam at ranges varying from 1 m to 33 m. Backscatter
energy from the telescope elements had to be controlled to avoid difficulties
with the detector. The design from these specifications resulted in a
Galilean telescope with an input negative singlet lens and an output doublet.
Range scanning is accomplished by translating the input lens of the
beam expander. If the input lens were scanned linearly with time, too much
time would be spent at closer ranges. Therefore, a nonlinear drive function
is electronically generated, which gives an output scan in which range
varies approximately linearly with time.
The conical scan is generated by rotating a wedge about the optical
axis. The wedge is driven by a motor, whose speed is variable from 0 to
2,300 rpm. The motor speed is adjusted by a potentiometer, and can be
displayed on the control panel. As the motor speed decreases below a few
hundred revolutions per minute, the laser beam is automatically blocked for
safety reasons, preventing propagation into the atmosphere. In addition to
measurements of the motor speed, signals are available every 2.8° of rota-
tion to indicate the angular beam position.
The receiver consists of a 10.6-um detector and the associated bias
circuit and amplifiers. The requirements of high quantum efficiency and
good frequency response dictated the use of a mercury-cadmium-telluride
(HgCdTe) detector. The detector is mounted in a Dewar flask, which holds
liquid nitrogen for a period of at least 6 hr. The electronic circuitry of
the receiver was designed to minimize loss of signal-to-noise ratio. The
function of the bias circuit is to provide a voltage supply to the detector
and to provide the proper matching between the receiver preamplifier and the
detector. A prime requirement in designing this circuit is to protect the
detector from excess voltages or currents that can result from removal of
507
-------
the local oscillator, too high a local oscillator power, or loss of coolant
in the detector Dewar.
The frequency tracker is shown in Figure 7. The input band is trans-
lated up to a center frequency of f by mixing the input with the output of
a voltage-controlled oscillator (VCO). The signal is then limited and fed
through a frequency discriminator centered near f . In tracking operations
the discriminator drives an integrator, the output of which drives the VCO.
With the tracking loop closed, the sum of the input and VCO frequencies is
constant. The input signal is periodically replaced by the output of a
crystal oscillator. The readout circuit makes use of measurements made
during this calibration interval to automatically correct for drift in the
discriminator and integrator.
The frequency counter counts the positive zero crossings of the VCO
during a 160-|js period once each angle clock pulse. The sample gate gener-
ator provides this accurately timed counter gate, along with a pulse immedi-
ately after it, to clock the output storage register.
The sensor is operated from a control panel, as shown in Figure 18,
located in the helicopter. All operations on the system are performed from
the control panel except filling the detector with liquid nitrogen, which
was done during aircraft preflight. The control panel also has a number of
diagnostic capabilities that may be used by maintenance personnel.
DATA ANALYSIS
It should be noted that HRWS was designed to measure to a range of
approximately 100 ft from the sensor origin (cone Apex). Further, the
sensor is a descendant of other conical scanning devices whose domains of
measure were considerably less turbulent and less subject to linear shear
than that in a helicopter's vicinity. Wind tunnel and other measurement
studies demonstrate the complexity of the helicopter flow-field.2
For this experimental system, it was cost-effective to locate the
entire sensor in a pod compatible with both rotary and fixed-wing aircraft
as well as ground-based installations. A pod using the hard points of the
AH-1 rocket pod was accordingly built. A picture of the system enclosed
within the pod with the pod attached to a UH-1 helicopter is shown in Fig-
ures 9, 10, and 11.
508
-------
Two simplified approaches have been applied in the data analysis:
(1) circular scan methods, and (2) point-pair methods. Considering the
latter, which is the simplest, any pair of points separated by a sufficient
angle can be used to deduce the wind components transverse to, as well as
parallel to, the axis halfway between them.
The point-pair method has been applied to derive downwash profiles as
running functions of range over a short sequence of range scans.
RESULTS
Measurements were made in hover and at forward flight velocities at
altitudes from 1 m to 500 m. Data were successfully obtained throughout all
flights and are currently.being processed.
A typical data presentation is shown in Figure 12. The helicopter is
hovering with its skids 1 m above the terrain. At a range of 7 m, the
radial velocity component is shown as the beam is scanned around a cone. A
function, approximately sinusoidal, is observed peaking at 180°. The peak
velocity is approximately 18 m/s and the mean approximately 6 m/s. Since a
17° cone half-angle is used, this corresponds to a velocity component paral-
lel to the aircraft axis of 6.3 m/s, a vertical component of 23 m/s, and a
zero cross component.
Comparison of wind tunnel and flight results are shown in Figures 13,
14, and 15. Figure 13 is a downwash profile taken in a wind tunnel using a
scale AH-1 aircraft.2
Figures 14 and 15 are the measured downwash profiles from a hovering
UH-1H aircraft at 100 ft above ground level and 20 ft above ground level.
The 100-ft hover case is clearly out of ground effect and aircraft physical
differences account for the slight profile variation. However, the 20-ft
hover is in ground effect, which accounts for the magnitude of differences.
The general curve slope remains basically the same for all three cases.
COMPARATIVE GROUND TEST
The HRWS was compared with ground-based wind instruments at Biggs
Optical Range (BOR). The comparison consisted of aligning the HRWS with an
annometer array as shown in Figure 16. A series of measurements were made
under several wind and turbulence levels and reduced to 1-min averages.
509
-------
Typical data are shown in Figures 17, 18, and 19. Figures 17 and 18 are
time history plots of average HRWS measurements to a 500-m range compared
with averaged annometer data and Laser Crosswind Sensor (LCS) data, respec-
tively. Figure 19 compares a single annometer at a 24-m range with the HRWS
focused at the equivalent range. Summarizing tabular data are shown in
Figure 20.
The agreement in the cases shown is quite good. Variations are readily
attributed to dissimilar instrumentation.
CONCLUSION
A C02 laser doppler velocimeter has been successfully miniaturized,
flown aboard a helicopter, and compared with other ground-based systems.
Data appear to be consistent with wind tunnel measurements and compares very
well with ground-based instruments.
A future system will be coupled to a microprocessor to obtain real-time
three-dimensional wind vectors as a function of range.
This miniaturized LDV can readily be used as a diagnostic tool for
aircraft wind field investigation and as a ground-based wind sensor to probe
the first few hundred meters of the atmosphere.
REFERENCES
1. K. A. Browning and R. Wexler. The Determination of Kinematic Properties
of a Wind Field Using Doppler Radar. Journal Applied Meteorology, 7,
1, February 1968.
2. Anton J. Landrelre and John C. Bennett, Jr. Investigation of the
Airflow of a Hovering Model Helicopter at Rocket Trajectory and Wind
Sensor Locations. United Technologies Research Center, Report No.
477-912573-15, July 1977.
510
-------
RANGE
SCANNER
LASER
HETERODYNE
OPTICS
TELESCOPE
DETECTOR
RECEIVER
CONTROL
PANEL
CONICAL
SCANNER
FREQUENCY
TRACKER
FREQUENCY
COUNTER
T
A
P
E
R
E
C
O
R
O
E
R
Figure 1. Block diagram HRWS.
Figure 2. Conical scan configuration on AH-1 Aircraft.
511
-------
ff
\
Figure 3. Laser doppler velocimeter transmission and reception.
LASER
SYSTEM
Figure 4. Doppler velocity — measurement geometry.
512
-------
Figure 5. Optical heterodyning.
Figure 6. 3-D measurement with conical scan.
513
-------
Search/Track
tn
Input
l-» 6"
MHz
Acquisition
Detector
*— Calibrate
Figure 7. HRWS frequency tracker.
-------
Figure 8. Control panel.
515
-------
Figure 9. HRWS in pod with access doors open.
-------
en
Figure 10. HRWS attached to UH-1H Aircraft.
-------
Figure 11. Completed installation of HRWS.
-------
20
o I?
o ''
OJ
i- >-
.— 0)
CJ Q.
o
—I en
UJ 1.
>2 8
£U
4 -*
0
«•
»
* M»
» *»»•»•*
NMM»
-i-
-I-
-i •
+ = Digital Data
* = Analog Data
M = Analog and Digital
Data Coincide
N»»»ttNMM»
• N»
»• •
• • »«
M <
» • •
•MMHt • » MM
• •
*•• •••
"72-
144 216
DEGREES
288 360
Figure 12. Full scan windfield measurements at 1-meter hover at a range of 7 meters.
40-
35
30
25
20
15
10
00
cz.
o
a
DOWNWASH PROFILE: WINDTUNNEL (LANDGREBE AND BENNETT)
I I
2.3
10 15 20 3050 100
Range (feet)
Figure 13. Downwash profile of AH-1 in wind tunnel.
519
-------
40
35
DOWNWASH PROFILE: 100 FOOT HOVER
30
in
ec.
25
20
15
10
4 f
' ' in
15 20 30 50 1
2.3
10
00
Range (feet)
Figure 14. UH-1H downwash profile out of ground effect hover.
520
-------
4C
35-
3C-
25
oo
ex.
1
ic-
DOWNWASH PROFILE: 20 FOOT HOVER
I
2.3
3
10
15 20 30 50 100
Range (feet)
Figure 15. UH-1H downwash profile in ground effect hover.
521
-------
horizontal
Jl slant length
ft range
horizontal separation of sample po
radius of conical scan » jj sin 17°
Ints . I/?
118°
V cone angle of separation of sample points
I' angular separation of sample points « 29°
O angle of elevation of sensor boreslght line • 11
Figure 16. Geometric layout of HRWS on BOB + annometer array and
axis of laser crosswind sensor.
522
-------
to
ce
o
LU
>
I 1
Time (minutes)
Figure 17. One-minute averages for 500-m range.
HRWSdata
BOR annometer data
523
-------
o
UJ
to
1/1
ee.
I I I I I I I I I I I I 1 I
Time (minutes)
Figure 18. One-minute averages for 500-m range.
HRWS data
LCSdata
524
-------
in
Of
r - _
I I
Time (minutes)
Figure 19. Single annometer at 24 m
HRWS focused at 24 m
Elapsed Time
(minutes)
1
2
3
4
5
6
7
HRWS (m/s)
Single Focus
5.9
4.4
4.8
5.0
4.0
4.1
4.6
500m AV
6.0
5.1
4.9
5.4
4.5
4.5
5.0
BOR (m/s)
Single Annometer
5.6
4.9
4.6
4.5
2.3
5.5
5.3
Average
5.2
4.4
4.3
4.7
3.4
4.8
5.2
LCS
5.0
4.2
4.1
4.7
3.5
4.5
5.0
Figure 20. Tabulated summary of typical HRWS/BOR comparison.
525
-------
526
-------
WORKSHOP REPORTS
527
-------
528
-------
WORKSHOP I
EXPERIMENT DESIGN, MEASUREMENT STRATEGY AND TACTICS, COMMUNICATIONS REQUIREMENTS
Chairman
Roy Evans
Environment Sciences Research
Laboratory
Abdul J. Alkezweeny
Kurt Anlauf
Donald L. Blumenthal
Richard Coulter
Rosa de Pena
Gary Eaton
Roy Evans
G. T. Gay
Paul Lioy
Walter A. Lyons
Michael MacCracken
Peter H. McMurray
Leonard Newman
Rapporteur
Cary Eaton
Research Triangle Institute
Participants
Brynjulf Ottar
Rudolf Pueschel
Joseph Sickles, II
Gary Stensland
George Sverdrup
Gary Tannahill
Frank B. Tatom
Edward Uthe
William M. Vaughan
Scott Wagner
Kenneth T. Whitby
George Wolff
529
-------
EXPERIMENTAL DESIGN, MEASUREMENT STRATEGY AND TACTICS,
COMMUNICATIONS REQUIREMENTS
INTRODUCTION
The overall purpose of the Sulfur Transport and Transformation Experiment
(STATE) program is to perform measurements to develop and to evaluate regional
models of the formation and transport of visibility-reducing aerosols. Models
will be used for interpretive or diagnostic purposes and ultimately for predic-
tive purposes to allow States and local agencies to estimate concentrations of
these pollutants, considering transboundary transport.
One goal of Project PEPE is the investigation of multiday gas-to-particle
conversion chemistry and aerosol formation processes; questions such as the
differences in chemical processes when urban or point source plumes are injected
into clean air and into dirty air (2-3 days old) should be addressed. The emphasis
is on the formation of sulfates and nitrates as opposed to the prediction of the
formation of ozone, although ozone may be used as a predictor for the formation
of sulfate and nitrate aerosols. Chemical measurements will not emphasize the
measurement and speciation of hydrocarbons, although limited sampling and specia-
tion should be performed. Visibility measurements should be a part of the
Project PEPE experimental program; some detailed chemical species determination
should be performed, probably through X-ray fluorescence (XRF). We would like
to differentiate between nitrate species such as nitric acid and ammonium nitrate,
but we may be forced to settle for total nitrate.
Another goal of project PEPE is to measure the vertical dispersion and
layering of the atmosphere to provide modelers with information to determine the
number of layers needed in air quality simulation models. We wish to follow
specific air masses for periods of several days to observe the buildup of concen-
trations in polluted air masses.
The time span of Project PEPE should be mid-July through August of 1980
primarily for logistical reasons. During this period we can expect perhaps two
PEPE's, each about 1 week in duration. In the interval between PEPE's, the
530
-------
experimental resources of the project will be devoted to studies of the dis-
persion and transformation of power plant and urban plumes.
The following mode of operation is envisioned: Certain facilities and
equipment must be supplied by the prime contractor for the project, as discussed
in the RFP. In addition, the Regional Field Studies Office (RFSO) may elect to
fund other governmental and university group participation in the project. It
will be the contractor's responsibility to provide some support to and coordin-
ation of these groups to produce a coherent program. An RFSO field team will
provide final project guidance in the field.
SCENARIO FOR STUDY OF A PEPE
The meteorological forecast data will be monitored to provide as much lead
time as possible for planning a PEPE experiment. It is expected that the occur-
rence of a region of slow-moving air can be anticipated 3 to 4 days in advance,
which would then start a "PEPE watch", i.e., notification that an appropriate
situation for study may be developing. If the shorter-period (36- or 48-hr)
forecasts confirm the previous indications, then the deployment and readiness
alerts would be issued for the actual experiment. The criterion for the start of
an experiment is that an air parcel will move slowly, or perhaps stagnate, over
an area of substantial emissions, with the subsequent motion of that parcel
forecast to remain relatively slow (less than 6 ms ) for at least 36 hr. The
definition of the initial parcel is the area of a 100- to 200-km square, or
roughly 10 counties.
When the experimental start is set, a tracer release crew would be dis-
patched to the anticipated center of the (temporary) stagnation area, where they
would release the tracer. For at least the first experiment, release would be
during the daytime hours of rapid vertical mixing, permitting release at the
surface. A tetroon release crew would go to the same point and release a cluster
of tetroons—perhaps two triads ballasted to float at 0.4 and 0.8 times the
mixing depth. Aircraft completely equipped for mechanism studies would start
sampling the parcel by means of spirals, perimeter flights, X-pattern flights,
and perhaps other patterns to measure the initial conditions within the parcel.
These aircraft would continue to sample the "parcel" at perhaps 6-hr intervals,
following the assumed centroid of the parcel wherever it might move for as long
as possible. Mobile PIBAL crews would be dispatched in leap-frog fashion to
531
-------
make frequent soundings within the parcel. Horizontal shearing and mixing will
change the air in the parcel considerably over the anticipated period of several
days that it may be tracked; the definition of the parcel will continue to be an
area about the assumed centroid of the initial parcel, to be identified by
tetroons when possible and by trajectory calculations otherwise. These measure-
ments in a Lagrangian frame will provide the data for mechanism studies as the
pollutants in the parcel age and are reinforced by freshly emitted pollutants
along the path followed.
Surface sampling vans would be deployed to perform stationary sampling at
locations the parcel is expected to cross over. To the extent possible, these
locations would be at existent monitoring stations with ground power available.
However, if the driving time to sampling time ratio were to be excessive, one or
more mobile vans would continue to make local measurements near the field base.
Sequential tracer samplers would have to be deployed well in advance of an
experiment; a regional grid would be the most feasible pattern for the samplers.
Notification would be provided for start time and sampling duration for each
point in the grid.
To provide a very large-scale picture of conditions before and during the
Lagrangian measurements, the NASA planes would make regional overflights of
roughly 2,500 km total length in flight patterns that might consist of parallel
traverses and cross flights. These overflights would start on the day of tracer
and tetroon release, or, preferably, on the preceding day, and continue at 12-hr
intervals for the duration of the total episode. Flights beginning at approxi-
mately noon and midnight would serve to characterize regional distributions near
times of maximum mixing and stability, respectively.
Other aircraft would make regional flights within the boundary layer to
provide direct compositional measurements. The flights would generally be made
over preselected flight paths, which would remain fixed for a number of flights,
to provide a temporal sequence in a Eulerian frame. These flights might total
1,000 km in length, and would cover a smaller grid contained within the grid
overflown by the NASA planes. While these two grids would probably be fixed for
the duration of an episode, they could also be shifted if the continued develop-
ment of the episode, either by formation or transport, were indicated to be
moving outside the initially specified location. The relative sizes of the
three regions discussed are shown in Figure 1-1.
532
-------
Since the end of an episode is frequently attended by the development of
showers as a new air mass approaches the area, the probability of "dirty" rain
is rather high, and an alert to an intensive cloud and rain sampling experiment
could be issued because the extent of the polluted air mass feeding into the
shower area would be known.
STUDY STRATEGY
The above scenario of a PEPE has the highest priority for the entire duration
of the study. However, because the weather sequences during the period chosen
for study cannot be preordained, considerable flexibility in the actual conduct
of the study is required to maximize the amount of useful information obtained
during the study.
Since a PEPE of sufficient predictability and extent may not occur during
significant intervals of the study, other kinds of studies will also be performed;
studies that can be completed without interfering with possible PEPE experiments.
The most frequent type will be plume experiments during which effluent from
isolated sources such as power plants or from a source complex such as an urban
area will be studied to better characterize the input into PEPE's. These experi-
ments will be conducted in a manner similar to prior plume studies, with spatial
scales of the order of 250 km downwind of the source and corresponding durations
of up to 24 hr on occasion. The aircraft equipped for mechanism studies will be
used primarily, with a number of specific scenarios to be used for individual
experiments. As one example, there may be days when plumes from two power
plants remain separated horizontally for a long distance, but one plume mixes
with the HC- and NO -rich air from an urban source. It would be of interest,
/\
then, to measure the differential influence of the urban source on sulfate
conversion in the two plumes. Other process, or mechanism, studies could be
performed, such as opportunistic studies of the pollutant composition of air
from recently dissipated clouds, or the role of pre-existent pollutants in
conversion processes to enhance data from the multiday Lagrangian measurements,
and many others. These experiments would be planned and conducted on a day-by-
day basis, with a scenario specified about 12 hr prior to initiation of an
experiment, and with active communication with experimental platforms, and
feedback to field headquarters during the conduct of each experiment.
533
-------
FACILITIES AND EQUIPMENT
Study facilities and equipment recommended by the working group are dis-
cussed in the following paragraphs. In general, the working group ha:; recommend-
ed numbers and general descriptions of equipment that should be used.
In Situ Chemical Measurement Aircraft
The working group recommended that a total of four in situ measurement
aircraft should be involved, two of which are envisioned as performing general
"Eulerian" frame measurements of the areal extent and concentrations of pollu-
tants within the dirty air mass. These two aircraft will be equipped with in-
strumentation for criteria air pollutants and other parameters in what we have
described as the "base package." Two additional aircraft should be more exten-
sively instrumented to investigate gas-to-particle conversion chemistry and
vertical and horizontal dispersion in a "Lagrangian frame," following the marked
air parcel as far as logistical considerations will permit.
The base package (to be present in all four aircrafts) should include at
least the following parameters:
Altitude (pressure altimeter)
Location (by DME/VOR)
Ozone
B .(with heated nephelometer)
S02 (1 ppb sensitivity)
Temperature
Dewpoint
Turbulence
High-volume samples for particulates
Filter packs for sulfates and S02 (BNL system)
Continuous sulfate monitor
Condensation nuclei monitor
This equipment, when possible, should be previously used, field-tested,
preferably off-the-shelf devices. An analyzer for N0/N02 should be included in
the base package if an instrument is available in the summer of 1980 that will
provide 0-2 ppb sensitivity. Otherwise, the cost in power and weight require-
ments may not be justified because the aircraft envisioned in this study will
be in the small, light, twin-engine category. The inclusion of a continuous
534
-------
sulfate monitor in the "base package" should also be evaluated, balancing the
gain in information against the increased maintenance and calibration require-
ments. Position location to ±5 miles will be satisfactory for PEPE. measurements,
although more stringent requirements may be imposed by the urban and power-
plant plume studies.
The two "Lagrangian-frame" aircraft should be more extensively instru-
mented, each one including some or all of the following capabilities:
Continuous sulfates
Electrical aerosol analyzer
Gas grab samples for hydrocarbons
Perfluorocarbon samples
Real-time tracer measurements
Cascade impactor (cut sizes: 0.1 pm, 0.3 urn, 1 urn)
Filter sampling capability to allow chemical
and elemental analyses of particulates
Radiometers—UV and visible light, upward
and downward looking for radiation budgets
The actual inclusion of all of the above equipment at once in each aircraft
may not be necessary depending on the objectives of the chemical investigations.
Other measurements that would be desirable (but may not be feasible) include
aldehydes and formaldehyde; total hydrocarbons by photoionization.
The aircraft should be capable of cruising speeds of at least 150 knots,
preferably 200 knots, and should have a flight range of the order of 5 hr or
more with a full complement of instrumentation. Aircraft with slower speeds or
shorter ranges will probably not be able to traverse the Eulerian frame within
reasonable times.
At least one subsidiary or remote base of operations should be prepared for
the Lagrangian aircraft to enable them to follow the marked parcels through a
multiday path. Since these aircraft will operate more or less continuously after
beginning to track a parcel, each aircraft is likely to need two full crews,
both pilots and scientific personnel. Remote bases may also be necessary for
the Eulerian frame aircraft.
Each aircraft involved in the project must be equipped with a radio link
separate from the airways communications frequencies to permit communication
with the Mission Control Facility at least once an hour.
535
-------
Lidar Aircraft
The working group recommends the inclusion of at least two aircraft equipped
with lidar type devices for measurement of range-resolved light scattering from
aerosols and of total (vertically integrated) concentrations of either $62 or 03.
It is assumed that NASA-Lang!ey will fly their Electra equipped with a DIAL
system and that one other airborne aerosol lidar, either the EPA-Las Vegas de-
vice or the SRI lidar, will be available. NASA will test the DIAL system during
the summer of 1979 to determine actual sensitivities of the device for S02 and
03. Measurement of S02 would be most valuable to Project PEPE, provided that
actual sensitivity for S02 is sufficient to quantitate expected ambient levels.
Ozone should be measured otherwise. Flights at altitudes of 10,000 ft will be
necessary for both lidar aircraft to include the total expected mixing depths;
this may or may not require revision of existing instruments.
If at all possible, the Electra should be equipped with in situ instru-
mentation of 03, NO, N02, S02, nephelometer, temperature, dewpoint, and BNL filter
pack for sulfates and S02. The other lidar aircraft should be fitted with an
S02-sulfate (BNL) filter pack. Both aircraft should be equipped with trans-
ceivers to allow communication with the Mission Control Facility.
Fixed Ground Sampling Networks
EPRI-SURE Network--
Data from the nine Class I sites of the existing EPRI network and six
anticipated additional Class I stations should be utilized in this study under a
cooperative agreement. The locations of the nine sites are: Montague, MA;
Scranton, PA; Indian River, DE; Duncan Falls, OH; Rockport, IN; Giles County,
TN; Fort Wayne, IN; Research Triangle Park, NC; and Lewisburg, WV. At least
four of these are in the Greater Ohio Valley region. These sites are equipped
with 03, NO, NO , and S02 gas phase analyzers and an integrating nephelometer.
/\
Other samplers collect particulate matter for data analysis of sulfate, nitrate,
and acid precipitation. EPRI plans to install Doppler acoustic sounders for
determination of winds. The workshop recommends that these be installed as soon
as possible and that additional three-wavelength nephelometers and radiometers
also be installed. Size-fractionating samplers would also be desirable.
It is recommended that data from the sulfate and nitrate samplers have a
time resolution of at least 6 hr during PEPE experiments.
536
-------
Close coordination between PEPE and EPRI network personnel would be es-
sential.
SURE and MAP3S Precipitation Network--
SURE Class I sites now have colocated precipitation samplers. The Multi-
state Atmospheric Power Production Pollution Study (MAP3S) system will have
similar samplers at approximately eight sites in 1980. Samples are analyzed
for sulfate, nitrate, pH, and some metals.
National Air Sampling Network—
The National Air Sampling Network (NASN) has several sites within the study
area that measure S02, 03, CO, and total suspended particulate by high-volume
sampler. Sulfate analyses are performed on the TSP filters. This information
would be useful in determining sulfate buildup.
Inhaled Particulate Monitor Network--
The Inhaled Particulate Monitor Network (IPMN) will have approximately two
sites per State. The systems employ dichotomous samplers with particle size
cuts at 2.5 and 15 urn.
Ohio River Basin Study--
The Ohio River Basin Study (ORBS) system will consist of three sites. The
emphasis of measurements will be on nitrates, nitric acid, and detailed analysis
of aerosols. Criteria pollutants will also be monitored.
Tennessee Valley Authority Network—
The Tennessee Valley Authority (TVA) maintains a sampling network around
each of its major power generation facilities. Emphasis is on S02 measurement
although high-volume samples, meteorological data, and some ozone data are also
available. Possibly four TVA sites might provide data to define the boundary
conditions along the southern edge of the Greater Ohio River Valley.
National Oceanic and Atmospheric Administration Turbidity Network--
Turbidity data from the National Oceanic and Atmospheric Administration
(NOAA) system are acquired by use of a pyroheliometer and should be part of the
PEPE data base.
National Weather Service Upper Air Soundings
National Weather Service (NWS) releases of rawinsondes are normally made
537
-------
twice per day. It is hoped that the frequency of releases could be increased to
four times per day at the following 10 sites: Greensboro, NC; Buffalo, NY;
Pittsburgh, PA; Huntington, WV; Nashville, TN; Dayton, OH; Detroit, MI; Salem,
IL; Dulles International Airport; and Fort Totten, NY.
Deployable Ground Sampling Systems
Instrumented Vans—
It is hoped that three vans, each equipped with sophisticated instrumenta-
tion, will be available for this study. These vans will be in use near urban
and industrial centers for plume study, and, upon prediction of a PEPE, can be
deployed to preselected sites where electrical power is readily available.
The recommended vans are Bob Stevens' EPA van, Kenneth Whitby's University
of Minnesota van, and the General Motors Corporation van.
Each of the vans will be assumed to be equipped with the following analyzers
or samplers: S02, NO-NO , 03, CO, LBL dichotomous sampler, electrical aerosol
/\
analyzer (EAA), optical particle counter (OPC), cascade impactor, heated and
unheated nephelometers, and COSPEC, as well as calibration systems and NBS
traceable standards.
In addition, it is suggested that the following equipment be installed in
one or more of these vans: HN03 filter, radiometer, pyrheliometer, 10-m meteoro-
logical tower, and continuous sulfate analyzer.
Additional Vehicles—
Automobiles could be equipped to measure ozone (Dasibi) and sulfate (by
filter pack). These vehicles could travel highways during a PEPE to provide
additional ground coverage of the event.
Tracer Sampling Network—
A tracer sampling network using samplers such as those described by Gilbert
Ferber should be included in the project and deployed over the region on a grid
determined by the number of samplers that might be available.
Mobile PIBAL
PIBAL releases should be made in both the plume study and in the Lagrangian
mode of the PEPE study. Suggested release points during a PEPE would be near
tetroons in the Lagrangian air parcel.
538
-------
Communications Network
Both the Project PEPE field measurements program, and the power plant and
urban plume studies between PEPE's, will require good communications between the
Mission Control Facility and the individual field units. Communications on a
minute-to-minute basis are probably more important for the plume studies than
for the PEPE field activities, but communications between all field units and
Mission Control at least hourly will be needed for all aspects of the program.
Two methods were suggested by the working group, either of which is probably
adequate. Two-way communications via a separate FM network operating at about
165 MHz and an airborne repeater in a dedicated aircraft would provide minute-
by-minute communications throughout most of the PEPE area. The repeater aircraft
would orbit at an altitude of about 15,000 ft during the entire PEPE and could
move with the Lagrangian parcel if necessary. Off-the-shelf equipment such as
the Motorola or the GE lines would be satisfactory for this purpose, provided
that acquisition was begun about a year in advance of the actual deployment and
that adequate resources were budgeted for the contractor to install and maintain
the equipment. There is some probability that the Lagrangian parcel would move
beyond the range of the repeater at the end of its trajectory.
Radiotelephones provide an alternative means of communication that would
operate over a broader geographic area but would provide only intermittent
communications between field units and Mission Control. Radiotelephones that
are satisfactory for aircraft installation are available either for purchase
(e.g., King Radio) or for rent (from Bell); however, lead times for acquisition
will probably amount to at least several months, and acquisition should be begun
at the beginning of FY80.
High-Altitude Air Force Aerial Photography
The working group recommends that high-altitude aerial photography of the
PEPE be obtained from the U.S. Air Force from ongoing U-2 flights. It is thought
that this information could be obtained at virtually no cost via a request from
EPA to the Air Force. The Las Vegas laboratory already has well-established
channels for requesting this assistance from the Air Force.
Mission Control Facility
An operations headquarters must be established at a central location within
539
-------
the PEPE area to coordinate STATE and PEPE field activities. Its functions will
include:
1. Forecasting the occurrence of the PEPE and the location of its
onset;
2. Acting as field data analysis center;
3. Serving as field laboratory;
4. Maintaining communications with all field units;
5. Assimilating data from satellite imagery to aid in maneuvering
field crews.
Chemical Tracer Injection Team(s)
Chemical tracer injection will include SF6 and up to three perfluorocarbon
compounds, provided satisfactory measurement instrumentation is available by
summer of 1980. Tracer injection support requirements were assumed to be estab-
lished by Workshop 2.
Tetroon Tracking
We have assumed tetroon parcel markers will be utilized in Project PEPE,
with tetroon release to be accomplished by NOAA. Tracking techniques and require-
ments were assumed to be established by Workshop 2.
DATA ARCHIVING
The contractor selected for conduct of the study will be responsible for
reduction and validation of all data collected as part of the contract and, in
addition, will process some data from other experimental platforms. These data
will be delivered to the RFSO on magnetic tape in formats to be developed by
mutual agreement. Data from other sources, including not only other partic-
ipants in the experimental studies, but supplementary data such as NWS meteoro-
logical data and pollutant monitoring data from existent networks, will need to
be included in the total STATE data bank on the PEPE study. A data management
system to incorporate all these diverse data sources and to make data available
for subsequent analysis should be structured before the data start flowing in.
This system should provide format specifications to all participants in the
study, and will require good coordination between the STATE data manager and the
sources of input data.
540
-------
Quality Assurance
With the diversity of sampling platforms and methodologies that will be
used to provide data during the course of the study, a procedure to ensure the
comparability and quality of the collected data is needed. Design of sampling
systems and performance of individual instruments can be checked by adherence to
good procedures, frequent calibrations, and checking of calibration sources.
To provide the final step in a quality assurance program, intercomparisons
should be made between complete sampling systems as used in their respective
platforms. To accomplish this end, all aircraft should make side-by-side
flights at several altitudes under various ambient pollutant conditions, with
any given craft making at least one comparison flight a minimum of three times
during the study. Selection of an open site, such as a small airport, would
permit comparison of aircraft systems with mobile ground systems, which
should be made. In turn, comparisons between mobile ground systems and certain
of the fixed monitoring stations should be included if the fixed stations are to
be used as locations for mobile sampling operations.
MISCELLANEOUS RECOMMENDATIONS
Historical Forecasting of PEPE
A program should be initiated now to review the data and statistical findings
from SURE and MAP3S in order to derive a reasonable technique for forecasting
the occurrence of a PEPE.
Determine Best Study Window Via Climatological Study
A climatological study should be part of the planning stage for two reasons:
(1) to determine the optimum location of tracer injection, and (2) to be a basis
for setting the time window for the study. At present, the study is planned to
begin in July 1980 and extend for 6 weeks.
Cooperative Study With Canada
The involvement of Canadian air pollution agencies, such as the Federal
Atmospheric Environment Services (AES) and the provincial Air Resources Branch
(MOE), is welcomed. EPA will establish contact. The suggested study area would
be immediately northeast of the Eulerian gridwork proposed in the scenario.
Specifically, it would be highly desirable that aircraft (AES) and ground-level
541
-------
support measurement (AES and MOE), including remote sensing and radiosonde
releases, be made in the Lake Ontario and Lake Erie regions. This would greatly
aid in defining the northern boundary conditions of the Eulerian grid.
EMISSION INVENTORIES
To relate the measurements obtained in the study to pollutant sources,
and to provide the requisite input for model development and evaluation, a
reasonably accurate and up-to-date emissions inventory for the region of study
will be required. The updating and improvement of prior inventories sponsored
and performed by EPRI and the MAP3S program should provide a useable inventory
of sulfur emissions. Some additional work will be required to compile an inven-
tory of NO and other pollutant emissions, with conversion of the EPRI/MAP3S
^
inventory for point sources and updating of NEDS for area source emissions--
particularly the mobile source component.
542
-------
A-Overflight
B—Eulerian
No.—Parcel
Km
300
Figure 1-1. PEPE scenario.
543
-------
544
-------
WORKSHOP II
METEOROLOGY, TRANSPORT, AND MULTIDAY PARCEL TRACKING
Chairman
Harry L. Hamilton, Jr.
Research Triangle Institute
Participants
Wendell Ayers
Michael Chan
Jason Ching
John F. Clarke
Walter F. Dabberdt
C. Raymond Dickson
James G. Edinger
Gilbert F. Ferber
Noor Gillani
James G. Haidt
Harry L. Hamilton, Jr.
Rudolf B. Husar
Edwin Keitz
William King
Bryan W. Lambeth
Rapporteur
William King
Research Triangle Institute
Mei-Kao Liu
Ron Meyers
David Patterson
Earl F. Pound
Perry Samson
Walter J. Saucier
Gene E. Start
Eugene Tong
Fred M. Vukovich
T. L. Waldron
Gerald Watson
Marvin L. Wesely
James J. B. Worth
Bernard Zak
545
-------
METEOROLOGY, TRANSPORT AND MULTIDAY PARCEL TRACKING
CONCLUSIONS
Meteorologists can provide a 24-hr prediction of the onset of a synoptic
situation conducive to the development of a Persistent Elevated Pollution
Episode, if it is possible to adequately define the conditions that are
sought. (See Recommendation 1.)
Within a developing area of elevated pollution concentrations associated
with a high pressure system, a release point can be selected that will pro-
vide a high probability that an air parcel marking device will be in that
area of elevated pollution concentrations for a multiday period. The resi-
dence time in the area will be dependent on the shape of the high pressure
system and the speed with which the system moves. With regard to areas of
elevated pollution now classified as channeling situations, the residence
time of an air parcel track in the area will be limited by the relationship
between the alongwind extent of area and the wind speed and on the positions
and movement of frontal systems on both the synoptic scale and the mesoscale.
All tracer systems suggested require some development. Accordingly,
generic techniques are considered. Two tracer techniques should be used:
(a) a technique that will approximate, the path of a parcel within the devel-
oping area of elevated pollution, and (b) a technique that will provide data
for the estimation of dispersion within the developing area of elevated pol-
lution. Technique (a) is conceived as a super pressure balloon(s) of some
configuration carrying a suitable device that, with ancillary equipment,
will permit position determination. Technique (b) must be a gaseous tracer,
probably one of the perfluorocarbons. (See Recommendations 3 through 8.)
546
-------
RECOMMENDATIONS
1. A study must be made to identify the synoptic situations that give rise
to PEPE's in the Ohio Valley. This study should use archived surface
and upper air meteorological and air quality data to improve our knowl-
edge of the aerometric conditions associated with identified PEPE's.
One task would be to determine the routine meteorological variables
that can serve as surrogates for actual pollutant measurements; these
surrogates would permit the study to cover a longer historical period
and provide a larger data base for the classification of pertinent
meteorological situations. In addition, this larger data base would
provide data for the determination of the climatological expectancy of
a PEPE in the area selected for the future program. At least one
recent year of meteorological data for which there are available rela-
tively detailed pollution data (and perhaps previously identified
PEPE's) should be reserved as a data set for testing the success of
synoptic type classification as the indicator of a developing PEPE.
This front-end research effort should provide the following as a minimum:
a. A classification of the synoptic regimes and wind flow (at
the surface and aloft) corresponding to the onset, peaking,
decline, and movement of PEPE's.
b. A set of guidelines for both objective and subjective fore-
casting of PEPE's and parcel trajectories.
c. On the basis of historical data and analysis, recommend
candidate regions that are most favorable for PEPE experiments
and most representative of an urban source complex and/or of
a rural industrialized source complex.
2. A study should be undertaken to determine the accuracy of now available
forecast trajectories (TDL) relative to trajectories determined from
actual data. If these forecast trajectories are a reasonable approxi-
mation of the "actual" trajectories, examination of their configuration
(speed, curvature) may provide a useful indication of an incipient PEPE
situation. This study must use current data because predicted trajec-
tories are not archived.
547
-------
3. The technique of locating a super pressure balloon that is being used
as an air parcel tag by radio direction finding from a chase airplane(s)
should be investigated to determine the ease of target acquisition, the
flight pattern required for position determination, and the accuracy of
position determination with and without visual sighting. A preliminary
feasibility investigation should be made to determine equipment devel-
opment time and costs. This investigation should consider including
temperature and pressure data sensing and data transmissions on the
homing signal.
4. The existing tetroon-transponder parcel tagging system is proven and
operational but to assure the capability of multiday tracking, improved
mobility of ground-based radar systems is required. A prototype radar
unit with improved mobility has been assembled, but lead time will be
required to similarly assemble the additional four available units.
5. The feasibility of using the Loran-C or the Omega navigational aid
systems for air parcel tags at 1 km above terrain in the area selected
for the proposed study should be investigated empirically. The develop-
ment of a sufficiently lightweight onboard receiver, data storage system,
and data transmitter presents no difficulty, but, as in the case of the
radio direction finding system, cost estimates, ease of acquisition of
position data from a chase airplane or ground station, and the coverage
provided by the navigational aid systems in the area proposed for the
PEPE system must be determined.
6. The use of a beacon transmitting to a satellite to provide location data
for an air parcel should be studied to determine the possibility of
significant weight reduction of the beacon and battery package so that
the system would be acceptable from a safety standpoint to the Federal
Aviation Administration (FAA). Further, the turnaround time on position
data from the satellite ground station must be determined. The costs
of this system, including data transmission costs and the lead time
required for weight reduction, must be determined.
7. The Sandia LAMP super pressure balloon, if proven as an operating
system, should be considered as an air parcel tracer and measurement
platform.
548
-------
8. The gaseous tracer system using perfluorcarbons requires the demonstra-
tion of the capability of a sequential sampling system. This is planned
for the spring of 1979. This instrument requires laboratory evaluation
of sample concentration. Real-time readout (5-min intervals) of perflu-
orcarbon concentration can be made by another instrument (two-trap) that
is suitable for fixed mobile ground operation or airborne operation (if
power is available). Fabrication lead time and possibly funding will
be required if a sufficient number of the sequential samplers for a
ground network of stations is required for the PEPE program.
The Meteorology and Transport and Multiday Parcel Tracking Workshop
considered at length the requirements for assimilating tagged parcel location
data, gaseous tracer concentration data, low-level wind data, etc., that
would furnish feedback essential to continuing operations. The conclusion
was reached that a time-shared or dedicated computer would be required.
Without attempting to specify the data input required, the group is
adamant that the project must have an onsite weather station with adequate
data sources and personnel to provide specialized forecasting services and
material necessary for thorough operational briefings as required.
To provide finer time resolution to modellers, the frequency of NWS
rawinsonde observations from a selected set of stations in the operational
area should be increased to four per day.
Special wind and temperature vs. height (to 2 km) observations for
plume studies, and subsynoptic time scale observations for Lagrangian parcel
studies will be required to update trajectory forecasts during operations.
A multistation mobile network will be required.
549
-------
550
-------
WORKSHOP III
MODEL DEVELOPMENT AND VALIDATION-MEASUREMENT NEEDS
Chairman
George M. Hidy
Environmental Research and
Technology, Inc.
Francis Binkowski
Samuel C. Coroniti
Kenneth L. Demerjian
Allen C. Dittenhoefer
Ron Drake
John Eckert
Noor Gillani
Robert G. Lamb
Mei-Kao Liu
Man's Lusis
Participants
Rapporteur
Joseph Sickles, II
Research Triangle
Institute
S. Harvey Melfi
Ron Meyers
David Miller
Brand L. Niemann
Francis Pooler
Joseph Sickles, II
Lowell Smith
Kenneth T. Whitby
William E. Wilson
551
-------
MODEL DEVELOPMENT AND VALIDATION—MEASUREMENT NEEDS
The overall goal of the PEPE project was stated to be the characteriza-
tion of transboundary, interstate air pollution involving large-scale regional
events. The indicator for such events is elevated particulate sulfate con-
centration. The scale of the events extends horizontally over 100 to 1,000
km and involves buildup of 2 to 5 days. Using modeling as an interpretation
tool, the specific scientific objectives of the study that are relevant to
model evaluation were given tentative priorities based on (a) importance to
the PEPE modeling, (b) available knowledge, and (c) scientific significance
of understanding. Priority one is highest. The topics listed below repre-
sent major areas that remain uncertain after preliminary assessment of
regional studies conducted to date, including those of the EPRI/SURE, MAP3S,
STATE, and the EPA/ITA.
1. Quantification of factors in long-distance transport of pollution
such as
a. The effects of rough terrain in the Appalachians on eastward
travel. (1)
b. Diurnal influences of the boundary layer. (1)
c. Action of clouds on mixing, conversion chemistry, and
removal. (1)
2. Characterization of PEPE air masses
a. Role of thermodynamic and dynamic properties such as
temperature, dew point, mixing, velocity, and sunlight-
cloudiness. (1)
b. Reasons for observed "sharp" changes in concentration
buildup and depletion in PEPE's. (2)
c. The air mass aging character and scale. (2)
d. Characteristics of plumes at distances on the order of
100 km; the nature of concentration gradient in the
PEPE. (3)
552
-------
3. Sulphur and Nitrogen Budgeting
a. Requirements for details in chemical conversion such as
the particle formation and growth rates, and photochemical
vs. other reaction classes. (1)
b. The location and intensity of emissions. NOX,HC (1)
SOX (2)
c. The significance of SO and NO background for regional
/\ /\
model initial and boundary conditions. (2)
d. The role of particle deposition and associated uncertain-
ties in removal of haze. (2)
e. The role of gas removal, including dry and wet deposi-
tion. (3)
4. The merging of regional scale and subgrid scale processes,
including their parameterization (2)
For refinement of conceptual modeling of PEPE behavior, three features
of long-distance transport require additional study at high priority. These
involve questions as to the importance of terrain features, especially the
Appalachians, and transport over the mountains. Also, the transport of
polluted air during conditions of changing stability at night and day needs
additional definition. Virtually nothing is known about the details of
vertical mixing and chemical processing by clouds. Additional characteriza-
tion of the PEPE air masses is needed to improve knowledge for modeling of
thermodynamic-dynamic interactions, especially the apparent sharp changes in
pollution levels and the definition of "plumes" at great distances.
Fundamental to the evaluation of existing air quality models is the
sulfur and nitrogen budgeting in PEPE's. The experimental design should be
aimed at improving knowledge in several areas including the chemical conver-
sion rates for gases and particle formation or growth, the location and
intensity of SO , NO , and NMHC emissions. Resolution of major emissions
/\ /\
contributing to the PEPE should be hourly if possible. Information on SO
/\
emissions and methods for obtaining hourly estimates appear to be available,
but attention is needed to improve NO and NMHC characterization. For
^
improvement of the detailing of chemistry, effort should be made to break
down NMHC emissions into reactivity categories. Removal processes for gases
553
-------
appear to be reasonably well defined at present for modeling purposes. How-
ever, controversy remains concerning the magnitude and availability of par-
ticle removal. This issue needs to be resolved for PEPE, although other
programs may already be studying these questions.
MEASUREMENT REQUIREMENTS
The evaluation and refinement of air quality models for interpretation
of PEPE behavior will require an elaborate set of measurements. The list of
desired parameters is given in Table III-l. These include a range of mea-
surements widely available for air quality and meteorological monitoring as
well as measurements that require special instrumentation of a highly exper-
imental nature. Recommended requirements are identified with model verifica-
tion and evaluation, chemical and aerosol formation processes, and regional
scale fluid dynamics. The measurements are assigned priorities for both
ground and aircraft observations. By suitable selection of sites, the PEPE
project can take advantage of existing meteorological and air quality data
collection including that of the NADB/SAROAD system, EPRI's (proposed)
Regional Air Quality Study (RAQS), the National Weather Service, and the
DOE/DOA/EPRI Precipitation Chemistry networks.
There is concern for the representativeness of observations. Measure-
ments of different spatial and temporal scales may be required. Cell-
averaged (Eulerian) data are needed to be compatible with model predictions.
Ideal measurements on this scale could be provided by satellite. Although
such data may not be available by the time of the proposed study, satellite
imagery may provide real-time definitions of the boundary of the polluted
air parcels.
Burden measurements of integrated species may be conducted either from
the ground or from aircraft. Although these measurements view the atmosphere
as a single layer, they will nevertheless provide useful data against which
model predictions can be compared. Measurements conducted on different time
and spatial scales are necessary for describing subgrid phenomena such as
chemistry and fluid dynamics.
Regional (secondary) pollutants such as sulfates and ozone are relatively
homogeneously distributed. Thus, a low density sampling network monitoring
554
-------
the routine parameters with the standard averaging times will probably be
sufficient for model assessment.
Although the study should capitalize on existing or proposed stations
of opportunity, the following additional station locations should be
considered:
1. Places most likely for PEPE to pass over.
2. Upwind, within, and downwind of selected urban area.
3. Appalachian mountains (complex terrain).
In addition to the "regular" monitoring stations that provide data for
model evaluation, one or two more comprehensive stations are recommended to
provide data needed to improve the current understanding of subgrid phenomena.
Such parameters as aerosol size distribution, aersol water, hydrocarbons, OH,
NH3, and HN03 should be considered.
Airborne measurements, as long as they are not considered to be repre-
sentative of the ground-level concentrations, will also provide useful input.
Their major value will be to provide information on vertical concentration
distributions. In addition, flights conducted in conjunction with tracer
releases will be used to study all parcel behavior in a Lagrangian framework.
Rawindsondes should be released with increased frequency. Six-hour
releases are recommended during the day; whereas three-hour releases may be
needed at night when the atmospheric structure is poorly defined. Additional
measurements using acoustic radar and/or lidar to determine vertical struc-
ture should also be considered at one or two sites.
Modeling Evaluation and Designs
Several models available for application to the PEPE's are summarized
in Table III-2. The models are classed as numerical grid schemes, trajec-
tory/plume, or statistical and other. Most of the models have capabilities
of dealing with transport, mixing, emission distributions, at least linear
oxidation chemistry for sulfur dioxide, and dry deposition. The models that
have been used directly for PEPE applications include the DOE/MAP3S series,
the Teknekron/Prahm scheme, and the ERT/SURE model.
Although the models could be exercised for design studies, the panel
felt that review of the SURE and MAP3S data and other historical aerometric
data would be most useful for design of the experiments. These data are
555
-------
becoming available and should be examined by interested investigators in the
next several months.
The concept of modularization of models was discussed. In general, the
panel believes that such a framework would be helpful for providing a uniform
format for investigator's results.
556
-------
TABLE III-l. DESIRED MEASUREMENTS FOR EVALUATION OF MODELS
(PARENTHESES ARE PRIORITY VALUE)
Parameter
S02
03
NMHC (including specific
breakdown & oxygenates)
NO, N02
PAN
HN03
NH3
Tracers (gases and tetroons)
OH
H202
S04
TSP (<15 Mm)
RSP (<3 Mm)
N03
Carbon
Organic carbon
Aerosol size distribution
Light scattering coefficient
Ground
(1)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(2)
(2)
(2)
(2)
Aircraft
(1)
(1)
(2)
(1)
(2)
(1)
(2)
(1)
(1)
(1)
(1)
(1)
(1)
(1)
(2)
(2)
(2)
(2)
Required for*
v,c
v,c
c
v,c
c
c
c
D.V
C
C
v,c
V
v.c
c.v
c
c
c
v,c
Availability
monitoring
EPRI/RAQS SAROAD
EPRI/RAQS SAROAD
EPRI/RAQS SAROAD
EPRI/RAQS SAROAD
EPRI/RAQS
EPRI/RAQS
EPRI/RAQS
EPRI/RAQS
EPRI/RAQS
Comments
Need some integral burden
Need some integral burden
At 1-2 locations;
need some information
Need some integral burden
At 1-2 locations
At 1-2 locations
At 1-2 locations
Some continuous desirable
Need some information
At 1-2 locations
*See footnotes at end of table.
(continued)
-------
TABLE Illrl. (continued)
3 km)
T
OP
Solar radiation
Cloudiness
Ground
(1)
(1)
(3)
(1)
(1)
(1)
(1)
(1)
(1)
Aircraft
(1)
(1)
(3)
(1)
(1)
(1)
(1)
(1)
(1)
Required for*
D,V
C,D
C
C
D,V
0
0
D,C
D,C
Availability
monitoring
Hydro serv. ,EPRI
MAP3S, NC-141
NWS
NWS
NWS
NWS
NWS
Comments
Integral vertical
aerosol burden
At 1-2 locations
Surface and aloft
Surface and aloft
Surface and aloft
*V = Model verification
C = Chemistry
D = Dynamics
-------
TABLE III-2. REGIONAL AIR QUALITY MODELS AVAILABLE FOR DESIGN STUDIES AND EVALUATION
Originator
BNL/MAP3S
ANL
ANL/Peters
SAI
en
en •
ID
ERT/SURE
TEKNEKRON/
PRAHM
Winds
Diagnostic-
Prognostic
Interpol
(Climate
only)
Interpol
Interpol
Interpol
Winds 100M
Grid
Layers / Dimensions Chemistry
3-5 X.Y.Z Linear
Nonlinear
Homogeneous
5 X.Y.Z None
? X.Y.Z Steady-
State
Homogeneous
3-1/2 X.Y.Z Linear
Nonlinear
Homogeneous
3-5 X.Y.Z Linear
Nonlinear
Homogeneous
Semi -empirical
In 03> H20
1-3 X.Y.Z Linear
Originator
BNL
PNL
NOAA/ARL
ERT/ELSTAR
SRI.
EUROMAP
LRTAP
Trajectory/pi ume
Layers Dimensions Chemistry
12 X.Y.Z Linear & Wet
Deposition
1-X X,Y Linear & Wet
Deposition
? X,Y Linear & Wet
Deposition
3+ X,Y,Z Linear;
Nonlinear
Homogenous
2 X.Y Linear & Wet
Deposition
X,Y Linear & Wet
Deposition
Statistical/other
Originator Description
ERT Principle
Component
Analysis-
Statistical
Time &
Space
(Note--
Needs data
base)
ANL Statistic
Trajector
(pseudo-spectral)
-------
WORKSHOP IV, PART 1
INSTRUMENTATION FOR STUDY OF TRANSFORMATION
Chairman
Robert K. Stevens
U.S. Environmental
Protection Agency
Frank Allario
Kurt Anlauf
Dagmar Cronn
Thomas G. Dzubay
Edward Friedman
Lee Langan
Peter H. McMurray
Participants
Rapporteur
Madhav Ranade
Research Triangle Institute
Millan M. Millan
Leonard Newman
Madhav Ranade
Chester W. Spicer
Gary Stensland
Robert K. Stevens
James Tommerdahl
560
-------
INSTRUMENTATION—TRANSFORMATION MEASUREMENTS
In conjunction with the Seminar on Persistent Elevated Pollution
Episodes (PEPE) held at Durham, N.C., on March 19-23, 1979, a workshop on
instrumentation needed for measurements under the PEPE program was conducted.
The objectives of this workshop were to identify the instruments presently
available and to define research tasks (needed to be funded in early FY1980)
for developing new instruments by the summer of 1980. In the first part of
the workshop, instrumentation for studying transformation in pollutant
plumes was considered. The measurement areas focused on during this discus-
sion were:
1. Sulfate speciation and acidity,
2. Nitrates,
3. Formaldehyde and aldehydes,
4. Particulate, and vapor phase organics,
5. Real-time particle size distribution measurements,
6. Halocarbons,
7. Adaptation of instruments to the aircraft platform,
8. Precipitation, and
9. Hydroxyl radical.
SULFATE SPECIATION AND ACIDITY
For the PEPE study, it is desirable to quantitatively monitor sulfur in
sulfate form and to know proportions of acid and ammonium sulfate.
Analytical methods presently used for sulfates and total acidity will
be sensitive enough for the PEPE measurements. It has been observed at
Brookhaven National Laboratory and EPA that size fractionation prior to the
collection of the sample is necessary to avoid influence of basic coarse
particles in the acidity measurements. It was suggested that separation of
particles above 2 |jm be effected before sample collection.
561
-------
The analytical methods available for total acidity include Gran titra-
tion, benzaldehyde extraction, and radiolabeled titration. Of these, the
radiolabeled titration being developed by EPA/ESRL is the most sensitive and
utilizes shorter sampling times.
Recommended collection media for sample collection are Teflon and acid
washed quartz filters because these do not produce artifact formation. The
filters could be used for nitrates as well as sulfates. If sampling for
carbon analysis is achieved on the same filter, quartz could be the filter
of choice. If used for sulfate and nitrate measurements alone, the Teflon
will be preferable.
Sampling periods of 1/2 to 1 hr at a rate of 500 1pm may be needed to
obtain samples for the sulfate, nitrate, and carbon analysis.
Tandem filters for ambient S02 measurements are needed as backup for
real-time S02 and particulate sulfur.
Continuous sulfur measurement by flame photometric detection (FPD) may
be used for real-time monitoring of sulfates and S02 by the following schemes:
1. Diffusion denuder for S02 followed by FPD to differentiate between
S02 and particulate sulfur;
2. Electrostatic separation of particulate sulfur to differentiate
between S02 and particulate sulfur;
3. Diffusion danuder including neutralization with NH3 to present
to the flame only (NH4)2 S04. (This avoids H2S04 losses in the
burner block and calibration difficulties.)
The FPD system adaptability to aircraft use is considered marginal at
present due to problems associated with pressure changes.
Information regarding particle size dependence of sulfate speciation is
desirable in the PEPE. At a minimum, speciation should be determined for
two classes—suboptical and optical. Three techniques presently available
are the vacuum impactor, micro-orifice impactor, and a diffusion battery
with a precutting slit. The vacuum impactor requires a large pump, provides
a small sample, and is not well-character!zed with respect to losses. The
micro-orifice impactor (University of Minnesota) collects particles below
0.1 pm on Teflon substrates suitable for XRF analysis. The performance of
this device needs to be documented by further laboratory and field testing.
The diffusion battery system can remove particles smaller that 0.1 pm from
562
-------
the sample and has been used in several previous studies. The Tandem Differ-
ential Mobility Analyzer (TDMA--University of Minnesota) may also be useful
to investigate the hygroscopic nature of the submicron aerosol sample in
several size ranges down to 0.01 urn.
NITRATES
The state-of-the-art of nitrate sampling and measurement needs consider-
able improvement. Samplers used for sulfur measurements may be used. Samp-
ling time should be kept low requiring high flow rates. Negative artifact
formation is currently under investigation. Sample stability over several
days has been observed; for longer periods, storage in a cold inert atmos-
phere (such as Argon) is recommended. The Teflon and acid-washed quartz
filter would be compatible with nitrate measurement.
The analytical methods include: (a) ion chromatography, (b) formation
of nitrobenzene followed by electron capture gas'chromatography, and (c)
flash vaporization followed by chemiluminescence. Of these, the last method
has received limited testing.
NITRIC ACID
Use of tandem filters utilizing nylon or cotton backup filters followed
by extraction for GC or 1C analysis is recommended for nitric acid measure-
ment. Use of nylon or cotton wool following a filter has been successful in
the past. A diffusion denuder for HN03 also has the potential to determine
HN03.
FORMALDEHYDE AND ALDEHYDES
Currently, formaldehyde analysis is conducted after collection in liquid
followed by colorimetry. Use of GCMS is required for higher aldehydes.
There are no reliable and relatively inexpensive methods presently available.
PARTICULATE AND VAPOR PHASE ORGANICS
Currently, total carbon measurement as well as differentiation between
volatile and nonvolatile carbon is practical. Development of a special high-
volume sampler is needed for GCMS work. Teflon filter analysis for carbon
by nuclear methods (PESA or GRACE) is currently under investigation. A
563
-------
cartridge to collect gaseous organics is being developed by RTI under EPA
sponsorship.
It was also expressed by the participants that development of a device
to allow total carbon analysis by size is desirable because the organic aer-
osol may influence the optical properties of the atmosphere significantly.
An inexpensive and real-time nonmethane hydrocarbon monitor (NMHC) is
being developed by ESRL/EPA, which may be useful for both ground-level and
airborne monitoring. The system will be available for field studies by the
summer of 1980.
REAL-TIME PARTICLE SIZE DISTRIBUTION MEASUREMENT
Currently the electrical aerosol analyzer (EAA) coupled with an optical
particle counter (OPC) is used for the real-time size distribution measure-
ment. Agreement between the mass data and the estimated means from the
volume distribution data from the EAA/OPC system is not always consistent.
Causes for this discrepancy may be due to humidity and should be investi-
gated before summer 1980.
HALOCARBONS
During the seminar, perfluorocarbon was proposed as a tracer for air
parcel movement. Technology for release and monitoring needs to be refined.
Work should begin immediately. Use of state-of-the-art technology to follow
urban blobs by using the freons emitted by the major sources was also sug-
gested.
ADAPTATION OF INSTRUMENTS TO THE AIRCRAFT PLATFORMS
From economic considerations, single- and twin-engine airplanes are
suited for aircraft measurement. Due to the weight and space limitations of
the payload, packaging of the instruments and auxiliary components needs to
be examined. A nominal 1,000-1b load capacity and 75 A at 26 V power capacity
is expected to be available.
Instruments considered for mounting in the aircraft need to be trimmed,
if possible, to allow efficient use of space and load. For example, light-
weight and/or d.c. pumps are available that .could reduce the weight for
several sampling devices. Spares for the instrument complement might be
564
-------
augmented in some cases by using subsystems such as burner-block, and PM tube
assemblies as opposed to complete instruments.
The aerosol sample inlets for aircraft need to be developed by summer
1980. State-of-the-art sampling inlets for use on airborne studies are not
optimized for aerosol sampling and probably have significant losses of par-
ticles <5mm in diameter. Provision of a 15-mm cutoff for the inlets should
also be considered to compare with the inhalable particulate matter (IPM)
measurements made at ground level.
PRECIPITATION
Sampling in clouds and at ground level was considered to be necessary to
study precipitation chemistry associated with PEPE. In airborne measurements,
HNO, in the cloud water should be considered. A study of the cloud sampling
with regard to location (e.g., at an edge or in the middle) and collection
efficiency is recommended.
Ground-based samplers are currently available that allow measurement of
acidity, SO,, NO.,, and trace metal analysis. However, questions regarding
effects of wind speed and location that may result in size selectivity need to
be resolved.
HYDROXYL RADICALS
Hydroxyl radicals may play a role in the S02 to sulfate conversion. The
resonance fluorescence method developed at Washington State University may be
useful. Estimated cost for one measurement is $5,000-$10,000.
SUMMARY
Table IV-1 summarizes the status of instrumentation discussed above.
565
-------
TABLE IV-1. STATUS OF INSTRUMENTATION FOR PEPE
on
O>
Oi
Parameter
S02
NOX, 03
NMHC*
Organics
HN03
H+
SOtT
S(aerosol)
N03=
Trace metals
Halocarbons
Hydroxyl
radical
Methods
Chemi luminescence
Tandem filter
Chemi 1 umi nescence
photometric
GFC/FIDt
GC/MS
(a)Chemi luminescence
(b)Sample & collect
l.Gran titration
2.Radiolabeling
Sample & collect
wet chemistry
l.FPD
2.XRF
Sample & collect
Sample & collect
Semi real time GC/EC
Sample & collect
Response time
1 s
30 min-24 hr
5 s
1 s
Collect &
analyze
(a) 1 s
(b) 2.20 min-
24 hr
30 min-24 hr
10 min-24 hr
30 min
Real-time
5 min
30 min
1-2 hr
10 min
2-24 hr
Sensitivity
1-5 ppb
2 ppb
30 ppb
1 ppb
0.1 ppb
0.1 ppb
1 yg/m3
1 yg/m3
1 yg/m3
3
1 yg/m
1:1 yg/m3
1-50 yg/m3
Pi cog rams
ppb
Availability
Now
Now
Summer 1980 *
Under development
Now
Now
Now
Now
Now
Now
Now
Now
Ground
level
X
X
X
X
X
X
X
X
X
X
X
X
Air-
borne
X
X
X
X
X
X
X
X
X
X
X
X
*Nonmethane hydrocarbons.
tGas filter correlation spectrometry.
*0nly one will be available.
-------
WORKSHOP IV, PART 2
INSTRUMENTATION FOR REMOTE SENSING
Chairman
Frank Allario
Langley Research Center
L. W. Chaney
Dagmar Cronn
Richard Coulter
John Eckert
Edward Friedman
James G. Haidt
Edwin Keitz
Lee Langan
Participants
Rapperteur
Madhav Ranade
Research Triangle
Institute
W. A. McClemy
Millan M. Millan
Earl Pound
Madhav Ranade
Robert K. Stevens
James Tommerdahl
Edward Uthe
567
-------
INSTRUMENTATION FOR REMOTE SENSING
INTRODUCTION
This Workshop was asked to address two questions:
1. What active airborne remote sensors are available for use in the
EPA 1980 Regional Field Studies with respect to the PEPE program?
2. What tasks are required for various sensors to prepare for the
1980 PEPE program?
The Workshop addressed remote sensors other than airborne active remote
sensors (e.g., long-path absorption measurements using laser and nonlaser
sources) and passive remote sensors because it was agreed that these types
of instrumentation may provide significant information on gaseous species in
the troposphere, and may provide a wider network of sensors to complement
currently planned in situ sensors discussed at the earlier workshop, Instru-
mentation for Study of Transformation.
The Following agenda was adopted by this Workshop group.
1. An overview discussion was conducted of the role of active/passive
remote sensor technology in the PEPE program, with respect to mea-
surements of gases, aerosols, extinction, meteorological parameters,
visibility, turbulence, and backscattering.
2. A shopping list of instruments (active/passive) with the potential
for application in the 1980 PEPE program was developed based on
the following guidelines. The instruments had undergone develop-
ment, laboratory testing, and sufficient field testing with either
calibration testing or correlative measurement testing, which pro-
vided reasonable results consistent with expected physical pheno-
mena being measured. Instruments not falling within those guide-
lines (based upon information available to the attendees) were
excluded or identified on the chart.
568
-------
3. Earth Observation Satellite Systems were not discussed at the Work-
shop since it appeared that sufficient discussion was generated
during the 2-day seminar portion of the meeting to provide an ade-
quate data base for assessing the capability of satellite imagery.
However, it was agreed that satellite observations of PEPE's could
complement the ground/airborne measuring network in identifying
regional boundaries of PEPE and to provide information (in conjunc-
tion with meteorological data) on the early formation of regional
haze.
The available active remote sensors are summarized in Table IV-2 and
have been categorized according to fixed wavelength lidar systems, tunable
wavelength lidar systems, or acoustical sounding techniques. Table IV-3
summarizes passive remote sensing technology, and instruments have been
categorized according to nondispersive and dispersive correlation radiome-
ters, interferometers, infrared heterodyne spectrometers, and spectroradio-
meters. Table IV-4 summarizes long-path laser monitoring systems using
retroreflectors to provide total integrated burden measurements along a 1-km
horizontal path, or a 200-ft vertical path deploying retroreflectors on high
vertical structures. A 20-m long-path airborne system has also been included
to provide areal coverage of gaseous species such as CO, N20, and NH3.
ROLE OF REMOTE SENSORS IN PEPE PROGRAMS
Two types of remote sensors were considered by the Workshop: (1) active
remote sensors in which the sources of radiation are electromagnetic (UV/VIS/
IR) and acoustical soundings; and (2) passive remote sensors in which natural
sources of radiation are utilized including upwelling thermal radiation,
solar-reflected radiation, and direct solar/sky emissions. Platforms consi-
dered were high-flying (>20,000 ft) and low-flying (<20,000 ft) aircraft and
mobile and stationary ground-based measurements. An extensive topic of
discussion involved the role remote sensors can play in providing complemen-
tary and independent measurements of gases, aerosols, meteorological parame-
ters, visibility, turbulence, and optical backscatter. Four major roles were
identified for the PEPE program.
1. Airborne remote sensors could provide positioning data to other
instrumentation (i.e., in situ) for the purpose of obtaining
569
-------
detailed measurements in localized areas within the large haze
region. Remote sensors can provide positioning data to other
instruments, which can obtain localized measurements with poten-
tially higher sensitivity and specificity, smaller spatial resolu-
tion and basic characteristics of aerosols (i.e., concentration,
size distribution, chemical composition). Remote sensing from
high-altitude, fast aircraft could provide early information on
the formation of regional episodes and define the boundaries, areal
extent, tracking information, vertical layering of aerosols, and
S02, and 03 concentrations in the regional haze.
2. Airborne remote sensors could serve as the primary instrument
technology for assessing characteristics of isolated plumes includ-
ing structure of the plume, dispersion, degree of Gaussian fit,
gaseous S02 and 03 concentrations, and boundary layer structure.
These measurements would be conducted from low-altitude, low-speed
aircraft and from ground-based stationary and mobile platforms to
perform plume characterization studies. For both these applica-
tions, on-line data processing of remote sensor data would be
required to provide operational data for other instruments.
3. Airborne remote sensors could provide information on the correla-
tion between aerosol concentrations in the regional haze and the
concentration of gaseous constituents such as S02 and 63. At
present the correlation between aerosol loading and the concentra-
tion of S02 and 03 in regional haze episodes is not well understood.
Remote sensors could be utilized in a unique operational basis
(large areal extent) to provide data for improved understanding of
the correlation for a particular episode. Further, active remote
sensors from airborne platforms can provide measurements of the
location and width of the vertical mixing layer, and potentially
provide data to understand the composition of relative "clear"
areas within the regional haze.
4. By comparing and complementing remote sensing data with data
obtained from the complete measurement network in PEPE, retrospec-
tive data analysis can be performed to improve understanding of
570
-------
chemical transformations, transport, and chemical processes in the
regional haze.
The Workshop participants agreed that remote sensors represent cost-
effective sensors for providing an extensive data base for regional haze
studies by providing large areal coverage, vertical layering, and integrated
burden measurements with high potential in model validation studies particu-
larly in the area of transport. The following unique capabilities of remote
sensors for the PEPE program were identified.
Capability to perform daytime/nighttime plume characterization.
Characterization of vertical structure in plumes and extensive air
masses.
Rapid characterization of plume dispersion.
Identification of boundary layer behavior over various terrain
characteristics.
Capability to detect invisible plumes.
Capability to identify aerosol/gaseous distribution along regional
haze boundaries with on-line information.
Provide parameters for model validation and transport processes
above the surface.
Provide measurements of S02 and N02 transboundary fluxes.
Provide total burden measurements for boundaries and areal extent
of regional haze.
ATMOSPHERIC PARAMETERS MEASURABLE WITH REMOTE SENSORS
The spectrum of atmospheric parameters that can be measured with remote
sensors was discussed. Included were gases and aerosols, optical backscat-
tering, meteorological parameters, extinction, visibility, and turbulence.
Gases and Aerosols
The Workshop group identified the capability to measure simultaneously
aerosols and S02 and aerosols and 03 with the potential to measure the three
species simultaneously with further development of existing DIAL systems.
Total burden measurements of 83 from aircraft platforms represent valuable
input parameters to regional modeling. Other species that are measurable
with remote sensors include N02, CO, NH3, H20, N20, CH4, C02.
571
-------
Optical Backscattering
Optical backscattering capability can be obtained with a variety of
active lidar systems, with multiwavelength capability in some cases (see
Table IV-2). The High Spectral Resolution Lidar (HSRL) has the unique
capability of distinguishing optical backscattering from aerosols and
molecules through a unique combination of high resolution Fabry-Perot
Filters. This instrument can obtain the vertical and horizontal distri-
bution of the atmospheric optical extinction coefficient, the aerosol-
to-molecule scattering ratio, and the aerosol backscattering phase
function from aircraft platforms, and can complement the interpretation
of multiple wavelength lidar instruments for the characterization of
atmospheric contaminant mass loading, size distribution, and composition.
Meteorological Parameters
Remote sensors are capable of measuring meteorological parameters
including temperature, winds, water vapor, atmospheric turbulence,
height of mixing layer, and clouds. Remote sensors with potential for
meteorological applications are tabulated in Table IV-2. Consideration
of measurements of meteorological parameters should include both active
lidar techniques and acoustical radar measurements.
Visibility
A primary objective of the PEPE program is to improve understanding
of visibility in regional haze. Due to the complexity of the visibility
problem, the participants did not agree on the capability of current
active remote sensors to provide detailed understanding of the visibility
problem. However, it was agreed that the multiple wavelength capability
of remote sensors should provide a data matrix that, in conjunction with
other information, could improve current understanding of the definition
of visibility. The potential for application of active lidar techniques
to improve understanding of visibility should be further investigated
and is recommended as a near-term task.
572
-------
Summary of Remote Sensing Instrumentation Charts
Although detailed information of remote sensors is given in Tables
IV-2, IV-3, and IV-4, this section will provide an overview of the
charts. Fourteen fixed wavelength lidar systems were identified, with
12 related to optical backscattering measurements and 2 related to
measuring gaseous species—the 03 instrument is currently operational
and the NH3 instrument may be operational with minor modifications. Of
the 14 fixed laser systems, the 2 gaseous species measurements utilize
aircraft platforms; the 12 optical scattering instruments are divided
into 4 aircraft systems, 5 mobile ground platforms, and 3 stationary
platforms. Three tunable wavelength lidar systems were identified, one
measuring optical backscattering and two measuring aerosol and gaseous
species (03, S02, H20, N02). One aircraft instrument with tunable
wavelength capability will be operational from an aircraft system and
will be capable of simultaneously measuring aerosols and 03 or aerosols
and S02.
Of the 11 passive remote sensors identified, only 2 were operational
from an aircraft system, namely the aircraft MAPS instrument (CO) and
the multispectral observation of pollutants (MOPS) for measuring S02,
03, and N02. However, some passive remote sensors (e.g., COSPEC) might
have the capability of aircraft deployment with further development, and
should be identified as near-term tasks for the 1980 PEPE study. Three
of the passive remote sensors are deployable in ground-based mobile
vans, and six operate from stationary platforms. A recommendation of
the Working Group was that a survey be made to identify those passive
remote sensors that, with minor modification, might be used in the
airborne mode to provide additional capability to active remote sensor
technology for measurement of gaseous species.
Three long-path laser absorption techniques were identified by the
Workshop group; one was an operational aircraft system and the remaining
two were long-path ground systems. The latter two systems have been
extensively tested over horizontal path lengths of 1 km and are currently
planned for testing with a retroreflector located approximately 200 ft
up a high tower, thereby providing the capability to complement ground-
based in situ measurements with vertical burden measurements.
573
-------
TABLE IV 2a. ACTIVE REMOTE SENSORS (FIXED WAVELENGTH-OPTICAL SENSORS)
Category lad
instrument
Fixed Wavelength
High Spectral
Resolution Lidar
(HSRL)
Laser Absorption
Spectrometer
Mark IX
Mark V
Mark XI
tn Multi-Wavelength
--J Infrared Lidar
Mark VI II
Lidar
Lidar
Airborne Lidar
System III
Airborne Ozone
Sensor
Van-Mounted
Lidar System
Doppler Laser
Velocimeter
Airborne Laser
Doppler Velocimeter
Species
Aerosol extinction
03.NH3
Aerosols
Aerosols
Aerosols
Aerosols
Aerosols
Aerosols
Aerosols
Aerosols
Ozone
Tropospheric
Aerosols
Aerosols
Aerosols
Spectral range
tan
0.4-0.45
9-12
(discretely tunable)
0.700
1.060
0.530 and 1.06
1.06,1.6.3.5,
10.6
0.700
0.700
0.700
0.53 and 1.06
simultaneous
9.5
0.35, 0.693, 10.6
microwave
to
to
Transmitting
laser
N2-pumped dye
CO; Waveguide
laser
Ruby
Nd:YAG
Nd:YAG
Nd:YAG end
C02
Ruby
Ruby
Ruby
Nd:YAG
TEA
C02 laser
Ruby, C02
Cu vapor
Cu vapor
Receiver
telescope
diameter
(cm)
36
15
15
15
35
40
15
—
20
35
25
=80
=25
«25
Viewing
mode
Atmospheric
backsc after
Reflected ground
(molecular absorption)
Atmospheric
backscatter
Atmospheric
backsc alter
Atmospheric
backscatter
Atmospheric
backscatter
Atmospheric
backscatter
Atmospheric
backscatter
Atmospheric
backscatter
Atmospheric
backscatter
Oownlooking
earth reflected
Atomospheric
backscatter
Atmospheric
backscatter
Atmospheric
bukscatter
Platform
A/C
A/C
Mob He van
Fixed station
Aircraft
(Queenaire)
Transportable
system
Mobile van
Mobile van
Mobile van
Small twin-engine
A/C
Large twin-engine
A/C
Stationary
Stationary
Aircraft/
side viewing
Current
status
System assembly for
8/79 flight (CV-9901
Operational
Operational
Operational
Operational
1.06 and
10.6 iaa operational
Unknown
Unknown
Operational
Operational
Prototype constructed
schedules tests 5/79
Operational
(ruby, microwave)
Operational
Operational
Field
application
Regional aerosol
extinction profiles
Total burden
Plume studies
Aerosols
Aerosols
Aerosols
Aerosols
Plume opacity
measurements
Plume opacity
measurements
Point source/urban
plumes
Total burden
ozone
Met. parameters
Plume studies
Atmospheric turbulence
Wind speed
Plume turbulence
Plume turbulence
Organization/
contact
NASA/LaRC
(E. V. Browell) +
University of
Wisconsin IS. Shipley)
JPL
(E. T. Menzies)
SRI
(E. Uthe)
SRI
(E. Uthe)
SRI/EPRlC)
(E. Uthe)
SRI
(E. Uthe)
EPA«>
(Ching)
EPA<3>
(Conner)
EPA/NEIC<*>
(Dybdahl)
EPA/Las Vegas
(Joe Eckert)
EPA/Las Vegas
(Joe Eckert)
NOAA/Boulder
(V. Derr)
NOAA/Boulder
(V. Oerr)
NOAA/Boulder
(V. Derr)
Remarks
Ground-based vertical
measurements to 3 km-
complete
Field testing in
SEV-UP studies
Real time data proc-
essing and display
Real time data proc-
essing and display
Scheduled to
fry in 1979
Complete capability
requires further development
-_
Used for plume
opacity studies
Plume opacity
Resolution element
30'-vertKal
30'-horizontal
Sensitive to
ambient 03 levels
Molecular and spherical
particle discrimination
through polarization
Met. parameters
Plume turbulence
measurements
-------
TABLE IV-2b. ACTIVE REMOTE SENSORS (TUNABLE WAVELENGTH-OPTICAL SENSORS)
en
«>j
01
Category and
instrument
A/C Dial system
DIAL system
Airborne Lidar
System II
Receiver
telescope
Sptttnl range Transmitting diameter Vumini
Species pm User (cm) mode
Current Field Organization/
Platform status application contact
S02.03, N02, H20. 0.28-0.9 Nd:Y AC pumped 36 DIAL upward/ A/C System assembly for Regional study of 63, S02, NASA/LaRC
aerosols, temperature. dye downward viewing 8/79 flight aerosols, trap H20 profiles (E. V. Browell)
pressure
S02.03 0.28-0.35 Nd:YAG pumped SO DIAL Mobile Available for testing Gaseous atmospheric SRI/EPRl'51
dye 1/80 species (Hawley)
Aerosols 0.59 Flashlamp pumped 30 Atmospheric Small twin-engine Operational Dimensions/location of EPA/Las Vegas
dye laser backscatter aircraft point source plumes (J. Eckert)
TABLE IV-2c. ACTIVE REMOTE
Category and Measured
instrument variable
RASS Temperature,
winds aloft
Angle of Winds aloft
arrival
Sodor network Mixing depth
SENSORS (ACOUSTIC RADAR)
Organization/
contact Remarks
SRI(Franhal) Operational
SRI (Russell) Possibly available
summer 1980
SRI (Uthe-Russell) Backscatter
sounder
Possibly available
summer 1980
Remarks
Field demonstration in
NASA SEV-UPS Program
Real time display
Resolution element
150 ft. -vertical
1 SOft.-riorizontal
-------
TABLE IV-3. PASSIVE REMOTE SENSORS
Category ind
instruments
Non-Dispersive Gas Filter
Correlation Techniques
A/C Maps
Fourier Transform
Spectrometer
Interferometry
HSI
IR Interferometer
Dispersive Correlation
Technique
COSPEC
(passive)
cn
vj COSPEC
CT> (active)
Infrared Helerodyna
Spectrometers
IHR
LHS
Spectroradiometers
Sun Photometer
Automatic
Telephotometer
Multispectral
observation of
of pollutants (MOPS)
Species
CO. CH4
CO, CHj. C02. N;0,
H20. S02. NO. ...
Spectrally scanning
Spectrally scanning
SO 2, N02
S02.N02
03.NH3
Continuous
spectral coverage
Aerosols, water vapor
Visibility
S02, 03. N02
Spectral conraee
um
4.524.8
2.016.0
1.0-14.0
2.0-14.0
0.3-0.5
0.3-0.5
9-12
(disc rate steps)
7.513.0
UV, VIS, Near IR
Visible
(0.30-0.60)
Field of
•Jew (degrees)
7.5
—
2
12
N/A
N/A
0.25 mr
0.25 mr
—
—
Multfrpectral
imaging
Vertical
resolution (km)
<5km
Integrated
burden
Integrated
burden
Integrated
burden
Integrated
burden
Integrated
burden
<5
<5
Intvgrittd
burden
Integrated
burden
Total burden
(1 ppm-m)
Viewing
mode
Thermal IR
Direct solar
Direct solar
Aurora's and
night glow
Vertical upward
viewing •
Una of sight
to source
Direct solar
Direct solar
Direct solar
Direct solar
Direct solar &
nadir
Platform
A/C
Mobile van
Balloon/gnd-based
Mobile van
Mobile van
possible A/C
Stationary
Stationary gnd
Stationary gnd
Stationary
Stationary
A/C
Current statin
A/C operational
Operational
Operational
Operational
Operational & tested
in 22 countries
Operational & tested
Operational
Operational
Operational
Operational
Operational
Organization/
contact
NASA LaRC
(H. Reichle)
EPA-RTP
(W. Hergett)
JPL
IB. Fanner)
Utah State University
(A. Steed)
Barringer Res. Inc.
& Contractors
Barringer Res. Inc.
& Contractors
NASA LaRC
(J. M. Hoell)
NASA LaRC
(f. Allario)
SRKUthe)
SRI-EPRI
(Uthe)
JPL
Remarks
Measured dispersion of urban plumes
outward from Chicago
Horizontal range «1 km depending
upon source intensity
Stratospheric balloon flights successful
Nightime measurements possible
Plume dispersion studies (400 km range)
Plume model verification
S02 , N02 transboundary mass fluxes
1-6 km range
Currently performing NH3 measurements
Laboratory engineering model
Automatic tracking and
digital recording
TV microcomputer system for
evaluating contrast reductions
Plume studies of S02
-------
TABLE IV-4. LONG PATH LASER MONITORING SYSTEMS
Category and
instrunwot Species
Tunable Wavelength
Differential Absorption N2
------- |