TT TT       ^     ^T A TT
Hazardous Wastes
and Hazardous
     Site Remediation • On-Site Treatment • Risk Assessment
  Contaminated Groundwater Control • Permitting • Monitoring • Incineration
    • Underground Leak Detection • Fixation • Cost/Economics

            WASTES  AND
     Site Remediation • On-Site Treatment • Risk Assessment
  • Contaminated Groundwater Control • Permitting • Monitoring •
Incineration • Underground Leak Detection • Fixation • Cost/Economics
         March 4-6, 1986 • Atlanta, Georgia
              U.S. Environmental Protection Agency
           Hazardous Materials Control Research Institute
                 Department of Defense
           Agency for Toxic Substances and Disease Registry
                Portland Cement Association
            National Environmental Health Association
                 National Lime Association

Printed in the United States of America
Library of Congress Catalog No. 86-80269
Copyright © 1986
Hazardous Materials Control Research Institute
9300 Columbia Boulevard
Silver Spring, Maryland 20910

  The National Conference and Exhibition on Hazardous Wastes and Hazardous
Materials:  Site  Remediation  •  On-Site  Treatment   • Risk  Assessment  •
Contaminated Groundwater Control • Permitting • Monitoring • Incineration •
Underground Leak Detection • Fixation • Cost/Economics required dedication and
talent from many individuals and commitment from a number of organizations. We
gratefully express our thanks and appreciation to the following conference affiliates
for all their contributions to a successful conference.

U.S. Environmental Protection Agency
Hazardous Materials Control Research Institute
Department of Defense
Agency for Toxic Substances and  Disease Registry
Portland Cement Association
National Environmental Health Association
National Lime Association

  We also wish to express our gratitude to all of these knowledgeable individuals for
their advice and guidance in planning and producing a highly effective and infor-
mative program:

Gary F. Bennett, Ph.D., The University of Toledo
Hal Bernard, Hazardous Materials Control Research Institute
Ken Gutschick, National Lime Association
Robert Knox, U.S. Environmental Protection Agency
Charles Mashni, U.S. Environmental Protection Agency
Thomas Potter, National Lime Association
Jerry Steinberg, Ph.D., Water & Air Research,  Inc.
Andres Talts, Department of Defense
Ralph Touch, Agency for Toxic Substances and Disease Registry

  Producing a document of the magnitude of  this proceedings requires a highly
skilled team, much cooperation and communication, and a tremendous amount of
effort by all involved. We are fortunate to have such a team and would like to
convey our special thanks to Dr. Gary Bennett, Professor of Biochemical Engineer-
ing, The University of Toledo, and Hal Bernard, Executive Director,  HMCRI, for
the excellent  editing; to the  typesetters, proofreaders, and graphic artists who
completed a tremendous amount of work in an incredibly short period of time; and
to the staff of HMCRI for coordinating the myriad details and activities of this

       The RCRA Amendments  of  1984 (the new RCRA) have
       radically changed hazardous waste management through-
       out the country and considerably improved the control of
 hazardous wastes. Whereas it was estimated that the old RCRA
 provisions would cost the regulated industry between $1  billion
 and $3 billion per year, the new  RCRA amendments, when com-
 pletely implemented and  fully in effect, may cost as high  as $20
 billion per year.
   The new RCRA includes 58 congressionally mandated  statu-
 tory deadlines that go into effect by 1987 or early 1988. Congress
 wrote, into the statute, provisions that are detailed and directive.
 The bill  does not wait for EPA to write regulations or issue
 guidance; provisions go into effect automatically on prescribed

 Land Application  Management
   The law establishes a hierarchy of management practices. There
 is a very strong presumption against the disposal of hazardous
 wastes on the land and a very strong preference for treatment and
 destruction of hazardous wastes.

 Site Remediation
   The bill provides EPA with "corrective  action authority" at
 existing hazardous waste  facilities which are very similar  to the
 enforcement authorities under Superfund. The bill also closes the
 gap in the existing  program by  expanding the regulated com-

 Leaking Underground Tanks
   A major new subtitle of the  legislation  dealing with leaking
 underground storage tanks went into effect on May 8, 1985. This
 section probably affects two to  five million underground tanks
 across the U.S. including tanks storing petroleum products and
 gasoline as well as  very  complex tank  storage systems storing
 hazardous chemicals.
   As of May 8, 1985, there is a ban on new tanks that are not de-
 signed to prevent  releases due to corrosion  or structural failure.
 An estimated 100,000  new tanks must meet these requirements.
 By May  1986, a  nationwide registration program will require
 state notification of the age, size, type and location of the tank, as
 well as its uses.

Burning Hazardous Wastes
  The provision concerning burning hazardous wastes and the
blending of hazardous wastes into fuel requires  notification of
the fuel user that they are receiving hazardous waste. Some can-
not be burned in residential and commercial boilers.

RCRA Permit Program
  The  bill makes permitting of a hazardous waste facility far
more difficult than it has been. Permit applicants must now sub-
mit exposure information on the potential for public exposure to
hazardous substances from landfills and surface impoundments.
This exposure information will be used to write new permit con-
ditions. EPA has the authority to write any permit condition
necessary to protect human health in the environment, inde-
pendent of whether  or not  regulations  are in place  for that
  Congress has also set some very strict deadlines for the issuance
of permits.  By November  1985,  all applications for land disposal
facility permits must have been  submitted to EPA or an author-
ized state.  Final determination on these  land  disposal  permit
applications must be made by November 1988.
  By November 1986, all incinerator applications  must be sub-
mitted  and EPA must make final decision on these by November
1989.  All remaining applications have to be in by November
1988,  and EPA must make a decision on these by November

Small Quantity Generators
  The first small quantity generator (SQG) requirement went into
effect in August 1985. Any SQG that ships wastes offsite must do
so by using the Uniform  National Manifest Form. Effective in
April 1986, SQGs must send their wastes to fully regulated haz-
ardous waste facilities.  These requirements will have a  tremen-
dous impact on small generators; costs for their waste manage-
ment will be considerably  greater.

Other RCRA Amendments
  There are many other provisions of the new RCRA. There is a
set of provisions that deal with non-hazardous  solid waste and
small quantity generator wastes disposed of in landfills and sur-
face impoundments. The  bill also provides for  federal enforce-
ment of the requirements for lagoons. There are  a series of waste
minimization requirements for generators to certify  that they are
doing everything economically feasible to reduce the amount of
wastes  generated.  There are requirements for listing additional
hazardous wastes. Delisting of wastes is much more difficult and
complex under the new bill. New tests for ascertaining toxicity are
included in the new RCRA.

Equipment for Data Collection at Hazardous Waste
Sites—An Overview for Environmental Management
Professionals	1
  James P. Mack & Thomas J. Morahan
A Statistician's View of Groundwater Monitoring	8
  Douglas E.  Splitstone
Data Evaluation in a Groundwater Study of Waste
Management Practices in the Phosphate Processing
Industry	13
  Edward W. Mullin, Jr., Jack S. Greber & William
  E. Thompson
Plant Cuticular and Dendrochronological Features
as Indicators of Pollution	17
  O.K. Sharma, Ph.D.
Detailed Stratigraphic and Structural Control: The Keys
to Complete and Successful Geophysical Surveys of
Hazardous Waste Sites	19
  H. Dan Harmon, Jr., P.O.
Detection and Measurement of Groundwater
Contamination by Soil-Gas Analysis	22  /
  H.B. Kerfoot, J.A.  Kohout &  E.N. Amick
Environmental Appraisals and Audits: A Case Study
of Their Application	27
  Anthony R. Morrell

Chlorinated Organics and Hydrochloride Emissions
Sampling from a Municipal Solid Waste Incinerator	31
  Thomas A.  Driscoll, James P.  Barta, Henry J. Krauss,
  David H. Carmichael & J. Maxine Jenks
CARE—Modeling Hazardous Airborne Releases	34
  M. Gary Verholek
The Use of PC Spreadsheet-Based Graphics to Interpret
Contamination at CERCLA/RCRA Sites	39
  George A. Furst, Ph.D.

Emergency Response to Toxic Fumes and Contaminated
Groundwater in Karst Topography: A Case Study	44
  P. Clyde Johnston, Mark J. Rigatti & Fred B. Stroud
Use of Low Flow Interdiction Wells to Control
Hydrocarbon Plumes in Groundwater	49  >
  John H. Sammons, Ph.D. & John M. Armstrong, Ph.D.
Computer Groundwater Restoration Simulation at a
Contaminated Well Field	58
  Shih-Huang Chieh, Ph.D. & Jeffrey E. Brandow, P.E.
Enhancement of Site Assessments by Groundwater
Modelling  	64
  Joseph R. Kolmer, P.E.  & John B. Robertson, P.G.
Alternative Treatment Techniques for Removal of Trace
Concentrations of Volatile Organics in Groundwater	69
  Mark E. Wagner & Brian V. Moran
Cost-Effective Soil Sampling Strategies to Determine
Amount of Soils Requiring Remediation  	76
  Gregory J. Gensheimer, Ph.D., William A. Tucker,
  Ph.D. & Steven A. Denahan
Land Treatment of Wood Preserving Wastes	80
  John R. Ryan & John Smith
Method for Determining Acceptable Levels of
Residual Soil Contamination 	87
  William A. Tucker & Carolyn Poppell
Objective Quantification of Sampling Adequacy and
Soil Contaminant Levels Around Point Sources
Using Geostatistics	92
  Jeffrey C. Myers
                ON-SITE TREATMENT
Innovative Application of Chemical Engineering
Technologies for Hazardous Waste Treatment	98
  Robert D. Allan & Michael L. Foster
Field Studies of In Situ Extraction and Soil-Based
Microbial Treatment of an Industrial Sludge Lagoon	102
  David S. Kosson, Erik A. Dienemann & Robert
  C. Ahlert, Ph.D., P.E.
Cleanup of Contaminated Soils and Groundwater
Using Biological Techniques	110
  Paul E. Flathman & Jason A. Caplan, Ph.D.
Physical/Chemical Removal of Organic Micropollutants
from RO Concentrated Contaminated Groundwater	120
  L. Simovic, J.P. Jones & I.C. McClymont
State-of-the-Art Technologies of Removal, Isolation
and Alteration of Organic Contaminants Underground	124
  Walter W. Loo & George N. Butter

 Assessment of Volatile Organic Emissions from a
 Petroleum Refinery Land Treatment Site	127
   Robert G. Wet herald, Ph.D., Bart M. Eklund,
   Benjamin L. Blaney, Ph.D. & Susan A. Thorneloe
 Technology for Remediation of Groundwater
 Contamination	133
   David  V. Nakles, Ph.D. & James  E. Bratina

 Anaerobic Biological Treatment of Sanitary
 Landfill Leachate	136
   A.K. Mureebe, P.E., D.A.  Busch & P.T.  Chen, Ph.D.
 On-Site Versus Off-Site Treatment of Contaminated
 Groundwater—An Evaluation of Technical Feasibility
 and Costs	143
   Kent L. Bainbridge & Daniel  W. Rothman, P.E.
 Optimization of Free Liquid Removal Alternatives in
 the Closure of Hazardous Waste Surface Impoundments	148
   David K. Stevens
 Assessment of Chemical Treatment Technologies and
 Their Restrictive Waste Characteristics	154
   Hamid Rasiegar, Ph.D.. James Lit, Ph.D. &
   Chris Conroy
 Treatment Technologies for Hazardous Materials	157
   Jeffrey M. Thomas & Phillip T. Jarboe
 Update on the PACT Process	163
   Harry W. Heath, Jr.
 Treatment of PCB-Contaminated Soil in a
 Circulating Bed Combustor 	171
   D.D. Jensen, Ph.D. & D.T. Young
 Treatment of Hazardous Waste Leachate	175
   Judy L. McArdle, Michael M. Arozarena, William
   E. Gallagher, P.E. & Edward J. Opatken

 Field Testing of a New Hazardous Waste
 Stabilization Process	180
   Richard H. Reifsnyder
 Innovative Techniques for the Evaluation of
 Solidified Hazardous Waste Systems	186
   Harvill C. Eaton, Marty E. Tittlebaum & Frank K.
 Site Characteristics and the Structural Integrity
 of Dikes for Surface Impoundments	
   Jey K. Jeyapalan,  Ph.D., P.E. & Ernest R. Hanna
 Use of X-Ray Radiographic Methods in the
' Study of Clay Liners 	
   Philip  G. Malone, Ph.D., James H. May,  Kirk W.
   Brown, Ph.D. & James C. Thomas
 Closure Design and Construction of Hazardous
 Wastes Landfills Using Clay Sealants	
   John F. O'Brien, P.E.,  Lonnie E. Reese & Ian
   Kinnear, P.E.
Soil Liners for Hazardous Waste Disposal Facilities	206
  D.C. Anderson
Slurry Wall  Economical  in Dewatering of
Sydney Mine Disposal Site 	210
  Bruce J. Haas, Mark R. Nielsen & Norman N. Hatch
Utility of Soil Barrier Permeability Data	216
  Walter E. Grube, Jr.
                                                             Quality Assurance and Quality Control Procedures for
                                                             Installation of Flexible Membrane Liners	221
                                                               James R. Woods,  P.E. & Salvatore  V. Arlotta, Jr., P.E.
                                                             Mechanisms for the Fixation of Heavy Metals in
                                                             Solidified Wastes Using Soluble Silicates	224
                                                               Ella L. Davis, James S. Falcone, Scott D. Boyce &
                                                               Paul H. Krumrine

Illinois Plan for On-Site Incineration of
Hazardous Waste	
  James F.  Frank & Robert Kuykendall
An Introduction to EPA's New Trial Burn Data Book	233
  M.P. Esposito & N.J. Kulujian
Low Temperature Thermal Stripping of Volatile
Compounds	2-**
  John W.  Noland, P.E..  Nancy P. McDevitt & Donna
  L. Kolluniak
Comparisons Between Fluidized Bed and Rotary Kiln
Incinerators for Decontamination of PCB Soils/
Sediments at CERCLA Sites	242
  Henry Munoz, Frank L.  Cross, Jr., P.E. & Joseph
  L. Tessitore, P.E.

New Requirements for Underground Storage Tanks	246
  A nna O.  Buonocore, P. E.,  Gerald F  Kotas &
  Kevin G.  Garrahan, P.E.
Cost-Effectiveness Evaluation of Leak Detection and
Monitoring  Technologies for Leaking Underground
Storage Tanks	251
  James Lu, Ph.D.. P.E. <$ Wayne Barcikowski
Underground Storage System  Assessment, Testing
and Remediation 	269
  Scott J. Adamowski, Angela J. Caracciolo, III
  & G. David Knowles,  P.E.
Case Study of Product Detection in Groundwater	273
  Pratap N. Singh, Ph.D..  P.E.
Underground Storage Tanks—Leak Prevention,
Leak Detection, and Design 	278
  Jey A. Jeyapalan. Ph.D., P.E. & James B.
  Hutchison. P.E.

                 SITE MANAGEMENT
Performance Evaluation of Commercial Hazardous
Waste Treatment Facility Operations	292
  Ronald J. Turner & Joan V. Boegel
An Examination of Siting Problems for
Off-Site TSDFs	298
   Douglas  B. Taylor, P.E.
Spatial Data Research for Hazardous Waste Sites	301
   Timothy  W. Foresman & Lynn K. Fensiermaker
Sensitivity Analysis of Remedial Action Alternatives
for Hazardous Waste Sites	305
  Elio F. Arniella, P.E. & E. Lawrence Adams, Jr., P.E.

Design of a Lateral and Vertical Expansion at an
Existing Interim Status Landfill	310
  Rodney T. Bloese & Thomas G.  Ryan, P.E.




Cost Estimating for RCRA/CERCLA Remedial
  Michael R. Morrison & Gregory P. Peterson, P.E.
A Modular Computerized Cost Model for Remedial
Technologies at Superfund Sites	
  William Kemner, John Abraham, P.E., Jack
  Creber & Jay Palmisano

                 RISK ASSESSMENT
Using Cost/Risk Analysis in Waste Planning:
The New England Pilot Project	
  Debora C. Martin
Assessment of Potential Public Health Impacts
Associated with Predicted Emissions of
Polychlorinated Dibenzo-Dioxins and
Polychlorinated Dibenzo-Furans from a
Resource Recovery Facility	
  David Lipsky, Ph.D.
Improving the Risk Relevance of Systems for
Assessing the Relative Hazard of Contamianted Sites	
  Ellen D. Smith, Lawrence W. Barnthouse, Ph.D.,
  Glenn W. Suter II, Ph.D., James E.  Breck, Ph.D.,
  Troyce D. Jones & Dee Ann Sanders, Ph.D.
Soil Cleanup Criteria for Selected Petroleum Products
  Sofia K. Stokman  & Richard Dime, Ph.D.
Risk Management: Personnel, Equipment and Indemnity
  Denny M. Dobbs & Arlene B. Selber

Innovative Approach to Site Remediation Involving
an RCRA Part B. Permit	
  Michael J. Conzett, P.E. & Michael E. Harris, P.E.
Overview of the Proposed Natural Resource
Damage Assessment Regulations	
  Richard J. Aiken,  Willie R. Taylor, Ph.D.
Regulatory Impact Analysis for the Toxicity
  John L. Warren
Preliminary Assessments and Site Investigations
Under the Corrective Action Authorities of RCRA:
Analysis of Early Experiences	
  John W. Butler &  Robert D. Volkmar
Assessment of the Application of RCRA Part 264
Standards  to CERCLA Site Remediation	
  Rebecca N. Fricke, P.E.
Small Quantity Generators: The Maryland
Approach to Regulation and Assistance	375
  Alvin L. Bowles, P.E.
Siting Efforts in Southern California	378
  Kieran  D. Bergin
The California Site Mitigation Decision Tree	380
  Paul W. Hadley & William Quan
Hazardous Waste Management: The Role of the
Local Health Department	384
  Michael J. Pompili & Philip G. Brown

U.S. EPA's Initiatives for Expanded Public
Involvement in the RCRA Permitting Program	386
  Vanessa Musgrave & Edwin Berk
Cleanup in the Sunshine: Florida DOT's Public
Information Program at the Fairbanks Site	390
  Robert C. Classen & Charles C. Aller
Public Involvement in the RCRA Permitting Process—
A Facility Perspective 	394
  Gordon Kenna
Public Participation in Siting Hazardous Waste
Management Facilities in Alaska	397
  Sharon O. Hillman
Developing  a Comprehensive Public Affairs Policy
for New and Existing Industrial Operations	401
  Ann E. Burke
Behind the "F"  in Public Education for Hazardous
Waste Management: A Case for Special "Tutoring"
—Plus Some Tips for Better Grades 	404
  Howard A. Coffin, Patricia Hunt & L.T. Schaper, P.E.
                REUSE & RECOVERY
Recycling of Dust from Electric Arc Furnaces—
An Experimental Evaluation 	409
  E. Radha Krishnan, P.E., William F. Kemner &
  Copal Annamraju, P.E.
Trends in Used Oil Composition and Management	419
  Jacob E. Beachey & William L. Bider
Used Solvent Elimination Program	424
  Renato G. Decal, P.E.
Waste Reduction Audit Procedure—A Methodology
for Identification, Assessment and Screening of
Waste Minimization Options	
  Carl H. Fromm, P.E. & Michael S. Callahan, P.E.
Minnesota Technical Assistance Program: Waste
Reduction Assistance for Small Quantity Generators
  Cindy A. McComas & Donna Peterson
Waste Minimization at Air Force GOCO Facilities	
  Douglas L. Hazelwood, Brian J. Burgher, P.E. &
  Charles Alford
Defense Environmental  Leadership Project  Study
of Industrial Processes to Reduce Hazardous Waste	
  Thomas E. Higgins, Ph.D., P.E. & Drew P. Desher
                                                             Exhibitors' List	450




                             Equipment  for  Data  Collection  at
                      Hazardous Waste  Sites—An Overview for
                      Environmental Management Professionals
                                                  James P. Mack
                                               Thomas J.  Morahan
                                          Fred C.  Hart Associates, Inc.
                                              New York, New York
  Industry and government are in the process of cleaning up
America. Industry is becoming increasingly concerned with ques-
tions of environmental liability. Regulatory agencies are actively
protecting public  health and the environment. Future liability
can be minimized by the use  of environmental audits. Present
liabilities, however, must be thoroughly understood in order to
develop cost-effective liability  management plans that meet the
goals and objectives set for environmental cleanup.
  This understanding generally is gained through environmental
investigations at individual sites  or facilities. CERCLA,  RCRA
and other statutes require regulatory compliance. In some cases,
the buyer or  seller of a property may run into unforeseen en-
vironmental liabilities. Some states have recently passed legisla-
tion  that requires  environmental conditions be defined and
problems investigated prior to real estate transactions. In cases of
corporate mergers or other processes where ownership of a facil-
ity changes, new owners or sellers may be held accountable for
the environmental liability at  a facility. While the  government
agencies  ensure that cleanups  are effective,  industries which  fi-
nance the cleanups can encounter huge expenditures. Business-
men and legal professionals handling environmental affairs stay
close to technical events, carefully monitoring costs and cleanup
strategies. These environmental  management professionals are
sometimes unfamiliar with techniques for data collection and the
creative use of equipment used for data collection activities. This
paper is directed toward business, legal and environmental man-
agement  professionals seeking knowledge of various types  of
equipment available for environmental data collection. An under-
standing  of the use of this equipment will help in the develop-
ment of cleanup strategies for a proper, complete and cost-effec-
tive cleanup.
  While  each site investigation is unique and requires a certain
degree of site-specific modification, there is a logical sequence of
events that is followed  in order to collect the desired informa-
tion. While the purpose of this  paper is to discuss the field equip-
ment available to perform the investigation, it is important  to
understand the sequence of events in order to define the situa-
tions in which the equipment should be used.
  The  first phase of the  investigation should be  devoted  to
collecting enough background data on site conditions as possible.
While there may not be much  data available  on actual site con-
ditions, there is usually a certain amount of regional information
available. Data can be obtained from local health departments,
state and federal environmental and geological agencies and uni-
versities.  This information  is important in defining the general
environmental setting of the site.
  The second phase of the investigation should be devoted to
understanding the source of the potential contamination. At this
point, certain indirect techniques such as surface  geophysical
instruments or portable analytical instruments  are very useful.
They can be used to define the possible horizontal extent of the
source, detect buried objects and initially determine the general
types of contaminants. Soil borings and soil sampling techniques
also are used to collect actual waste or contaminated  soils for
chemical analysis to further define the characteristics of the con-
  Once it has been determined that there is a contaminant source,
the question arises as to whether the contaminants are migrat-
ing. This usually requires more extensive sampling of surface
water, sediment, soils and groundwater. Normally, groundwater
is an important issue and becomes the focus of further investiga-
tions. Wells usually are installed adjacent to the source to deter-
mine  the subsurface  conditions as well as to detect any immed-
iate migration. Proper soil sampling  and well drilling techniques
are essential to the adequate completion of this task. Downhole
investigative  equipment becomes useful in defining subsurface
geology. Water levels are measured in wells to define flow direc-
tions. As the scope of the potential problem develops, additional
wells may be required at further distances to define long-term mi-
gration pathways.
  The final component is the data evaluation and risk analysis.
This phase is as important as the field effort. The data must be
properly organized and the actual risks correctly established in
order to arrive at a cost-effective remedial solution.

  Surface Geophysical Techniques have been used for  years to
give indications  of subsurface rock  and soil conditions. These
techniques vary greatly in theory and technology, resulting in very
specific conditions where their use is warranted. In recent years,
these techniques have been applied successfully to investigations
at hazardous waste sites.
  Surface geophysical  surveys are performed as indirect tech-
niques to provide information to investigators in advance of
direct sursurface investigatory techniques such as test pitting, test
borings or well drilling activities. At hazardous waste sites, these
techniques generally are used in three different ways. First, some
methods can be  employed to provide information about subsur-
face conditions  such as depth to bedrock, depth to the water
table or thickness of certain soil units. Next, some methods can
be  employed  to examine the migration  of contamination in
groundwater. Finally, some techniques can be utilized to search
for buried metal, specifically containerized wastes such as drums
or tanks which could act as ongoing sources of contamination.
                                                                                                     MONITORING    1

   Magnetrometry has been used successfully in the past to iden-
tify trends in regional geology. More recently, it has been ap-
plied to search for buried metals at hazardous waste sites. The
proton precession magnetometer measures the total intensity of
the earth's magnetic field at one point. The earth's magnetic field,
which is fairly constant over a small site area, is affected by fer-
rous metals.  Ferrous metal objects create their  own magnetic
fields which change the intensity of the  magnetic field over a
site. Using the magnetometer, these ferrous metal objects can be
   Several key steps in a magnetometry survey must be observed
to obtain quality data. First, all measurements must be referenced
to a grid system. Second, a point free of magnetic inferences
must be located, known as a reference point. The magnetic  field
of the  earth  changes  throughout the day. These changes arc
known as diurnal changes. A full record of diurnal changes must
be kept for corrections of the  intensity measurements taken at
grid points during the day. Often it is necessary to have a  base
station magnetometer dedicated to this task. These measurements
also can guard against the collection of poor quality data such as
that collected during magnetic storms.
   Once correct data are plotted, it is sometimes possible to model
the depth, mass, shape and angle of orientation of an object.
Theoretically, that  degree of data interpretation is possible  only
under ideal conditions.
   Variations on  this technique are possible. Gradiometers can be
used to measure the point changes in a field, instruments can be
hand held, sensors  can be placed on staffs and  instruments may
be towed along traverse lines for continuous readings.
   Limitations are critical to this type of survey. Magnetic induc-
tion caused by alternating current in power lines will affect read-
ings. In this case, or in  the case  of extensive surface metal inter-
ference, surveys should not be performed. Additionally, it is pref-
erable to have a map of shallow buried surface metal from a metal
detection survey to  aid in data interpretation. With these precau-
tions in mind, magnetometry can be a useful tool  to indicate the
presence of buried ferromagnetic objects.

Metal Detention
   Metal detention  has been used for years to locate pipes and
other buried metal  objects.  It relies on the effects of a conduc-
tive metal object or a radio signal which it broadcasts and re-
ceives.  The effective depth of penetration is only a few feet, but
it can be used very  accurately to locate shallow buried metal tar-

Electromagnetic Conductivity
   Electromagnetic conductance, generally  referred to as "EM,"
also responds to  the conductive properties of materials.  EM, also
called terrain conductivity, usually is used to indicate the presence
of a conductive contaminant plume.
   EM does not require ground contact. The EM transmitter coil
radiates an electromagnetic field which induces eddy currents in
the earth. Each of these eddy current loops, in turn, generates a
secondary electromagnetic field  which is intercepted by the re-
ceiver coil. An output voltage fis produced which is linearly re-
lated to subsurface conductivity.
   Continuous survey  instruments are available with  transmitter
and receiver fixed in the same instrument, but their use is limited
to shallow depths. Instruments are available with separate trans-
mitter and receiver coils for deeper surveys. Initially, several EM
soundings are made by varying the intercoil spacing at each loca-
tion.  After  a determination of  background  conductivity  and
appropriate  intercoil  spacing, EM profiles can  be completed
along previously surveyed transverse lines.
Electrical Resistivity
  Electrical Resistivity (ER) allows a measurement of  subsur-
face resistivities in soil, rock and groundwater. Application of the
method requires that an electrical current be  introduced into the
ground through a pair of surface electrodes. The resulting poten-
tial field  is measured at the surface between a  second pair of elec-
trodes. The electrodes can be arranged into several different con-
  Initially, several ER soundings can be made to establish sub-
surface background conditions and determine  the appropriate
electrode ("A") spacing  for  the ER profile lines. The sounding
data will be collected  by establishing a central reference  point
and varying the spacing between the current  and potential elec-
  ER profiles  then can be assembled to indicate areas of low re-
sistance,  indicating the presence of potential contaminant plumes.
Areas of low resistivity also could indicate perched water zones
or low lying topographic areas.
  ER is best used, as with most geophysical  techniques,  in con-
junction  with other indirect testing. Soundings always should be
done to determine the depth to the water table and depth of any
clay layers  that might be present. ER is a quick  and inexpensive
technique generally available for site investigations.

Ground Penetrating Radar
  Ground penetrating  radar  (GPR) is a  reflection technique in
which high frequency radio waves  are radiated downward into
the subsurface and then  reflected back to a receiving antenna.
Variations in the return signals occur when subsurface materials
have different electrical properties. An interface between two lay-
ers having sufficiently different electrical properties will show up
in the radar profile. The effectiveness of GPR can be limited by
the penetration depth  of the radio waves. The effective pene-
tration is highly  site  specific and  dependent upon subsurface
boundaries, water content and clay content.
  The radar system electronics usually are mounted in a van, with
the antenna towed behind. An impulse radar transmits electro-
magnetic pulses of short duration into the ground from a broad-
band antenna. Pulses from the antenna are reflected from vari-
ous interfaces in the subsurface and then are picked up by the re-
ceiver section of the antenna.
  Unfortunately,  GPR systems are very expensive and  usually
are  used to locate contaminant plumes which are lighter than

  Hazardous materials emergency response has spawned the de-
sign of many field analytical instruments due  to the added safety
requirements  inherent  in  the performance of such dangerous
work. Many of these instruments have been readily adapted for
use in site investigations to collect data which can be used in the
assessment  of a  site.  These  instruments can be used either to
gather data directly or as a screening technique to indicate where
further investigation is needed. This section of the paper dis-
cusses the more common types of instruments which can be used
to gather field data and their uses and limitations.

Organic Vapor Analyzers
  Two general types of instruments are currently used to analyze
for  the presence of volatile organic compounds.  Both detectors
rely on the ionization potential of the organic  compounds. Flame
ionization detectors (FID) rely  on a flame  which burns  com-
pounds to break chemical bonds to release  energy. Photoioniza-
tion detectors (PIDs) rely on ultraviolet light which is capable of
breaking the chemical bonds to release energy. The major differ-
ence in their application is that methane and  some very light  or-

ganic compounds cannot be ionized by ultraviolet light. As a re-
sult, PIDs can be used only if methane concentration is not a con-
cern. If qualification of methane is a  concern, a flame ioniza-
tion detector should be used to measure its presence.
  PIDs, such as the HNU portable photoionizer, and  FIDs, such
as  the Century Systems (Foxboro) Organic Vapor Analyzer
(OVA), have been used  in hazardous waste site investigations to
survey for the presence of organic compounds. Since  compound
separation and identification are not possible, total  concentra-
tions are measured as the benzene response equivalent. When
used in the  survey mode, the OVA is an efficient and accurate
indicator of total organic compound concentrations on a con-
tinuous sampling  basis  with a  response time of one  to  two
seconds. If instantaneous information is required on the presence
of methane  and non-methane hydrocarbons,  both instruments
can be used  in tandem. An example  of this application is a meth-
ane-vinylchloride  gas mixture which can  be common at landfill
  In cases where compounds identification or quantification is
required, portable gas chromatographs can be used.  The Model
128  OVA is  a portable FID gas chromatograph. The OVA has an
injection port which enables a sample to be injected from the air
intake line or by a syringe. As an initial approach to sample eval-
uation, this  technique can quantify the amount of volative organ-
ics  in a particular sample. This technique is useful when screen-
ing soil or water samples to select screen placement or selecting
worst  case samples for laboratory  analyses. This type of field
screening can lead to significant savings. If a gas chromatograph
(GC) column is present,  separation of each of the individual com-
ponents in a mixture is possible for identification and quantif-
ication. The OVA provides data in the 1 ppm concentration
  Low ppb concentrations  can be  measured with a Photovac
portable  PID gas chromatograph. Several Photovac models cur-
rently are available. The most advanced is the model 10S70 which
has  two columns, is fully programmable, has a 100-compound
memory with computer peak search  and integration and an inter-
face to allow communications with  a home-base computer. This
type of  instrument can  separate  non-methane  organic  com-
pounds, identify them and perform  integration to provide accu-
rate concentrations. When used with prepared standards and lab-
oratory analytical checks, the Photovac can provide confident
analytical analysis on-site in a matter of minutes without the ex-
pense of an on-site laboratory.

Applications of Portable Analytical
  Portable analytical instruments typically are employed in sev-
eral different approaches during site investigations. One approach
is an investigation  of the lateral of extent of shallow  volatile
organic soil  contamination. This is commonly done by opening a
hole in the ground with a slam-bar or other tool and inserting the
probe  of a  survey instrument to measure the concentration of
total organics. If measurements are referenced to a grid, they can
provide an indication of areas to be sampled for analytical analy-
ses or remediation.
  Another common use is to screen samples taken during an in-
vestigation.  To prevent costly  analyses  of every sample, only
certain samples are sent  to the  laboratory based on  the sample
screening. A small amount of sample is placed in a VOA vial and
heated to volatilize the organics into the air, or "headspace," in
the  vial. An aliquot of air is removed by a syringe and injected
into a GC column for analysis. This technique can also be used to
assist site geologists in deciding the depths at which to  finish wells
designed to monitor for volatile organic contamination.
  The photovac is gaining widespread use as a quantitative ana-
lytical tool to monitor drilling discharge water to prevent releases
of contaminants into the environment.
  Portable units to scan for concentrations of trace metals are ex-
pensive and relatively insensitive. They can be useful for sites with
widespread metals contamination in the 100 ppm and up range.
The  Colombia X-Met 700 portable x-ray  fluorescence (XRF)
spectrophotometer recently has been used successfully for this
application. Although it is expensive, the portable XRF currently
is gaining acceptance as a field analytical tool.
  Other direct-reading instruments which could be adapted from
safety-oriented tasks to portable analytical data gathering tools
include scintillometers, mercury vapor  analyzers and Hydrogen
Sulfide indicators.
  In addition to safety related instrumentation, some field ana-
lytical tools have been adapted from the laboratory and have been
around for some  years. They include pH meters, conductivity
meters, specific ion probes  for the measurement  of  dissolved
metals and colorimetric test kits for indicator parameters such
as chlorides and COD.
  The advantages of  using these techniques  are time and cost
savings. Field analyses provide data rapidly and allow data to be
factored into an  investigation while work is still taking  place.
Plans can be developed concurrently  for further investigation.
The number of field analyses need not be limited in these situa-
tions. A limited number of samples can be selected and analyzed
by the laboratory  for confirmation  of contamination, saving ex-
pensive laboratory time and reducing lab costs. Scheduling and
costs  are crucial on any project but are  paramount at sites  re-
quiring hydrogeologic investigations or soil investigation and ex-
cavation projects.

  A large  portion of the effort devoted to any site investigation
should be toward understanding the geologic and hydrologic
properties of the soils or rock, the  extent of soil contamination
and the potential  for contaminant migration through the unsat-
urated zones. Since most hazardous waste site investigations  in-
volve surface or near surface disposal of liquid and solid ma-
terials, one of the  primary questions that needs to be answered is
how much of the material remains in the soil and how much con-
tinues to leach. Source investigations involve the combined appli-
cation of surface geophysics and direct soil sampling.
  Additionally, contaminant migration potentials within the un-
saturated zone and upper water table are important issues to be
addressed. This determination requires the collection and analysis
of subsurface  soil and rock samples. Many methods are avail-
able  for advancing boreholes to obtain samples or details  of
strata; the selection of a method is  dependent on the  extent of
potential contamination, the complexity of the site geologic con-
ditions and  the risk of contaminant  migration. The  principal
methods in use are: hand augers, power augers and rotary core
drilling methods.

  The auger technique is useful as a tool to collect preliminary
samples, to use with an OVA, HNU or  Photovac in a soil gas
survey or to collect near surface waste samples. At least six types
of light portable augers are  available  for sampling  soft to stiff
  Hand augers may be used by one or two people. The hole is
advanced by pressing down on the cross bar as the tool is rotated.
Once the auger is full or has  collected sufficient material, it is
brought back to the surface and the soil is removed. The most
commonly used augers are the post hole, the helical and the spiral

  Hand augering usage is commonly limited to 3 to 7 ft. In tills
or clays that contain gravels or cobbles, hand augering may be
impossible. In uncemented  sands or gravels, it will not be pos-
sible to advance the hole below the water table since the hole will
continually collapse. Only samples of very limited size can be ob-
tained from the hole.
  In addition to collecting shallow subsoil samples, hand augers
also are used to create small diameter holes for a soil gas survey.
The holes are installed based on a grid system and the probe of an
OVA, HNU or Photovac is inserted into the hole. A sample of
the gas is collected and analyzed. This technique is a useful initial
method to define  the extent  of soil contamination by volatile
organic compounds.
  Because  hand augers are limited by soil conditions, depth of
penetration, quantity of sample, size of borehole and thickness of
the unsaturated  zone, they usually are confined to preliminary
surveys intended to generally define the limits of the contamina-

Power Augers
  The most common power  augering procedures used at  haz-
ardous waste site investigations are the continuous flight  solid
stem and hollow stem augers. Another type of augering  tech-
nique  is the bucket  auger. This procedure  uses open-topped
cylinders with base plates which have one or two slots reinforced
with cutting teeth which break up the soil and allow it to enter
the bucket as it is rotated. However, because of the limited depth
of penetration and the expense involved with operating a bucket
auger rig, it is rarely used.
  Solid stem augers drill much deeper holes with fewer problems.
However, this technique presents a serious problem when sub-
surface soil samples are required. To obtain the sample, the auger
string  must be pulled from the borehole.  Unless the soils are
stiff clays  or silts above the water table, the borehole is likely
to collapse and any sample  obtained would not be representative
of the stratum.
  Hollow stem augers consist of 3- to 6-in. diameter pipe with
continuous flights attached to the outside. Auger tailings are re-
moved from the hole by traveling up the flights and emerging at
the surface. When samples of the soil in advance of the augers
are required, the drilling is  stopped  and  a sampling device  is in-
serted into the pipe. The device is gently lowered to the bottom
of the hole and then either driven or pressed into the soil ahead of
the augers. Hollow stem augers allow drilling and sampling below
the water table, the acquisition of samples without pulling the
drill stem and the collection  of "undisturbed" samples.
  Auger rigs are mounted on four wheel drive trucks and all ter-
rain vehicles, allowing access  to most locations. Auger rigs will
not drill through rock and certain  cobbly  or till soils. Augers
usually are restricted to 200 or 250 ft in depth and still require cas-
ing if the hole is to remain after auger removal.

Rotary Core Drilling
  The  most common  use of rotary drilling  in site investigations
is to obtain intact samples of rock.  To do this, a "core-barrel"
fitted with a "core-bit" at its lower end is rotated and grinds
away an  annulus of rock. The stick of rock in the center of the
annulus passes up into  the core-barrel and is subsequently re-
moved from the  borehole when the core-barrel is full. The length
of core drilled before it becomes necessary to remove and empty
the core-barrel is termed a "run."
  In its simplest form, the core-barrel consists of a single tube
with an abrasive lower edge  which is rotated against the  rock
while fluid is passed under pressure.  However, most rock drilling
in  the United  States utilized the double  tube  swivel  type  core-
barrel. The importance of this tool is  that it contains an inner
barrel connected to the outer tube at the top via a swivel which
allows the inner barrel to remain stationary while the outer barrel
is rotated. The rock core "rides up" into the inner barrel.
  Double tube rock core barrels normally obtain a core between
5 and 10 ft long. Rock cores should be at  least 2 1/8 in. diameter
or larger. In soft to moderately hard rock, cores 12 in. in diameter
or larger can  be taken. A core barrel that collects a  2 1/8-in.
diameter rock core will produce a 3-in. borehole.
  Deep rock coring can be expedited with the use of the "wire-
line" technique. A wireline can produce  rock cores without the
removal of drill rods on the  outside core barrel. The inside core
barrel is retrievable as a separate  unit with the  use of a method
cable, or "wireline." An  empty  barrel can be sent down on  a
messenger to continue coring.  This type of technique saves time
and money for deep continuous coring operations.
  Core barrels should be held horizontally while the cores are ex-
tracted onto a  rigid surface.  The core should be properly tagged
and then placed in a core-box. Wooden spacer blocks should in-
dicate the top and bottom of each run.

Soil Sampling Devices
  The primary purpose of soil augering is  to collect subsurface
samples for visual examination and chemical and  physical tests.
In unconsolidated material capable of being drilled with an auger,
an important  component  of  any investigation should be  the
collection of soil samples ahead  of the  drill bit.  Hollow stem
augers are the most  common  drilling  technique,  but soil sam-
ples also can be collected ahead of drilling while using hydraulic
rotary methods.
  A  large number of samplers are available, most adapted from
geotechnical or soil  survey  investigations.  However, the  most
common device used in conjunction with hollow stem augering is
the split spoon sampler. In this device,  the sampler barrel is split
longitudinally  into two halves. The device  is  lowered into the
inner space of the hollow stem auger and driven into the soil by
repeated blows of a 140 Ib hammer falling through 30 in. Dur-
ing driving, the longitudinal halves are held together by the shoe
and  head which are screwed into each other.  The split barrel
allows easy examination and extraction of soil samples.
  Soil obtained with the split barrel sampler can be subject to  a
wide variety of field and laboratory tests. Samples can be  field
screened with an OVA  to determine presence of volatile organic
compounds. Soil should  be  logged  by  describing its texture,
color, grain size distribution, moisture  content and odor. If the
split  spoon has been properly decontaminated before sampling,
soil can be obtained for chemical analysis.
  Occasionally, large diameter "undisturbed" soil samples are re-
quired for complex physical  tests  such as hydraulic conductivity
determinations. Shelby or  push tube samplers often are used to
obtain undisturbed samples  of medium to  stiff consistency co-
hesive soils. The sampler is pressed into the soil. Care should be
taken because the sampler can be damaged,  either by buckling or
blunting or tearing the  cutting edge, when driven into very stiff,
hard or stony  soils. Undisturbed  samples of soft or loose  soils
must be taken  with a piston  sampler. Undisturbed samples  con-
taining much gravel or soft rock must  be obtained with a Deni-
son sampler.

  One of the  primary  waste migration pathways is via surface
runoff to rivers, streams and lakes. The contaminant may move
as a dissolved constituent in  the runoff or attached to entrained
soil particles. One of the most cost-effective means of determin-
ing the extent of waste migration  is to sample surface water and

Surface Water Sampling Techniques
  Surface water sampling locations are selected on the basis of
their probability for showing contaminants migrating from the
site. Prior to sampling, the surface water drainage in and around
the site must  be characterized using all available background
maps, topographic maps, serial photographs, river basin surveys
and other sources.
  Either grab or composite samples may be collected.  Grab sam-
ples are collected at one particular point and at one time. Flow
or time-weighted composited  samples are composed  of more
than one specific aliquot collected at various sampling sites and/
or at different points in  time. Because of the unknown safety
risks, as  well as the changes in chemical nature of the sample
that  may occur through  compositing, samples  containing haz-
ardous materials at significant concentrations shall not be com-
  If it is necessary to enter the water to  obtain the sample, it
should be done carefully  to leave the bottom sediments undis-
turbed. If the water is moving, a grab sample should be obtained
by pointing the open end of the container into the direction of the
flow. An attempt should  be made  to obtain the sample in the
middle of the stream and at mid-depth.
  The choice of a particular sampling device is dependent upon
the size of the water body,  the purpose of the sample and the
types and concentrations of wastes anticipated to be in the water
body. The types of samplers available include open tube, pond
sampler, manual hand pump, weighted bottle sampler, kemmerer
sampler and extended bottle  sampler. Of these, the pond sampler
and weighted bottle are used most often.
  The pond or dip sampler consists of a container attached to the
end of a long pole by an adjustable clamp. The pole can be of any
non-reactive materials such as  wood, plastic or metal. The sam-
ple is collected in a jar or beaker made of stainless  steel, glass
or non-reactive plastic. Preferably, a disposable beaker which can
be replaced at each station should be used.
  The weighted bottle sampler consists of a glass bottle, a weight
sinker, a bottle stopper and a line that is used to open the bottle
and to lower and raise the sampler during sampling. The sam-
pler can be either fabricated or purchased. This sampler is used to
take discrete samples  at predetermined intervals. The  pond sam-
pler can be used to develop composite samples from a single loca-
tion by collecting individual samples at regular intervals.

Sediment Sampling Techniques
  Sediment  samples are valuable  for locating pollutants  of low
water solubility and high soil binding affinity. Heavy  metals and
high molecular-weight halogenated hydrocarbons are examples
of contaminant groups which might be  found in greater con-
centrations in  sediment. A background sediment sample  should
be obtained  from sediments  upstream from the suspected source
for comparison. This is  especially  important if contamination
with heavy metals is suspected, because they occur naturally.
  Very simple techniques usually  are employed  for  sediment
sampling. Most samples will be grab samples from one particu-
lar locations although, for preliminary studies, several locations
may  be composited to reduce  the analytical requirements. Sug-
gested techniques include:
• In small, low-flowing streams or  near the shore of a pond or
  lake, the sample  container (typically an 8-ounce wide-mouth
  glass jar) may be used to scoop up the sediments.
• TO obtain sediments from larger  streams or farther from the
  shore of a pond or lake,  a Teflon beaker attached to a tele-
  scoping aluminum pole  by means of a clamp may  be used to
  dredge sediments.
• To obtain sediments from rivers or in deeper lakes and ponds,
  a spring-loaded sediment dredge or benthic sampler may be
  used by lowering the sampler to the appropriate depth with a
  rope. The  sediments thus obtained  are then placed into  the
  sample container.

  Geophysical borehole logging has been used for oil, coal and
mineral exploration for years. Recently,  these techniques have
gained widespread acceptance in the performance of hydrogeo-
logic investigations. There are many different techniques avail-
able. Lithologic logging can be used  to measure properties of the
rock or soil in a borehole or well. Hydrologic logging can be used
to measure the properties of the fluids in  a borehole.  Downhole
television also can be used  to examine downhole  conditions. A
review of these  techniques and their  applications is presented

Lithologic Logging
  Lithologic logging is used to measure the properties of down-
hole soils and rock. As with all logging techniques, it  is first ad-
visable to log at least one hole by physical examination of either
split spoon samples from soils or rock core samples for bedrock.
Rock core samples are more useful  than  chips because fracture
occurrence and widths (which are important conduits of ground-
water flow in bedrock) can be determined. Once lithology of at
least one hole is  known, the information can be correlated to
other holes.
  The simplest of the lithologic  techniques is caliper logging.
The caliper tool measures the diameter of a rock borehole using
three spring loaded prongs.  As the tool is  pulled up the hole, the
prongs spread out indicating the width of the borehole. Results
ared recorded electrically at the surface. The caliper method in-
dicates the relative hardness of  rock units. Caliper logging also
can indicate fracture location and occurrence. Caliper logging can
be the best tool to examine the  presence  of bedding plane frac-

Spontaneous Potential Logging
  Spontaneous potential logging is one form of electric logging.
It serves to measure the natural  electrical  potential that develops
between the  formation and the borehole fluids. This logging
technique must be performed  in  an open borehole  filled with
fluid. The  logging device consists of a surface electrode  and a
borehole electrode with a voltmeter to measure potential.
  Generally,  SP logs are read in terms of positive and negative
deflections from  an arbitrary  base line  which might correlate
either with permeable or impermeable zones. Information  re-
garding zones of higher permeability will  indicate likely contam-
inant pathways that may require further  investigation. The rate
at which the particular logging tool  is  lowered into the borehole
is another important criterion in  the data evaluation. The rate
should be sufficient  to detail specifics of the geology and pos-
sibly detect isolated  zones  of contamination or  likely contam-
inant pathways.

Resistivity Logging
  Resistivity  logging is another electric logging technique. Two
types of resistivity logging  techniques  are commonly employed
in downhole methods.  The  single point resistance  log is the sim-
plest form of electric log. In this  method, a single electrode is low-
ered into the hole and the return path for the current flow is furn-
ished by the  ground electrode. The single point resistance  log
measures the total resistance of earth materials.

  The second type of resistivity logging involves measurements
taken in the borehole in a similar manner as surface resistivities
investigations. In this system, four electrodes are commonly em-
ployed, two for emitting current (I) and two for potential meas-
urement (P). A number of different electrode configurations can
be used in resistivity logging to provide specific information. The
short normal spacing indicates the resistivity of the zone close to
the borehole where the drilling fluid might be an influence. The
long normal spacing  has more distance between the electrodes
and thus measures the resistivity further away from the borehole,
presumably beyond the influence of the drilling fluid. Both con-
figurations, short  and long-normal, measure a greater radius  of
influence than the single point resistance.
  A third configuration involving lateral devices utilizes widely
spaced electrodes for measuring zones  that are far from the bore-
hole. Because of the wide  spacings, lateral devices will not de-
tect thin beds of different  resistivity.  Boundaries of formations
having different resistivities are located  most  readily with short
electrode spacing, whereas  information on fluids (e.g., contam-
inated groundwater)  in permeable formations can  be  obtained
best with long spacings.

Density Logging Techniques
  Many downhole techniques are available to  measure the den-
sity of rock units. Natural gamma ray logging is perhaps the most
common. Natural gamma  logging measures  the  quantity  of
gamma rays naturally emitted by radioisotopes contained in clay
minerals. The higher the clay content of a rock, the more gamma
rays will be  emitted. This information is useful in indicating the
permeability of rock  units, since lower  permeability units gen-
erally contain more clay size particles.
  Other techniques which can be used to indicate density are high
resistivity density,  gamma-gamma density  and  neutron logging.
As with natural gamma ray logging, these techniques can be run
in cased boreholes. In other words, if the geology at a well needs
to be investigated and no records exist,  a cased hole can be logged
to gather qualitative  information. Natural gamma ray  logging
can be especially useful to  locate the  depth and thicknesses  of
bentonite seals on grout packs in monitor wells.

Vertical Seismic Profiling
  Recently,  some success has been met utilizing vertical seismic
profiling, or "cross-hole seismic." This  technique utilizes  an
array of detectors set in a  borehole to produce a three-dimen-
sional image of subsurface structures such as fractures. A seismic
source is used on the surface at some distance from the borehole
to create the seismic movement measured by  the detectors. Al-
though this technique is still under investigation, it shows promise
for  hydrogeologic investigation in the area of the development  of
three-dimensional data.

Hydrologic Logging Techniques
  The two most common hydrologic borehole logging techniques
are fluid conductivity and temperature logging.  These open bore-
hole techniques examine the properties of fluids within a bore-
hole and can provide  important information on  the natural cir-
culation of fluids in a borehole or the presence of contamina-
  These two probes are commonly available together and can  be
run as a hydrologic suite. Fluid conductivity logs provide a con-
tinuous record of  the conductivity or resistivity of fluid in the
borehole, which may be related to the conductivity of fluids in the
adjacent formation. A temperature log made simultaneously with
a fluid conductivity log allows  the most accurate conversion  to
specific conductance and also may identify contaminated or more
permeable zones in the formation.
   Fluid conditions in the boreholes should stabilize prior to im-
plementing this technique. The longer the period of time between
the drilling of the borehole and the logging of the hole, the more
accurate the results will be. Hydrologic logging can detect geo-
thermal gradients in groundwater in holes on the  order of 100 ft
thick.  Fractures contributing to groundwater movement some-
times  can  be noted  by changes  in  the geothermal  gradient.
Changes in temperature or conductivity also may be an indication
of contamination.

Downhole Television
   Downhole television has been  utilized  successfully to gather
in-situ information on boreholes and  wells in several  ground-
water  monitoring  programs. Borehole  television surveys are a
viable  alternative to  other  downhole instruments in the subsur-
face investigation stages of a groundwater monitoring program.
   Miniature borehole television cameras, developed for use in the
examination  of  nuclear reactor cores,  have been modified for
use in  borehole investigations. The lens attachments are capable
of looking sideward  or downward and include built-in lighting
  The in situ characterization of fractures that can provide path-
ways  for contaminant migration is critical in some investiga-
tions. Borehole television inspection can provide information on
the frequency, size and orientation of these fractures.  Vertical
correlations of rock cores in areas where voids are present (i.e.,
deep mining  or karst  topography) also can be simplified by this
technique.  Borehole  television can be used to check monitoring
well integrity. Casing inspections are especially useful  for con-
struction inspections when construction details are not known.
Well screens may be inspected in place to determine if rusting has
enlarged the screen openings or if screens have been damaged dur-
ing emplacement or well development operations. This informa-
tion may be  invaluable in the decision to decommission a well.
This technique is quick, inexpensive and creates a permanent
record for potential court cases.

Packer Assemblies
  Downhole  packers were developed for in situ measurements of
permeability  in geotechnical investigations such as dam  building
where data are critical. The packer assembly consists of two sets
of packers, upper and lower,  separated  by a perforated  pipe.
Each  packer is surrounded by a rubber seal  which is  fastened
to the assembly. Packers are inflated with compressed  air from
the surface to seal against  the  sides  of the borehole. Particular
zones then can be isolated for testing.
   For in situ  permeability testing, water is pumped in at a low and
constant pressure,  creating a constant  artificial head. Operating
pressures are stabilized with a gate valve until they remain oon-
stant.  Pressures then  are checked with a pressure gauge and  re-
corded for later use in permeability calculations. Each zone is
tested  for a defined period of time, and the amount of water re-
ceived by the zone is recorded with a standard water meter.
  Another  recent development is the insertion of electric or blad-
der pumps inside the assembly to remove water from the  bore-
hole. A particular zone can be isolated and tested with a  portable
GC for identification of zones carrying volatile organic  contam-
ination. Screens then can be set at optional levels for the study of
groundwater contaminant migration.

   Equipment which was not discussed in this manuscript includes
air sampling equipment. Generally,  air sampling programs are
highly specific. Particulates such as asbestos can be collected on
filters  by high volume air  samplers, while organics  can be col-
lected  in packed tubes for  thermal desorption  by the use of low

volume air samples. Air investigations generally include the use of
continuously reading portable meteorologic stations.
  Air sampling technology recently has been extended to plume
tracking through the analyses of soil gas. Tubes are placed in the
ground, usually for several days. Organic compounds which vol-
atilize  from groundwater or soil travel upward over time in the
unsaturated zone of the soil and collect a trace levels in the tubes.
The tubes then are removed,  thermally desorbed and  analyzed
for volatile organics. The technique is not yet accepted as a stan-
dard methodology, but has been  useful in many cases and cur-
rently is gaining wide acceptance.
  Groundwater  pumps continue  to be refined, especially dedi-
cated systems. These systems include small diameter electric sub-
mersible pumps and bladder pumps. Materials used in the pumps
are being researched. Many bladder  pumps currently  available
for purchase are built almost completely of Teflon.
  Technological developments continue in  the area of contin-
uous  measurement devices for groundwater levels in monitor
wells. These devices range from single well units with solid-state
memories to full-scale 16-channel devices. Some multi-channel
devices can be operated up to one mile away from a well, making
them useful for aquifer tests.  Some actually have  built-in com-
puters, disk drives, monitors and printers and can analyze aquifer
test data, keep records on each well and even generate reports.
  The general structure of an investigation requires an analysis of
data gaps, particularly with regard to source definition, pathway
examination and assessment of potential for contaminant migra-
tion. Cost and time constraints involved in collecting certain types
of data can be balanced against the usefulness of the results. For
instance, it sometimes may be less expensive to install additional
wells than to use a costly indirect geophysical technique.
  Investigations generally rely on  indirect investigatory tech-
niques to start, followed by necessary confirmation steps. Spe-
cifically,  a technique such as magnetometry can be used to search
for buried containerized wastes, but this generally is confirmed
through test pitting and subsurface waste analyses. It is impor-
tant to note, however, that the limitations inherent in some tech-
niques may preclude their use in certain situations. A knowledge
of these limitations  is critical to cost-effective investigations and
proper cleanup strategies.
  Knowledge of indirect measurement techniques and field ana-
lytical equipment is critical to the collection of proper data and
the control of costs. Innovative technologies constantly are being
considered for their usefulness in these investigations. Use of this
type of equipment to paint a complete picture of site conditions
can  help the environmental management  professional  reduce
long-term liability and provide  useful data  for risk assessments
and feasibility studies. We hope that we have provided a basic
overview of the equipment,  its  limitations and its usefulness  to
environmental site investigations.

               A Statistician's  View  of Groundwater  Monitoring
                                              Douglas E. Splitstone
                                                  IT Corporation
                                            Pittsburgh, Pennsylvania
   Owners and operators of hazardous waste facilities are required
 to monitor the groundwater around the hazardous waste facility
 and perform prescribed statistical tests on the resulting data to
 detect the existence, if any, of groundwater contamination. The
 statistical test prescribed by the U.S. EPA, which is referred to as
 Cochran's Approximation to the Behran's-Fisher Student's t-test,
 has  received increasing  criticism  as  being  inappropriate  and
 resulting in the allegation of groundwater contamination when, in
 fact, contamination did not occur. In addition, its specified ap-
 plication requires the existence of a well that  clearly monitors
 groundwater upgradient  of the site. The  U.S.  EPA recognized
 these problems. In publication of its regulations for owners and
 operators of permitted waste facilities, the U.S. EPA indicates the
 admissibility of statistical procedures alternate  to the Student's
   The inadequacy of the  Student's t-test is in part due to the lack
 of recognition that background groundwater quality varies both
 spatially and temporally. This paper describes a statistical model
 appropriate for describing spatially and temporally varying back-
 ground  water quality. An alternate statistical test procedure for
 assessing groundwater contamination as  a  deviation  from this
 background model is proposed. Use of this test procedure permits
 the owner/operator to control  the risk o' "false positive" test
 results to a small specified probability. Thus, the risk of unfairly
 being required  to conduct  an expensive  groundwater quality
 assessment program is  controlled while  assuring  protection
 against groundwater degradation.

  A sampling program to determine whether a particular hazar-
 dous waste facility is contaminating groundwater  should be
 designed so that any contribution from the particular facility can
 be distinguished from other sources. These other sources, which
 may be both anthropogenic and natural, provide what is loosely
 referred to as the background. The design of a sampling program
 to distinguish the contribution  of the site of interest from the
 background is complicated by temporal and spatial variations in
 the background.
  Temporal variations in  groundwater may, in part, be explained
by recharge rates which  are weather related. Spatial  variations
among wells may be due to a variety of  reasons.  Some of this
variability may be reduced by ensuring that the wells monitoring
the aquifer beneath a particular waste disposal  site are located
relatively close together and drilled to the same depth. Neverthe-
less, spatial variations may occur even if there is no contribution
from the site being monitored. The variations might be expected
to be horizontal gradients across the site.
  The Behrens-Fisher Student's t-test recommended by the U.S.
EPA for monitoring  the groundwater near possible sources of
contamination1-2  has  increasingly become the topic of critical
reviews.3-4-5 The U.S. EPA procedure is based on establishment of
a background level before the disposal facility is put into opera-
tion and is, therefore, insensitive to spatial and temporal varia-
tions in the background. It can be used, however, if variations in
the background levels are taken into account.
  Considering the hypothetical situation in which measurements
of groundwater  quality  are  obtained without  sampling or
measurement  error,  the  sample variance  of upgradient  well
measurements, SB2,  reflects  only  real background  variations.
Measurements  on subsamples from  the sample  taken from a
downgradient well all have the same value Xm. The test statistic
given by U.S. EPA, which is referred to as Cochran's approxima-
tion to the Behrens-Fisher Student's t-test,' reduced to:

                   (xm  ' XB)/(SB2/MB)1/2               (1)
where XB  is the mean of the background values and NB is the
number of background values. This test is inappropriate because
in the absence  of contamination, Xm can vary as widely as the
background values. The test is appropriate for the inappropn tte
null hypothesis that Xm  equals the mean of the background
values. The appropriate null hypothesis for testing is that Xm is a
member of the same population as the background values. The
appropriate statistic for this test is:
  Compared to the first test statistic, the estimate of the standard
deviation used in the second is considerably larger. Therefore, for
many sets of measurements, while the first test statistic may er-
roneously indicate  that there is contamination, the second test
statistic may indicate the opposite.
  The U.S. EPA has recognized this problem. In publication of
its  regulations for  owners  and  operators of permitted waste
facilities,  the  U.S.  EPA acknowledges  the admissibility  of
statistical procedures alternate to the Students t-test.1 An alter-
nate statistical  procedure is described below.

  Although the background quality of groundwater varies both
spatially  and temporally,  temporal variations in background at

nearby points are expected to be related. Extending this assump-
tion, temporal variations (apart from sampling and measurement
error) are the same at nearby points. This procedure leads to the
expression of a mathematical model for the background quality
of groundwater around a waste disposal site. In this model, the
background quality is given by the sum of a general mean (/*), a
spatial effect (Wi) and a temporal effect (Tj).
  If contamination of the groundwater occurs  from the  waste
disposal facility, and if the monitoring wells surround the site so
that they are located both upgradient and downgradient from the
site, then contamination is unlikely to change the levels of a con-
stituent in all  wells by the same amount. Thus, contamination
from a disposal site represents a deviation from  the background
model, and such a deviation  can be detected.
  Monitoring groundwater quality around a disposal site is an ex-
periment with two factors: a spatial factor (or well factor)  and a
temporal factor. In conducting such an experiment, samples are
collected and measurements  made at each well at each sampling
time.  If the  background model holds,  these  measurements,
denoted by Yij, can be represented by:
                         M  +  W.
                           = 1
                           = 1
 where N is the number of wells, M is the number of sampling
 times and Ejj represents the random sampling and measurement
 error. The measurements can be arranged in a two-way table and
 can be analyzed by a two-way analysis of variance.6
   Estimates of n, Wi and Tj may  be obtained from  the data.
 Denoting these estimated by m, w; and tj, respectively, they are
 given by:


 The differences between the observed values, yij, and the values
 predicted by the model are the residuals
                 rij = Xij - wi  -  tj - m               (7)

  As mentioned above, the contamination of groundwater by a
 waste disposal site will result in a deviation from the background
 model. Therefore, it is the residuals which contain the answer to
 the question of whether the site is contaminating the  ground-
 water. However, the residuals also are  affected by the sampling
 and measurement error. Because the assumed background model
 does not indicate whether the residuals  occur as a result of sam-
 pling and measurement error or are indicative of contamination,
 the model is incomplete.
  The analysis of  groundwater data  can  proceed no  further
 because there is no way to decide whether the residuals from the
 background model can  be  attributed  solely to  sampling and
 measurement error. The solution to this problem is the collection
 of replicate samples.  Replicate samples differ from each other
 only because they each contain an independent replicate of the
 sampling and measurement error. For example, a sampling design
 for groundwater monitoring might  specify collection   of two
 replicate samples from each well at each sampling time. Replicates
 are important in environmental sampling because the adequacy of
 environmental models is always in question.7 Replicates allow the
 variations unexplained by the model to be compared with the
 variations caused by sampling and measurement so that model in-
adequacies can be detected.
  The exact specification of the replicates that are needed is a
complex process since it depends on the environmental model.
For this reason,  replicate laboratory measurements obtained  by
splitting samples are rarely adequate because they do not reflect
sampling error. Specification of replicates often involves choosing
the time period between successive replicate samples. The model
should account for variations over long time spans with the term
Tj. Sampling  error usually accounts  for local and short-term
  Assume that true replicate samples can be obtained by resam-
pling the wells after a specific short time period. Consider a sam-
pling  program design which requires L replicate samples from
each well at each occasion that the wells are sampled. In most
cases, two (L = 2) replicates seem reasonable.  The measurements
on the samples may then be described by the  following model:

where, as before,
    i = 1 . . .  N (number of wells)
    j = 1 . . .  M (number of sampling periods)
    k =  1  . . . L (number of replicates)
  This model is the same as Model (3)  except that another
subscript has been added to index the replicates and a term (WT)jj
has been added to account for the various changes that occur
when the site is contaminating the groundwater. This term is
referred to as the well by period interaction and describes changes
in the relative relationship of the wells between time periods.
  If the sampling and measurement errors Eijk are statistically in-
dependent and normally distributed with the same variance, then
measurements obtained under this model can be analyzed by the
two-way  analysis  of variance appropriate  for models  with
replicate sampling.6 Estimates of /t, Wi, Tj and (WT)jj, which are
denoted by m, wj,  tj and ry, can be obtained as before with yij
replaced by:
The appropriate analysis  of variance (ANOV) table is given in
Table 1.
  The crucial comparison of the interaction with sampling error
which is indicative of groundwater contamination is made using
the F-statistic.6 The value of this statistic is given by:

                  NT(L-l)    45rij2           ,        (7)
            F =
  If the interaction is found to be significantly greater than the
sampling error, then one concludes that the wells do not vary in
the same way over time, and thus there is local influence on the
measured contamination.
  True replicate samples can be obtained by sampling the same
well after a fixed period. The duration of this period may be as
short as hours or as long as days. The determi.iation of this period
can be established for a given well system with an initial sampling
experiment. Such  an experiment has been conducted on the
monitoring wells of a major industrial facility.

  The need for determining what constitutes a replicate sample
has been discussed above. It also was indicated that replicate
analyses of the same sample do not provide for the determination
of sampling variability. To determine true replicate sampling for
the monitoring wells, a sampling experiment was conducted over
a four-day period. This experiment required the sampling of each
well once a day on each of the four days. The samples were col-

                         Table 1
           Appropriate Analysis of Variance (ANOV)
                 for Groundwater Monitoring
               Table 4
Groundwater Monitoring Study TOC (mg/l)
Between Wells
Between Sampling
Sampling Error
Periods T-l
Table 2
Groundwater Monitoring Study pH
B i
t 1
LN1 t2
1 1
, II 2
L. . r. .
'J ']
"J ("ijk" " ij.)
"i t"ijk m'2
(Standard Units)
B2 B3 B4
1 1
2 1
3 1
4 1
                         Table 3
          Groundwater Monitoring Study Conductivity
                    OtMHOS @ 25 °C)

1 1
2 1
3 1
4 1
B2 B3








B2 B3
Study TOX
B2 B3

                                                              lected after three well volumes of water were removed from the
                                                              well. Each sample was analyzed four times for pH, conductivity,
                                                              total organic carbon (TOC) and total organic halogens (TOX).
                                                              The data collected during this experiment are found in Tables 2
                                                              through 5.
                                                                 The sampling  plan was  designed to compare the variability
                                                              among samples from a given well taken within a relatively short
                                                              time span (approximately 24 hr apart) to the variability among
                                                              repeated analyses on the same sample. The  sampling plan is,
                                                              therefore, hierarchical, and the data were analyzed accordingly
                                                              using the appropriate  ANOV technique.' The  results of these
                                                              analyses are found in ANOV Tables 6 through 9.
                                                                 The values of the F statistics  shown in these  tables are all
                                                              greater than the tabulated critical values of F for the appropriate
                                                              degrees of freedom.* They clearly indicate that highly significant
                                                              variability exists  in all the measured constituents among the wells
                                                              and among samples from the same well during a comparatively
                                                              short period  of  time. The  significance of variability along the
                                                              wells is as expected for reasons given above. The significance of

the variation among analyses of a given sample clearly illustrates
that sampling variability is important.
  To  further illustrate this point, assume that the four samples
taken from Well Bl describe the background groundwater quality
for TOX. Thus, the background  mean TOX is 16.19 /ig/1, and
standard deviation  of the background TOX  is 6.11  /ig/1.  Now
assume that downgradient Well B4 was to be samples on one of
the four days of this experiment chosen at random. The mean and
standard deviation of the four  repeated analyses for TOX of the
daily samples is given in Table 10.  Also presented in this table are
the calculated and critical values of the Student's "t" statistic as
prescribed by Appendix IV of 40 CFR 264.
  Comparing the calculated and critical values of the Student's
"t" statistic for these four consecutive days serves to illustrate the
importance of the  sampling variability. On two of the four days
(Day 2 and Day 3), the calculated value of Student's "t" exceeds
the critical value. Thus, a conclusion of groundwater degradation
would be made. However, this is not true of the samples taken on
Day 1 and Day 4. If the quarterly sampling day were picked at
random  from these four, there would be a 50% chance of con-
cluding that groundwater degradation occurred solely due to ran-
dom variation introduced by the act of sample collection.



Table 6
3 6.0162 2.0054 51.03
12 0.4713 0.0393 12.61

48 0.1500 0.0031

63 6.6375
Table 7
Conductivity ANOV
3 3,578,628 1,192,876 127.81
12 111,990 9,333 43.21

48 10,348 216

63 3,700,968
Table 8
3 48,326 16108.70 44.22
12 4,372 364.32 132.96

48 132 2.74

63 52,829




Table 9
Well 3 2,918 972.63 17.26
Sample 12 676 56.36 12.44
Analyses 48 218 4.53
TOTAL 63 3,811

                Table 10
TOX Concentration for Downgradient Well B4
                                                                                          Table 11
                                                                        Expected Mean Squares for Groundwater Monitoring
                          EXPECTED MEAN

                                                                           Samples Within

                                                                          Analyses Within
                                                                                          Table 12
                                                                    Variance Components for Groundwater Monitoring Parameters
Analyses, o.
Sampling , o
                                                                   The selection of a sampling plan for continued groundwater
                                                                 monitoring can be made with the information generated by the
                                                                 analysis of variance. By equating the mean squares estimated by
                                                                 the analysis of variance to their theoretical expectations, one can
                                                                 estimate the contribution of each component (e.g., sample collec-
                                                                 tion or analysis) to the  total variability of the measurement. If a%
                                                                 and as are used to symbolize the variability due to analysis and
                                                                 sample collection respectively, then the expected mean squares for
                                                                 the analysis of variance are given in Table 11.
                                                                   In Table 11, W2 symbolizes the natural background variation
                                                                 among wells.  The resulting estimates of a2, and a^are given in
                                                                 Table  12 for the indicator parameters observed.
                                                                   The characterization of the contaminant concentration for a
                                                                 well undergoing quarterly monitoring is by the mean of the obser-
                                                                 vations taken during the quarterly monitoring period. A sampling
                                                                 plan should be designed to minimize the variation in this mean. If
                                                                 S samples are taken during the quarterly monitoring period and A
                                                                 analyses are run on each sample,  the variance of the quarter mean
                                                                 is given by:               2     2
                                                                                                        MONITORING    11

  The effect on this variance of varying the number of samples,
S, and analysis per sample, A, then can be investigated. If the
analytical cost is constrained by requiring the S + A £.4, then the
minimum  variance, and  hence maximum  precision,  of the
quarterly mean value for the indicator parameters for a given well
is obtained for the same analytical cost if the well is sampled four
times each  quarter approximately 24 hr apart and one analysis is
performed  on each sample. These results are given in Table 13.
The information  on the last line of the table indicates  that the
precision of the quarterly mean is approximately doubled (the
variance halved)  from  the U.S. EPA  recommended  sampling
scheme of one sample per quarter by sampling each well twice and
analyzing each sample once. This reduction in analytical cost may
be sufficient to compensate for increased sampling costs.
                           Table 13
                 Variance of Well Quarterly Mean

   The salient conclusions are summarized as follows:
 • The portion of the total measurement variability due to the
   act of sampling is much greater than that due to repeated chem-
   ical analyses of the same sample. Thus,  repeated chemical
   analyses of the same sample do not necessarily improve quar-
   terly characterization of groundwater.
 • Adequate replicate samples may be obtained by repeated sam-
   pling of each well approximately 24 hr apart.
 • The precision  of the quarterly measurement for groundwater
   contamination will be greatly improved from that specified in
   Federal Regulations if each well is sampled twice within a four-
   day time span and only one  analysis is performed on each
   sample. This sampling  scheme also  will permit the use of the
   analysis of the variance statistical procedure to adequately test
   the hypothesis of groundwater contamination.
•  Continued use of the statistical procedure specified in the Fed-
   eral Regulations  will, more  than likely, yield results alleging
   that groundwater contamination has occurred when the appli-
   cation of adequate statistical  procedures  will conclude that
   there is not contamination.
1.  U.S. EPA,  "Interim Status Standards for Owners and Operators of
   Hazardous  Waste Facilities." 40 CFR 265.92.
2.  U.S. EPA, "Regulations  for Owners and Operators of  Permitted
   Hazardous  Waste Facilities," 40 CFR 264.97.4.ii.
3.  Liggett, W., "Statistical  Aspects of Designs for  Study Sources of
   Contamination," Quality  Assurance for Environmental Measure-
   ments,  ASTM STP867, June 1985.
4.  Miller, M.D. and Kohoul.  F.C., "RCRA Ground Water Monitoring
   Statistical  Comparisons:  A  Belter  Version of Student's T-Test,"
   Proc. of the NWWA/AP1 Conference on Petroleum Hydrocarbons
   and Organic Chemicals in Ground Water Prevention, Detection, and
   Restoration, National Water Well Association, 1984.
5.  Ross, L. and Elton, R., "Maximizing the Statistical Performance of
   Ground Water  Monitoring Systems,"  Proc.  of  the NWWA/APl
   Conference on Petroleum  Hydrocarbons and Organic Chemicals in
   Ground Water  Prevention.  Detection, and  Restoration, National
   Water Well Association, 1984.
6.  Snedecor, G.W. and Cochran, W.G.. Statistical Methods, 5lh ed.,
   The Iowa State University Press, Ames, I A, 1950.
7.  Krumbein,  W.C., "Experimental Design in the  Earth  Sciences,"
   Transactions of the American Geophysical Union,  36, 1955, 1-11.
8.  Kempthorne, O., The  Design and Analysis of Experiments,  John
   Wiley and Sons, New York. NY. 1952. 104-109.

                       Data  Evaluation  in a  Groundwater Study
                               of Waste  Management  Practices
                          in  the  Phosphate Processing Industry
                                             Edward W. Mullin, Jr.
                                                  Jack  S. Greber
                                              William E. Thompson
                                               PEI Associates, Inc.
                                                 Cincinnati, Ohio
  The phosphate processing industry  has  been the focus  of
several groundwater monitoring programs within the past several
years. As a result of these studies, a relatively large data base has
been   assembled  which,   in   addition  to   evaluating  the
characteristics of the various wastes generated by this industry,
also includes groundwater monitoring data  which were used to
evaluate the impact of the industry on groundwater quality. The
interpretation of the monitoring data involved the use of several
different evaluation techniques which will be discussed.
  Student's t-test is one traditional method of evaluation that was
used.  This method compares the average and standard deviation
of a set of paired values. This method has been used extensively in
hazardous  waste-related groundwater  studies  and has been
criticized for its lack of sensitivity.
  Another traditional method of evaluation is the comparison of
Stiff diagrams.  These diagrams are graphical representations of
the basic quality of a water sample. This method is very useful
when dealing with the wastes from the phosphate industry,  but
the method is somewhat subjective.
  Several newer methods also were used in the  course of these
projects. These  methods included the calculation of ion  indices
and the use of cluster analysis.
  The ion indices are calculated using basic water quality com-
ponents such as  sodium and chloride. Through the comparison of
indices from impacted wells with indices from background wells,
an evaluation of contamination and attenuation can be made.
  Cluster analysis was used to identify any trends or relationships
between monitoring wells.  Cluster analysis  is a collective term
covering a variety of statistical techniques for delineating natural
groups or clusters. Evaluations were made based upon the degree
and the order in which the individual wells were clustered. This
method proved  to be  quite useful in confirming trends which
other methods also identified.
  Over the past several years, PEI Associates has been involved
with evaluating the waste management practices in the phosphate
industry. These evaluations have involved the  installation  of
groundwater  monitoring  programs  at  phosphate processing
facilities. In the course of these projects, eight different facilities
with a total of 85  wells were monitored. Samples were collected
from these wells at varying times; each well was sampled at least
twice and analyzed for a wide variety of water quality parameters.
The interpretation of the monitoring  data involved the use  of
several evaluation techniques which will be discussed.
  Groundwater monitoring data collected for the purpose of ful-
filling RCRA groundwater monitoring permit requirements have
been evaluated using Student's t-test. The Student's t-test is the
principal criterion for testing hypotheses concerning two popula-
tion means. The t statistic is the deviation of a normal variable (X)
from its mean (/*) measured in standard error units (Sx).
    t = (X - M)/SX                                     (1)
  This test compares the average and standard deviation of the
measurements for each parameter with the same parameter as
measured in another well, usually the background well. This test
was used on some of the data. However, it was not used on all the
data since  there  were  not enough replicates  to make the test
statistically valid. At least three or more replicates are required to
make the test valid. Due to various constraints on the projects,
only two sets of samples were collected at some of the sites. In ad-
dition to the lack of sufficient replicates, the Student's t-test has
been  shown  to  generate false  positive  results  as  has been
documented  in  the Federal  Register.  Therefore,  additional
methods for evaluating the data were used.

  Stiff diagrams  are a traditional method of depicting the basic
water types. Only the major cations and anions typically are used
in the  diagrams (Ca + 2, Mg+2, Na+, K + , C1-, alkalinity and
SO4-2). The  concentrations  of the ions are converted to milli-
equivalents/liter and then plotted on a Stiff diagram.  The shape
of the resulting polygon is indicative of the type of water being
evaluated.  For example, the Stiff diagram of a sodium carbonate
water  is distinctly different from  a calcium sulfate  water.
Therefore the comparison of a Stiff diagram of a sample of liquid
waste from a processing facility to the diagrams from monitor
wells will enable  one to graphically determine if there has been
any impact on the groundwater from the liquid waste.
  Figure 1  shows  such a comparison for one of the test sites. The
waste sample  (GL-1) and the background  well are compared to
two different wells from the site (PEI-2 and PEI-3). Neither of the
diagrams for PEI-2 or PEI-3  resemble the diagram for the back-
ground well, indicating that the water quality in these wells differs
from background conditions. The Stiff diagram for PEI-3 bears a
close resemblance to the diagram of the waste (GL-1), indicating
that some  contamination  has occurred in this  well. The Stiff
diagram for PEI-2 does not  resemble the diagram for GL-1 as
closely as  PEI-3, but  is  more similar to GL-1  than  to the
background well,  and therefore, it can be concluded that PEI-2 is
contaminated but not as severely as PEI-3.
                                                                                                    MONITORING    13

          50       25       0       25       50
           Cations (mg eq'l)  Amons (mg eq/l)
             BACKGROUND vs. PEI-2
          100      50       0       50      1UU
           Cations (mg eq/l)  Anions (mg eq/l)
               GYP LIQUID vs. PEI-2
                                                                50       25      0       25       50
                                                                 Cations  (mg eq/l)  Anions (mg eq/l)
                                                                   BACKGROUND vs. PEI-3
                                                                100      50      0       50      100
                                                                 Cations (mg eq/l)  Anions (mg eq/l)
                                                                    GYP LIQUID  vs. PEI-3
                                                        Figure 1
                                              Stiff Diagrams from the Test Site
   Another method which has been used with some degree of suc-
 cess is the calculation of ion indices. These indices were developed
 by the U.S. Geological Survey during an evaluation of phosphate
 mining and processing in  Florida.1 The indices are a numerical
 evaluation  of the concentration of the major anions and cations
 within a sample and in some cases the sample concentrations were
 compared to the concentrations in the background samples.
   The two  most useful indices are the index of Nonmineral Input
 (INMI)  and the Percentage  of Ion Increase (PII). The  INMI
 measures the percentage increase in concentrations of sodium
 above chloride and calcium plus magnesium above sulfate plus
 alkalinity compared to the same parameters in the background
                                C1 + A1
Mg-Mg ') - (S04-S04 '+Alk-Alk ') I - 1C-A ] - | C' -A'
              c, + Ai
                                              x 100

  Index of Nonmineral Input, in percent above
  Alkalinity, in milliequivalents/1
  Cation sums, in milliequivalents/1
  Anion sums, in milliequivalents/1
  Background conditions
                                                      The  PII is used to evaluate the increase in ions above back-
                                                    ground concentrations and is based solely on the sum of selected
                                                    cations and anions.
     PII =  Percentage of Ion Increase, in percent above
      C =  Cation sums, in milliequivalents/1
      A =  Anion sums, in milliequivalents/1
       ' =  Background conditions

  The senior author of the U.S. Geological Survey report1 consid-
ered an INMI of approximately 5% to be above background
levels for the groundwater systems in Florida and to reflect an im-
pact from the waste management practices. The relationships be-
tween the INMI and PII also can be used to evaluate the extent of
seepage and whether  or not the soil is attenuating the seepage
through ion exchange. For example, if the INMI and PII are both
significantly elevated above background conditions, this probably
indicates that the seepage has not been altered by ion exchange. If
the INMI is low and the PII is high, then this result may indicate
that the seepage is being  altered through ion  exchange with the
soil or that another source of contamination is influencing the

  At  the  test site  previously  mentioned,  ion indices  were
calculated for each sampling period. Table 1 shows the indices
(calculated with PEI-1 as background) for the second sampling
round. As can be seen from the data, the conclusions concerning
PEI-2 and PEI-3 are supported by the indices.  These two wells
have the highest indices for both INMI and PII. The lower INMI
for PEI-2  indicates that more attenuation has  taken place  at
PEI-2 than PEI-3,  while the other wells show some degree of im-
pact. The high INMI for PEI-5, the second background well, is
most likely a result of it being completed in a different lithology
than PEI-1, rather than it showing the effects of seepage.

                          Table 1
                   Ion Indices at the Test Site
  These methods have proven to be quite useful for evaluations
of the phosphate industry where the waste being studied consists
mainly of calcium and sulfate, both of which are included in the
Stiff diagrams and  the indices. However, these methods would
not be very useful in evaluating seepage  from wastes which did
not impact any of the major cations or anions, such as  organic
contamination, and these evaluations often may be subjective
when there is not a great deal of dissimilarity in the diagrams or
the indices.

  The ground water data collected during these projects also were
analyzed by cluster analysis to identify any water quality trends or
relationships between the monitoring wells and the waste samples.
Cluster analysis  is a collective  term covering a wide variety of
statistical techniques for delineating natural groups or clusters in
data sets.2 The objective of this type of analysis is progressive
grouping of the variables into clusters according to their degree of
statistical similarity.
  The  cluster procedure in  the  BMDP  Statistical  Software
package was used for this analysis.3 The exact procedure used was
Cluster Analysis of Cases (Method 2M). This procedure uses a
hierarchial approach to building the clusters. Initially, each  case
(an individual sample from a monitoring well or waste sample) is
treated as a single member cluster; then, the two most closely
related cases are combined into a new cluster and the "distances"
between the new cluster and the remaining ones are computed.
  The  distance  measure calculated is the  Euclidean  distance
radius between the  clusters. The  general formula used for this
calculation is as follows:
  djk   = distance between two cases of clusters j and k
  Xjj   = value of the ith variable in the jth case
  This process continues until all the cases are combined into one
large cluster. Figure 2 is a graphical representation of this process;
the initial cases are placed on a line in an order that is determined
by the program. Vertical lines then are drawn from each case until
it joins with another. By using the computed distance between
clusters as the vertical scale, the vertical location of each joint or
cluster shows the dissimilarity of the joined groups. Therefore,
clusters formed  near the  top represent homogeneous groups,
whereas clusters  formed further down represent clusters  formed
only because the process  goes to its logical end  of one large
                                                                                  Cluster Diagram
                                                                      E  --

                                                                                            Figure 2
                                                                                 Example of Hierarchial Clustering

                                                                    For the purposes of these studies, each sample from a monitor-
                                                                  ing well or waste sample was treated as one case, and all samples
                                                                  from a given sampling period were  analyzed together; i.e., no
                                                                  clustering was conducted over the different sampling periods. The
                                                                  variables  used in the analyses included the basic  water quality
                                                                  parameters (Ca, Mg, Na, etc.)  and several parameters chosen as a
                                                                  result of their  concentrations in the waste (sulfate,  fluoride,
                                                                  Ra-226, etc.).
                                                                    The results of the cluster analysis were evaluated based on the
                                                                  criterion that any increase in distance between clusters that was
                                                                  twice the  previous distance (a  100% increase) was considered to
                                                                  be a significant difference between clusters.
                                                                                     PEI  PEI  PEI  PEI   PEI  PEI   0L
                                                                      ~   100-1-
                                                                      =   150-f-
                                                                      £   200- -
                                                                    a E   250- -
                                                                    O (ft
                                                                    S »   300- -
                                                                           1000- -
                                                                           2000- -
                                                                         20000- -
                                                                         30000- -
                                                                                           Figure 3
                                                                                        Test Site Cluster
                                                                                                         MONITORING    15

  The cluster analysis  of the data represented previously in the
Stiff diagrams supported the conclusions of the Stiff diagrams.
As  shown  in Figure 3, the six  wells analyzed showed definite
clustering.  The background wells (PEI-1 and PEI-5) were deter-
mined to be similar on the basis of the clusters, as were PEI-4 and
PEI-6 (these wells were furthest downgradient  from the waste
source). Wells PEI-2 and PEI-3  also clustered together and were
determined to be the least similar to the other well clusters and the
most  similar to the waste sample GL-1.
  At  this site, all the data evaluation methods provided the same
conclusions as  to which wells were being affected by the waste
management practices. Cluster analysis is a worthwhile method to
assist in the evaluation of data when traditional methods may not
have  the necessary statistical power to sort out all the variables.
  The power of the cluster analysis may be increased if a factor
analysis is performed on the data set prior to the cluster analysis.
The factor analysis will identify which analytical parameters are
responsible for the majority of the experimental variance. Use of
factor analysis also will prevent the researcher from inadvertently
weighting the cluster  analysis by the inclusion of two or more
parameters which represent the same effect. The primary variable
should be identified  by the  factor analysis,  and  only those
variables should be included in the cluster analysis.
  As can be seen, there are a wide variety of methods available
for the evaluation of groundwater monitoring data. The use of
cluster analysis with or without factor analysis is a valuable tech-
nique which can help identify trends in the  data that otherwise
may not be recognized.

  The authors would like to thank the U.S. EPA and the Florida
Institute of Phosphate Research for their assistance in funding the
projects  which were discussed in this paper.

I.  Miller, R.L. and Suicliffe, H.. Jr., "Effects of Three Phosphate In-
   dustrial Sites on Ground Water Quality  in Central Florida, 1979 to
   1980," U.S. Geological Survey Water-Resources  Investigations Re-
   port 83-4256.
2.  Andberg, M.R., Cluster Analysis for Applications, Academic  Press,
   New York. NY, 1973.
3.  Dixon, W.J.,  Ed., BMDP Statistical Software, University of Cali-
   fornia Press, Berkeley, CA, 1981.

               Plant Cuticular and  Dendrochronological  Features
                                      as  Indicators  of Pollution

                                                G.K. Sharma, Ph.D.
                                                Biology Department
                                              University of Tennessee
                                                  Martin, Tennessee
                                                 Harvard University
                                             Cambridge, Massachusetts
   In a highly industrialized world, it is imperative that scientists
explore different methods of monitoring environmental contam-
ination. One new field of environmental research, begun in re-
cent years, may do this. This field is the study of leaf cuticle and
annual growth rings in woody plants.
   Several plant taxa growing in habitats characterized by varied
levels of environmental contamination were studied by the author
of this  paper for  dendrochronological and cuticular patterns.
Growth rings, ring and porous wood, and related features were
analyzed with a dendrochronograph. Cuticular patterns such as
stomatal frequency, stomatal size and  trichome frequency and
type were studied in relation to environmental contamination in
various habitats. Most plant populations under investigation were
in the industrial areas of Nashville and Memphis, Tennessee—
known for their high levels of environmental pollution.
   Statistical analysis of the data revealed the significance of plant
features as  indicators of environmental pollution in  the  area.
Growth  rings were,  for  example, extremely narrow  in highly
polluted areas, while the plant populations exposed to relatively
low levels of pollution exhibited fully  developed annual incre-
ments. Growth rings tended to be unusually irregular in plants
growing in highly polluted areas. Cuticular patterns such as tri-
chome frequency  and type were quite diagnostic for determin-
ing pollution levels in the environment.
   With an increase in pollution level, the stomatal frequency de-
creased  and the trichome frequency increased—possibly a re-
sponse to offset the detrimental effects  of pollutants on various
metabolic reactions and reproductive strategies in plants. In addi-
tion,  floral productivity  seemed  to  be adversely  affected in
polluted areas. The results demonstrate that tree ring character-
istics and leaf cuticular patterns can be used as monitors for eval-
uating the growth  responses of a plant species to varied levels
of environmental pollution.
  Analysis of annual increment variation  can be used to deter-
mine the degree of stress exerted by environmental pollution on
plant growth. Results of several cuticular and experimental stud-
ies conducted by the author substantiate  the findings reported
herein. These results suggest that plants can be used as extremely
reliable indicators  of environmental pollution, including hazard-
ous waste sites, provided  long-term monitoring is done for  a
variety of plant species.

  Studies3'5 have demonstrated the relationships between plants
and the environment. Damaging effects of fluorides and sulfur
 dioxide on plants have been documented.'
   In 1934, Chamberlin1 determined that industrial pollutants of a
 large mid-western city in the United States were detrimental to
 coniferous flora of the area. He especially referred to the fatal
 effects  of industrial pollution  on  Pinus banksiana.  Hill and
 Thomas3 observed that there was a decrease in alfalfa yield after
 exposure to sulfur dioxide.  Pyatf suggested lichens as possible
 indicators of air pollution in a steel-producing town in Wales. His
 studies  revealed that thallus size decreased and the lichen flora
 decreased in the number of  species present with increasing prox-
 imity of the pollution source.
   While working in the forests of the northwestern United States,
 Scheffer and Hedgcock5 found that sulfur dioxide from smelters
 produced  a characteristic mottling of leaves.  Coniferous  plants
 were more affected than deciduous species.
   Feder2 reported that geranium and  carnation plants had re-
 duced branching and retarded floral productivity after exposure
 to  low  levels  of  oxidant-type pollutants. Although  numerous
 plant taxa have been studies to determine the significance of
 various plant  morphological features as indicators of  environ-
 mental pollution,  relatively little work has been done to establish
 relationships between  cuticular dynamics and environmental
 pollution, especially hazardous waste sites. Recent studies7'8 on
 the subject suggest the potential of leaf cuticular dynamics in en-
 vironmental pollution research.
   The purpose of this study was to collect data for three plant
 species in  the relatively clean environment of Reelfoot Lake and
 to be compared to samples collected in the industrialized areas
 of Memphis and  Nashville. The present investigation is,  there-
 fore, a  continuation of a comprehensive study involving plant
 cuticular features as  indicators of environmental pollution with
 special reference to hazardous waste sites.

  Plant  populations of Polygonum pensylvanicum (smartweed),
Platanus occidentalis (sycamore) and Catalpa bignonoides (catal-
pa) were studied; comparisons of samples taken from the relative-
ly unpolluted sites of Reelfoot  Lake and Martin in  northwest
Tennessee  were made with ones collected from the-contaminated
sites in  middle  and northwest Tennessee—the latter  sites were
characterized t>y exposure to  additional pollutants emitted by
vehicular traffic and the industrial complex of the  surrounding
metropolitan areas.
  Twenty-five randomly selected leaves  were gathered from each
plant species.  Their  lengths,  widths,  internodal  lengths and
length-width ratios were determined. Gross morphological meas-
                                                                                                     MONITORING    17

urements were supplemented with cuticular data.
  For cuticular studies, representative leaves were washed with a
mild detergent and distilled water. After air-drying, the upper and
lower surfaces of the leaves were coated with Duco-cement™.10
Upon drying, fine layers of Duco-cement showing the cuticular or
epidermal complex of leaves were removed; slides were prepared
from the central portion of the layers for microscopic analysis.
Cuticular dynamics representing in stomatal frequency stomatal
size, epidermal  wall  undulation, trichome  frequency, trichome
length  and type  and  subsidiary cell  complex were recorded
(N = 25) for the  upper and  lower surfaces representing upper and
lower surfaces of leaves using 43x objective and lOx oculars.
  Dendrochronological studies were conducted on  the  woody
taxa by getting core samples in the two sets of habitats.  Annual
increments were recorded from the core samples. The entire data
were analyzed by a computer.

  Statistical analysis  of the data  revealed several significant dif-
ferences in the gross morphological, cuticular and dendrochron-
ological features of the two sets of samples (polluted versus rela-
tively unpolluted). Plants growing in the contaminated areas had
smaller leaves with both their length and width significantly re-
duced. In addition,  internodal length  also  was reduced. Plants
growing in the rural,  Reelfoot Lake area were healthier and had
larger leaves and longer internodes. These leaves were dark green,
while the leaves of plants from the contaminated sites had yellow-
ish,  light green  pigmentation.  It is  obvious that the photosyn-
thetic productivity of the plants at the contaminated sites was ad-
versely affected  by the  environmental pollution prevalent in  the
  A reduction in the total  biomoass of plant populations in  the
contaminated  areas  was an additional evidence to  suggest  the
detrimental effects  of environmental  degradation on  plant
growth. A reduction  in metabolic activities, especially in  photo-
synthesis, must mean reduced bimoass; this  result was evident in
all the populations  of contaminated sites. Cuticular measure-
ments revealed additional features affected by the environmental
contamination of  the  sites under  investigation.  Stomatal fre-
quency values  of the lower leaf surfaces in all plants under study
were high in the rural, less polluted habitats, while the plants
from the contaminated sites exhibited extremely  low stomatal
frequency values.
  In addition, trichomes on the leaf surfaces of plants from  the
latter sites had more numerous and much longer trichomes with
large bases. Similar results have been found in earlier studies6'7
in which affected plant populations were growing in sites char-
acterized by heavy vehicular traffic.  Stomatal size  differences in
the two sets of samples were not significant. Subsidiary cell com-
plex remained  unaffected by environmental  contamination in all
three plant taxa, and hence must be regarded as a reliable feature
for identification purposes of the species.
  The fact that some cuticular  and morphological features  of
some plant species were different in the two habitats is indica-
tive  of the evolutionary  trend exhibited  by plants growing  in
areas exposed to various kinds of environmental contamination.
Monitoring of these plant features over a long  period of time is
needed. This  study is underway in order to determine if these
morphological and cuticular trends are permanent.  If so, these
and  similar plant species might be used as indicators of environ-
mental complex, especially at contaminated sites.
  Preliminary investigations of the core samples of woody taxa
revealed a general decrease in the size of annual  increment. Addi-
tional  detailed studies are  underway to determine the usefulness
of plant features as monitors of environmental  pollution,  espec-
ially al hazardous  waste sites. Such  investigations might  reveal
plant species that  are able to  withstand  specific contamination
and  hence  be  of significance for  environmental and commercial

  Financial support for this work under the Research Grant Fund
of The University  of Tennessee at Martin is gratefully acknowl-

 I. Chamberlain, C.J., Gymnosperrru: Structure and Evolution, Dover,
   New York. NY. 1934.
 2. Feder, W.A., "Plan! Response to Chronic Exposure to Low Levels
   of Oxidant type Air Pollution," Environ. Pollut. I. 1970, 73-79.
 3. Hill, G.R.  and  Thomas,  M.D., "Influence of Leaf Destruction
   by Sulfur dioxide and Clipping on Yield of Alfalfa," Plant Physiol.
   8. 1933.334-345.
 4. Pyatt, B.F., "Lichens as Indicators of Air Pollution in a Steel pro-
   ducing Town in South Wales," Environ. Pollul. I, 1970, 45-55.
 5. Scheffer, T.C. and Hedgcock, G.C.,  "Injury to  Northwestern For-
   est  Trees by Sulfur dioxide from Smelters," U.S. Dept. Agr. Tech.
   Bull. 1117. 1955.
 6. Sharma, O.K. and Butler, J., "Leaf Cuticular Variation in Trifolium
   repens  L. as Indicators  of  Environmental  Pollution," Environ.
   Pollul. 5. 1972.287-293.
 7. Sharma, O.K.. Chandler,  C. and Salemi, L., "Environmental Pollu-
   tion and Leaf  Cuticular Variations  in  Kud/.u (Purrario  lobalo
   Willd.)," Ann. Botany. J5. 1980. 77-80.
 8. Sharma, O.K. and Tyree,  J., "Geographic Leaf Cuticular and Gross
   Morphological Variations in Liquidambar slyraciflua L. and their
   Possible Relationship to Environmental Pollution," The Botanical
   Gazette, 1.1-1, 1973, 179-184.
 9. Solbcrg. R.A. and Adams, D.F., "Histological Responses of some
   Plant  Leaves to  Hydrogen fluoride and Sulfur  dioxide," Am. J.
   Botany.  43. 1956, 755-60.
10. Williams, J.A.,  "A Considerably Improved Method  for Preparing
   Plastic Epidermal Imprints," The Botanical Gazette.  134,  1973,

                  Detailed Stratigraphic and  Structural  Control:
              The  Keys to Complete and  Successful Geophysical
                             Surveys  of  Hazardous Waste Sites
                                           H. Dan Harman, Jr., P.G.
                                            Engineering-Science, Inc.
                                                 Atlanta, Georgia
  Since the passage of RCRA and CERCLA legislation, count-
less geophysical surveys have been conducted at hazardous waste
sites as part of site investigations in an attempt to delineate the
horizontal and vertical extent of groundwater contamination as
well as to delineate the limits of the sites. Of utmost importance
during a geophysical survey is the identification of Stratigraphic
and structural control features  which influence the  potential
migration of leachate away from a site.
  Stratigraphic features such as  sand zones, clay lenses, outwash
valleys and top of rock zones are critical in  the overall hydro-
geological assessment of a site. Structural features such as faults
and fractures are equally critical in the  assessment of a site.
Proper use and interpretations of geophysical data can result in a
better understanding  of a site's subsurface characteristics even
before a drilling program begins.
  The geophysical data interpretations explained in this paper,
using  the "Modified  Wenner" method, are empirical  and have
resulted in the site-specific identification of subsurface details
which have been reasonably accurate as compared to subsequent
drilling results. The acquisition of detailed knowledge of a site's
subsurface is  the key to  a complete and successful geophysical
site survey.

  Geophysical surveys consist of surveying a site and its immed-
iate vicinity utilizing one or more remote sensing techniques. The
techniques  may  include  earth  resistivity,  electromagnetics,
magnetometry, seismic, metal detection and ground penetrating
radar.  Each technique has its  own advantages and  disadvan-
tages  at a particular  site. When groundwater contamination  is
a suspected problem, either earth resistivity or electromagnetics
normally is chosen as the technique in a geophysical survey.
  Among the many methods utilized within the  earth  resistivity
technique, the "Modified Wenner" method1 has been reasonably
accurate in terms of  depth investigated as compared  with sub-
sequent drilling results.  By utilizing the "Modified  Wenner"
method and  empirical interpretations,  site-specific identifica-
tions of Stratigraphic and structural control features are possible.
The identification of these features  is the key to performing a
complete and successful geophysical survey of a hazardous waste

  Stratigraphic and structural control is the understanding of a
site's  subsurface characteristics. When Stratigraphic and struc-
tural control is understood, a site's geological characteristic can
be understood. Furthermore, a site's hydrological characteristics
can be understood better in terms of aquifers, confining layers
and groundwater migration routes. Also, this  control enables
the investigator to more effectively perform  resistivity profiles
in an attempt to define the horizontal and vertical extent of iden-
tifiable contaminate plumes. By understanding the Stratigraphic
and structural control, the resistivity profile measurements can
be effectively placed in the appropriate subsurface zones in which
contaminant plumes are suspected.
  Stratigraphic  and  structural control is normally  established
only after an exploratory drilling program, but the control can
be established by the proper interpretation of resistivity sound-
ings. Soundings are  measurements of the earth's resistivity  at
various depths at a single land surface point.

  Resistivity  soundings,  utilizing  the  "Modified  Wenner"
method and empirical interpretations, have resulted in  site-spe-
cific subsurface identifications which have correlated well with
existing and/or subsequent drilling data. The key factor in the
interpretations of "Modified Wenner" method soundings is that
the potential or inner electrode spacing across the land surface
very closely approximates the depth below the land surface  at
which a measurement is taken; or, in general terms, "electrode
spacing equals depth below ground."
    Formula lor Apparent Resistivity
                   Current Meter       Battery
                  	(T)	1 ill	
                         Volt Meter
    SOURCE' Carrtaglon A wmon. lost
                          Figure 1
                   "Modified Wenner" Array
                  Diagram of Electrode Spacing
                                                                                                    MONITORING    19

  The electrode arrangement in the "Modified Wenner" method
is shown in Figure  1. In this arrangement, the current or outer
electrodes (C and C) are stationary while the potential or inner
electrodes (P and P') are moved  at equal distances from the
center of the array. The arrangement of the electrode array can
be varied depending  on the objective and depth to subsurface
  Numerous investigators have discussed the advantages and dis-
advantages of  empirical versus theoretical interpretations,  and
a discussion here will not be attempted. A partial list of refer-
ences  is included at the end of this paper.2"4 The author has ap-
plied the "Modified  Wenner" method and  empirical interpre-
tations  in  numerous subsurface investigations  because  of the
method's ease of operation and the relative simplicity of the inter-
pretations. The correlation between the sounding interpretations
and actual subsurface drilling data has been reasonably accurate.

  Five example soundings  are  presented to illustrate the appli-
cation of the "Modified Wenner" method and empirical inter-
pretations. The method has proven to be an asset in solving strati-
graphic and structural control problems, the  solutions of which
have  improved the effectiveness of subsequent resistivity profil-
ing, monitoring well  placements and groundwater contaminant
plume tracking.

  Resistivity soundings at a landfill in the  mid-western  states
were  conducted to aid in the  understanding  of the stratigraphy
underlying the  landfill as well as determining the thickness of the
landfill cap. Figure 2 illustrates a shallow sounding to 5 ft below
ground. An  apparent resistivity change is evident at approxi-
mately a 2 ft depth which was interpreted as the thickness of the
landfill cap. Similar soundings over the landfill were conducted,
and the apparent resistivity responses varied only slightly (mainly
due to the variation of cap saturation and lithology). A limited
number of shallow borings on the landfill had confirmed the cap
thickness to be  approximately 2 ft deep. Resistivity profiling iden-
tified variations across the landfill  which aided in the determina-
tion of possible recharge occurring through the cap.
                          P-Pl SPACING  (FEETI
  A deep sounding to 100 ft over the landfill was conducted to
aid in understanding the  deep stratigraphy underlying the site.
Figure 3 illustrates this sounding showing interpreted features
such as the water table, sand/clay lenses, limestone rock and pos-
sible fractures in the rock.  The actual drilling logs from both land-
fill borings and a nearby water supply well are presented to show
the correlation  with  the interpretations. Note the depth correla-
tion of the apparent resistivity graph to the boring logs, especially
the deflection of the  graph at 2 ft (cap-fill interface) and between
90 and 96 ft which correlates with the 90-ft water-bearing zone
identified by the water well log. The deep sounding at this land-
fill site aided greatly in  the  understanding of the  site-specific
stratigraphic and structural control.
 I  40

        -—3	j-sf—-I
                           «     SO    60

                           P-Pl V1CI-* !<•££']
                           Figure 3
                   Deeper Sounding at Land Till
  Soundings were conducted downgradient of a leaking lagoon at
a site in the southeastern states. This site is underlain by crystal-
line  rocks. The groundwater downgradient of the site contains
highly conductive salts which created special problems during the
soundings. Figure 4 illustrates -a 60-ft  sounding conducted with
all four electrodes in the center of the highly conductive ground-
water plume.  The graph  is essentially  flat,  and interpretations
are impossible to make.
                                                                                             !4    JO    36

                                                                                             P-Pl SPACING tfEETI
                          Figure 2
           Results of Resistivity Soundings at a Landfill
                           Figure 4
        Results of Sounding Downgradient of a Leaking Lagoon

  Figure 5 illustrates a 200-ft sounding conducted with the cur-
rent electrodes placed outside the plume. This graph is more rep-
resentative of the subsurface than is the graph in Figure 4. Note
that the low apparent resistivity between 50 and 69 ft on Figure 5
is not seen on Figure 4. The low resistivity values at this zone
were confirmed  to represent contaminated groundwater within
weathered and fractured crystalline rock. The sounding in Figure
5 aided in the understanding of the  stratigraphy and structure
below a highly conductive groundwater plume.
  200  ,




ui 120

I 100

i  eo

      0    20    10    60    30    100   120    140   1EO    160   200
                           P-P1 SPACING (FEET)

                            Figure 5
     Results of Sounding Near and Leaking L.agoon with Electrodes
                        Outside the Plume
  Understanding the stratigraphic and structural control under-
lying and in the vicinity of a waste site is critical in the overall
hydrogeological assessment of a site. Surface resistivity sound-
ings aid this understanding by yielding data which can be corre-
lated with existing and/or subsequent drilling results. Subsequent
geophysical surveys, resistivity  profiles for  example, can be
planned by the proper interpretation of the soundings in terms of
site-specific stratigraphy and structural features.




5  900

§  750
£  500




                           P-P1 SPACING (FEET!
                            Figure 6
  Results of Soundings in Sedentary Rocks to Identify Solution Cavities
   Soundings were conducted to various depths at a waste site in
 the southeastern states to aid in identifying solution cavities with-
 in sedimentary rock. The rock at this site is limestone with solu-
 tion cavities commonly occurring. The solution cavities are pos-
 sible avenues  of contaminant migration; therefore  the  under-
 standing of the subsurface structure at the site is critical. Figure 6
 illustrates one of the soundings conducted at the site.  The sound-
 ing graph correlates well with the actual drilling log. Note the
 low resistivity between 11 and 29 ft and the zone of staining and
 vugs present at 14.5,19.8 and 24.9 ft.
   As the degree of weathering in the rock decreases  with depth
 to 30.3  ft, the resistivity graph shows an  increase (more resistive
 rock). Yet, between 40.1 and 48.0 ft where a solution pitted zone
 and small vugs are present, the resistivity graph shows a corres-
 ponding decrease (more solution and weathering). The monitor-
 ing well subsequently installed  27  ft deep  at the sounding sta-
 tion  yielded groundwater  with higher conductivity  levels than
 background wells.  The resistivity sounding in Figure 6 aided in
 the understanding  of the solution  cavity distribution below the
   Application of the "Modified Wenner" method and empirical
 interpretations has resulted in reasonably accurate correlations
 between the interpretations and actual drilling data as well as
 groundwater quality data. The ability to predict the presence of
 subsurface features with reasonable accuracy is a real advantage
 and time-saving element in the decision-making processes con-
 cerning a hazardous waste site.

 1. Carrington, T.J. and Watson, D.A.,  "Preliminary Evaluation of an
   Alternate Electrode Array for Use in Shallow-Subsurface Electrical
   Resistivity Studies," Ground Water, Jan./Feb., 1, 1981.
 2. Moore, R.W., "An Empirical Method of Interpretation of Earth-
   Resistivity Measurements," American Institute of Mining and Metal-
   lurgical Engineering, Technical Publication No. 1743, July, 1944.
 3. Muskat,  M.,  "The Interpretation  of  Earth-Resistivity Measure-
   ments," American Institute of Mining and Metallurgical Engineering,
 4. Moore, R.W., "Geophysical Methods of Subsurface Exploration in
   Highway Construction," Public Roads, 26, Aug., 1950, 49-64.
                                                                                                               MONITORING     21

                     Detection and Measurement  of Groundwater
                            Contamination by Soil-Gas Analysis
                                                   H.B. Kerfoot
                                                   J.A. Kohout
                                                    E.N. Amick
                      Lockheed Engineering and Management Services Company, Inc.
                                                Las Vegas, Nevada
   A soil-gas sampling probe which can penetrate calcareously
 cemented alluvium was developed and evaluated above a chloro-
 form-contaminated groundwater plume. A linear correlation of
 greater than 95%  significance between groundwater and soil-gas
 chloroform  concentrations was observed. The precision of the
 method is controlled by sampling, with relative standard devia-
 tions between closely spaced samples ranging from 12 to 43%. A
 linear chloroform  depth profile was observed; the profile, as de-
 lineated by  the probe, agreed  with calculations made assuming
 diffusion-controlled  vertical flex of soil gases through the vadose

   Contamination of groundwater is a problem of increasing con-
 cern. In efforts to detect and measure groundwater contamina-
 tion, groundwater sampling and analysis is the traditional method
 of choice. Recently, preliminary surveys using remote-detection
 methods have been used as a tool for planning more cost-effective
 groundwater monitoring networks.' These surveys typically have
 been made  using  geophysical  instruments such as electromag-
 netic techniques that can be effective for locating inorganic spe-
 cies  but are  not useful in detecting organic compounds. For the
 detection and measurement of subsurface contamination by vola-
 tile organic compounds (VOCs), a new technology, soil-gas meas-
 urement, has been  developed.
   Soil-gas surveying originally was developed for oil exploration.1
 Applications of the technique to delineate subsurface contamina-
 tion  by VOCs have been developed by several workers.3'4
   Measurement of soil gases for detection of groundwater con-
 tamination takes advantage of Henry's Law, which states that
 the concentration  of a volatile compound in vapors that are at
 equilibrium with a VOC solution is directly proportional to the
 VOC concentration in solution. The relationship between these
 two concentrations is quantitatively described by the Henry's Law
 constant for that compound. The magnitude of this constant is
 directly proportional to the vapor pressure  of the compound and
 inversely proportional to the compound's solubility in water.
  Because of their  relatively low water solubilities and high vapor
pressures, VOCs in contaminated groundwater tend to be present
 in soil gases above the source. Table 1 lists the 17 substances most
 frequently encountered at Superfund  sites and their Henry's Law
constants; of these chemicals, 10 are amenable to soil-gas survey-
ing. In addition, major components of petroleum products which
can leak from underground storage tanks can be detected using
this technique.'
  This paper describes a study undertaken to validate equipment
and procedures for soil-gas surveying. The validation of the sys-
tem involved assessing  the bias of the technique for  indicating
chloroform contamination of groundwater, as well as assessing
the precision and short-range variability of the method.
                         Table 1
    Most Frequently Identified Substances at 546 Superfund Sites*
                                Henry*e L*" Constant  Percent
                                    (ppbv L/y9)      of aitea
   1   Trichloro«thyl«n*

   2   Le«d and compound*

   3   Tolu«n«

   4   B«nz«ne

   5   Polychlorinated blphcnyl* (PCB*)

   6   Chloroform

   7   Tctr*chloro«thylene





              20 b
Araenlc and coapounde
Cadpdun and coe^xMinda
Chroadua and coapounda
1,1. 1-Trlchloro«chane
Zinc and coag>ounda
Bthy Ibeniene
Hethylene chloride
trana- 1,2-dlchloroethylene
« 1
"Source: Kcrfoot, H.B. and Barro*\, L.J.. "Soil Gas Measurement for ihc Detection of Sub-
surface Organic Contamination," U.S EPA, l.*u Vegas, NV, 1986,
 Compound amenable 10 soil-gas surveying.
  The objectives of the study were to assess the bias and the pre-
cision of this newly developed technique for detection and meas-
urement of groundwater contamination by VOCs and to study the
vertical distribution of VOCs in the vadose zone. The study loca-
tion has a known chloroform-contaminated groundwater plume

                           Figure 1
             Survey Site Location, Henderson, Nevada
and an existing system of groundwater monitoring wells.  These
monitoring wells are separated by 200 ft and lie along a line per-
pendicular to the direction of groundwater flow.
  For evaluation of the method bias, soil-gas samples were taken
at a 4-ft depth at four locations 20 ft to the north, east, south and
west of  four monitoring wells  within  the boundaries of the
groundwater plume. To delineate the edges of the plume, addi-
tional samples were taken at a 4-ft depth 20 ft to the east of a well
at the western edge of the groundwater plume and 20 ft to the
                                                                            END CAP
                                                                         1/8" TUBING
                                                                       3/4" THREADS
                                                                       TUBING NUT
                                                                      1/2" THREADS
                                                   Figure 3
                                             Sampling Probe Design
   Well Location
   Water Table
   Sand and Gravel
Elevation (feet)
1640 —
1620 —
1580 —
~^^~ T— — • -r *
^ [ -] -j- •" 	 1 — r--T — '
655 650 645 640

•Sand 8. !
Gravel j
9' deep ! \
,,-S_ ~
^^^^ " .i
i ! I 1 ! i i i**^~
-_J 	 H 	 1 	 IWate Tabe 	 l__}-'l,^""
635 630 625 620 615 610
Stations Teit Well Eagt
0       600
 Scala in Feat
0          200
 Scale in Meter*
                                                             Figure 2
                                            Subsurface Hydrogeology at the Pittman Lateral
                                                                                                            MONITORING    23

west of a well on the eastern edge. Samples were taken in triplicate
and were analyzed by gas chromatography on site. The triplicate
analyses were used to assess the method precision.
  At one sampling location, four points separated by 3 ft were
sampled as a check on the short-range variability of the method.
Halfway between two of the wells, three points separated by 3 ft
were sampled in duplicate at depths of 1, 2, 3,4 and 5 ft.

  The geohydrology  of the study area  is relatively simple. Un-
confined groundwater occurs at a depth of 7 to 15 ft in calcified,
unconsolidated alluvium overlying a clay acquiclude. The ground
surface, water table and acquiclude all slope downward about 1
degree to the north, and the groundwater moves northward about
1.5 ft/day. Figure 1 shows the site location, and Figure 2 shows
the geohydrologic profile along the line of our survey.

  A shallow probe was used  to obtain soil-gas samples from a
depth of 4 ft. The probe consists of an outer pipe constructed of
3/4-'m. o.d. high-strength steel that ends in a tapered head with six
horizontal sampling ports, each 1/8  in. in diameter. A  1/8-in.
o.d. stainless-steel tube is connected  to the sampling ports and
runs through the pipe to  the sampling  manifold. The sampling
manifold is assembled from commercially available stainless-steel
fittings. Figures 3 and 4 show the probe design. Figure 5 shows
the sampling manifold.
  After insertion into the soil, the sampling manifold was purged
with soil gas.  Samples were then  withdrawn from the manifold
using gaslight syringes and were transported to a mobile labora-
  Soil-gas  samples were analyzed on-site  using an  Analytical
Instruments Development Model 511  gas chromatograph with a
3H electron-capture detector. The gas chromatograph was oper-
ated at 43 °C, and the output was processed using a Shimadzu C-
R3A integrator. Calibration standards were prepared by serial
dilutions of chloroform headspace vapors.
  Figure 6 shows the  soil-gas sampling locations relative to the
monitoring wells. Groundwater samples were taken and analyzed
in April and August of 1985. The samples were analyzed by the
purge-and-trap  gas chromatography/mass  spectrometry  (GC/
MS) method specified in the U.S. EPA Contractor Laboratory

  The bias of the method was evaluated by comparing the mean
soil-gas chloroform concentration measured to the results of the
groundwater analyses.  The soil-gas chloroform concentrations
correlate strongly with the groundwater concentrations. Table 2
lists the soil-gas and groundwater chloroform concentrations at
each well. Figure 7 is a plot of these data. A linear regression of
the soil-gas concentrations on the groundwater concentrations in-
dicates a correlation of better than 95% significance (r  =  0.85,
n = 6).' In addition, the spatial delineation of the chloroform
groundwater plume, as shown in Figure 8, is very good.
  The mean chloroform concentration from the four locations
around each well was used in the above calculations; however, the
relative standard deviation (RSD) of these mean values was often
quite high (above 100%). We interpret this result to be an effect
of spatial variability in  the soil-gas chloroform concentrations
over the  20- to  40-ft lateral distances separating these sampling
                                                                                 'x/s */",?./"•:*!	i.
                                                                                 D      vl	-...
                                      SHAFT *4130 COLO DRAWN, CHROM-MOLY. CONDITION N STEEL
                          END CAP *400 CHROM-MOLY
                           "'i r
                                                                             -I -«IO	
                                                                       INNER TUBING
                                                          Figure 4
                                                Probe Construction Specifications

                                                                                                 Table 2
                                                                             Groundwater and Soil-Gas Chloroform Concentrations
                            Figure 5
                       Sampling Manifold
                                                                                              Soil-gaa concentration (ppbv)a
                                                                              concentration  20  ft.    20 ft.    20 ft.     20 ft.
                                                                         Well     (U9/L)       West     North     East      South






                                                                                           28(5)   72.9(0.1)  124(53)c   45.6(0.2)d   67
                                                                                           (WSW)     (WNW)      (ENE)   (25 ft.  SSE)
                                                                                           266(6)   326(10)
100 ft.
E of

555 115(6) 12(5) 6(3) 27(2) 40
28 10.5(0.3) 10.5
                                                                      triplicate analyses; standard deviation in parentheses.
                                                                      Not detected; 5 ng/1 used in regression.
                                                                      ^Mean of four closely spaced points (see text).
                                                                      Duplicate determinations.
                                                                 The effect of short-range geologic variability on analytical re-
                                                              sults was assessed in  two locations. At one location,  four sam-
                                                              ples were taken at a depth of 4 ft, at separations of between 3 and
                                                              7 ft. Figure 6 shows the pattern of the sampling points. The rsd
                                                              value of the chloroform concentrations among these points was
                                                              42%. At  another location, three points 3 ft apart along a north-
                                                              south line were sampled at five depths in 1-ft increments between
                                                              1 ft and 5 ft. The mean rsd value of the mean chloroform con-
                                                              centration for all  five depths  at the  three locations was 12%.
                                                              These results indicate the short-range  geologic variability can be
                                                              a major factor in soil-gas  surveys and that the magnitude of this
                                                              factor can vary.
  The precision of the method was evaluated by analysis of mul-
tiple samples.  The analytical  precision, based on multiple daily
analyses of calibration standards, showed an rsd value below 4%.
The combined sampling/analysis precision,  based on triplicate
samples from each location sampled at a 4-ft depth, was charac-
terized by a mean rsd of 12 %.


     O Well Location
      • LGAS Probe Location
                   Figure 6
Soil-Gas Sampling Points (Sampling Depth  = 4 feet)
                                                                      S   100-
                                                              a 1
                                                              "o "H.
                                                              o -^
                                                            Log~Ground Water Chloroform Concentration
                                                                                                Figure 7
                                                                               Plot of-Chloroform-Soil-Gas Concentration and
                                                                                        Groundwater Concentration
                                                                                                                MONITORING     25

                                                   Ground-Watp r
                                                   Concent rat ion
            631     629      627      625      623     621      t, I0
                                                                                                Table 3
                                                                                    Chloroform Concentration Vertical
                    629      627     62i      62]     621      619
                           Figure 8
          Spatial Distribution of Chloroform in Soil-Gas and
                     Groundwater Samples

   The results of an evaluation of the chloroform concentration
 vertical profile (Table 3) showed a linear dependence of concen-
 tration upon sampling depth. A correlation coefficient indicating
 greater than 99% significance was obtained (r =  0.999, n  =  5).
 This result is in agreement with a model proposed  by Swallow
 and  Gschwend'  that  attributes  vertical  transport of  VOCs
 through the vadose zone to gaseous diffusion. The water table
 below the sampling locations  was at a depth of approximately

   Soil-gas surveying accurately indicated groundwater chloro-
 form contamination at a site in Pittman, Nevada. The precision
 of the technique is controlled by sampling and short-range geo-
 logic variabilities. At  the  site  studies,  results of a  depth study
 agreed along with a model for vertical transport of volatile organ-
 ic compounds through the vadose zone by gaseous diffusion.

   The authors would like to acknowledge contributions made by
 J.W. Curtis, K.L. Ekstrom, and L.J. Barrows to this study.
   Although this work was supported in part  by the U.S. EPA, it
 has not undergone review by  that agency and does not reflect
 agency policy.

 1.  Walther, E.G., LaBrecque, D.J., Weber, D.D., Evans, R.B. and van
   Ee, J.J.,  "Study  of  Subsurface Contamination  with  Geophysical
   Monitoring Methods  at  Henderson,  Nevada," Proc.  Fourth  Na-
   tional Conference on Management of Uncontrolled Hazardous Waste
   Sites, 1983, Washington, DC, 28-36.
3 ft. N
6 ft. N
3 ft. N
r, ft. N
3 ft. N
6 ft. N
3 ft. N
6 ft. N
6 ft. N
3 ft. N
Mean chloroform
(ppbv) (SOa)
23.0 (0.1)
22.9 (0.8)
19 (1)b
76 (1)
70 (2)b
58 (3)
109 (1)
111 (1)
99 (14)*>
153 (9)
149 (1)
132 (14)
205 (2)b
183 (8)
167 (23)
236 (29)
 Duplicate measurements unless noted
 triplicate measurements
2. Horowitz, L., "Geochemical Exploration for Petroleum " Science
   229,(1985), 821-827.
3. Marrin, D.L. and Thompson, G.M.,  "Investigation in the Unsatur-
   ated Zone Above TCE Polluted Groundwater."  U.S.  EPA, Ada,
   OK, U.S. EPA Project Number CR811018-01-0.
4. Voorhees, K.J., Hickey,  J.C. and Klusman,  R.W., "Analysis  of
   Groundwater Contamination by a New Surface Static Trapping/Mass
   Spectrometry Technique," AnalyticalChem.. 56. 1984, 2604-2607.
5. LaBrecque, D.J., Pieretl, S.L., Baker, A.T., Scholl, J.F. and Hess,
   J.W.,  "Hydrocarbon  Plume Detection  at Stovepipe Wells, Cali-
   fornia," U.S. EPA, Las Vegas, NV, 1985.
6. McGhee. J.W., Introductory Statistics. West Publishing New York
   NY, 1985.

7. U.S. EPA, "Chemical Analytical Services for  Multi-Media Multi-
   Concentration Organics  GC/MS  Techniques," WA-85J680 US
   EPA, Washington, DC, 1985.

8. Swallow, J.A.  and  Gschwend, P.M.,  "Volatilization  of  Organic
   Compounds from Unconfined Aquifers," Proc.  National Symposium
   on Aquifer  Restoration  and  Ground water Monitoring^ National
   Water Well Association, 1983, 327-333.

                         Environmental  Appraisals and Audits:
                            A Case  Study of Their  Application

                                               Anthony R. Morrell
                                          U.S. Department of Energy
                                       Bonneville Power Administration
                                                Portland, Oregon
  The Bonneville Power Administration (BPA) has had an envir-
onmental audit and appraisal program for 2 years. BPA's unusual
approach in implementing its appraisal program resulted in the
program's quick acceptance and organizational effectiveness.
  The first audits conducted under the appraisal program iden-
tified problems  in the use of polychlorinated biphenyl (PCB)
equipment at BPA substations. Subsequent U.S. EPA inspections
confirmed these same findings. In response to this situation, the
U.S. EPA and BPA entered into a Memorandum of Agreement
(MOA) specifying corrective actions to be taken.
  The U.S. EPA, recognizing the strength of BPA's appraisal
program, agreed to  forego their normal ad hoc  inspections of
BPA facilities. Instead, the U.S. EPA deferred to BPA's own
audits with U.S. EPA inspections limited to followup or verifica-
tion inspections.
  In addition to ensuring compliance with environmental regula-
tions, the Agreement's reliance upon BPA's appraisal program
maximizes the effective use of limited staff resources in both
agencies. The Agreement interjects a  large measure of predic-
tability and  manageability into  BPA  efforts to achieve  com-

  As a power marketing agency within the U.S. Department of
Energy (DOE), the Bonneville Power Administration (BPA) sells
and transmits the electrical output from 30 federal hydroelectric
dams in the Columbia River Basin. The BPA service area includes
300,000  square miles primarily in the  states of Oregon, Wash-
ington, Idaho and Montana, with small service areas in Califor-
nia, Nevada, Utah and Wyoming. The  BPA transmission system
serves as the backbone  for the  interconnected utilities in the
Pacific Northwest and is connected with 18 other transmission
systems at over 100 locations. The BPA transmission system con-
sists of approximately 14,200 circuit miles of  high  voltage
transmission lines and about 400 substations. Approximately 100
of the BPA substations contain polychlorinated biphenyl (PCB)
equipment, primarily PCB capacitors. A total of about 140,000
PCB capacitors are in service at these substations.
  In 1983 BPA instituted an environmental appraisal  and  audit
program. Upon adoption of the program, BPA decided as a mat-
ter  of  policy to  use  its appraisal and  audit program  as  a
mechanism to ensure that BPA facilities comply with applicable
environmental standards.

  Equal to the importance of its adoption is the special approach
BPA used in implementing its new program. BPA is convinced
that, given its own organizational culture, any other implementa-
tion  approach  would not have been as  readily accepted and
therefore would have been much less successful in achieving the
objectives of the program.

   BPA appraisal program objectives include the following:

 • Assure that  DOE environmental policy and requirements are
   appropriately interpreted and implemented by BPA and BPA
 • Help line managers achieve BPA's compliance commitments.
 • Increase employees' awareness  of environmental regulations
   and BPA's commitment to compliance.
 • Provide management with objective, timely and reliable in-
   formation on  BPA and BPA-contractor  environmental per-
   formance, including significant achievements and efficiencies.
 • Evaluate the effectiveness and efficiency of BPA's implemen-
   tation of measures to avoid, minimize,  rectify or otherwise re-
   duce adverse impacts to the environment and measures to com-
   pensate for impacts.
 • Provide management with recommendations for improvement
   of BPA's  environmental program performance.
 • Develop and recommend long-term solutions to current  en-
   vironmental  problems in anticipation of future standards or
 • Evaluate the accuracy  of  environmental  analysis or  impact
   predictions and identify methods for improvement.

   These objectives are relatively standard. What distinguishes the
 BPA program  is the unusual approach taken in its implementa-
 tion. Without  compromising the objectives of the program, the
 approach taken was a positive one designed to find the facilities
 audited in compliance. As a practical matter, this simply meant
 that every effort was made  to communicate, beforehand, what
 was required and what would be audited. This was accomplished
 by providing advance copies of checklists to be used during the
 audit and by meeting with the affected organizations prior to their
 actual audits.  This approach was very well received by the af-
 fected field organizations. Given BPA's functional organization
 and culture, any other approach would have met with resistance.
 Other factors which ensured successful implementation were pro-
 visions to: (1) resolve problems at the lowest organizational level;
 (2) include representatives of affected offices on each appraisal
 team;  and (3) resolve smaller problems in the field (i.e., labeling
 equipment), so that they need not be mentioned in the written
 audit reports. Taken together, these provisions were viewed as a
 sincere attempt by central headquarters  staff to assist the field
 facilities to come into compliance without unnecessarily embar-
 rassing them organizationally.
                                                                                                    MONITORING    27

         PHASE 1
     PHASE 3
                              Figure 1
                 BPA Environmental Appraisal Process
   A useful strategy in the initial development  of the Environ-
 mental  Appraisal Program was to seek management consensus
 regarding both the scope of the program and the priorities of sites
 to be appraised.
   To begin the program, meetings were  held with  all affected
 managers. The objectives and  purposes of the appraisal program
 were discussed.  As a result of these discussions, it was agreed that
 BPA's  appraisals  initially  would  focus  on  operations  and
 maintenance activities.
   It was also agreed that the first year's program should examine
 the requirements of those environmental  laws and  regulations
 dealing  with  oil  spill prevention,  hazardous and toxic waste
 management, noise  pollution control, safe drinking water protec-
 tion and  commitments made  to the  public through documents
 prepared to comply with the National Environmental Policy Act.
   After a schedule was developed for the first year and appraisal
 teams were identified, training  was conducted for those  who
 would  participate  in  the program.   BPA  sought  contractor
 assistance in conducting  the initial training and obtained the ser-
 vices of a nationally recognized firm that had published informa-
 tion in the area of environmental auditing.
   Through the  training,  the teams  obtained  skills  regarding the
 use of checklists and how to document audit findings in a report.
 Also, environmental  staff  members  contacted  other  electric
 utilities who had completed environmental  audits or appraisals in-
 order to  utilize  their  experience in  performing environmental
 audits on facilities associated with the electric utility  industry.
   In the program's  first  2 years, 23 field appraisals were  con-
 ducted.  Each of these appraisals required 1 -2 days of field work at
 the  site  followed by the  preparation of  a written  report;  the
 equivalent of approximately four full-time positions have been re-
 quired to  implement the  appraisal program.
   Figure 1 illustrates the basic steps used  in conducting  the ap-
 praisal process.

  The BPA appraisal program contains three components. Each
component has  a different focus (site versus program)  or em-
                             Figure 2
           BPA Audit/Appraisal Program TSCA Checklist

                Polychlorlniled Blphenyli (PCBi) In Service

Definitions:  PCBs in service al BPA facilities may  fall into any of the following
categories: (I) non-PCB equipment, where the level  of PCBs is below 50 ppm; (2)
PCB-contaminaled equipment, levels of JO to 499 ppm PCBs; and (3) PCB equip-
ment with levels above 500 ppm. This checklist addresses all PCB articles in service,
whatever the distinction in PCB levels. If appropriate, remarks regarding the levels
of PCBs should be made in the Remarks Column or on the back of the page.

Nole 1: Definition of Transformer: All transformers (e.g., power transformers,
potential transformers, current transformers) are defined as any piece of equipment
which contains the description "transformer" on the manufacturer's nameplate.
  a)  The transformers that have no PCB ppm  tag, and which have the word "oil"
     marked on the manufacturer's nameplate, arc assumed to be 50-499 ppm PCB
     until these are lagged with the actual PCB ppm level.
  b)  The transformers that have no PCB ppm tag.  which have markings such as
     "liquid-filled" or "contains dielectric" (other than the word "oil") or do not
     give any indication of the content of the transformer, are assumed to contain
     more than 500 ppm PCB until these are tagged with the actual ppm level.
                                                                         Yea  No
                                                                         Yes  No
                                                                         Yes  No
                                                       0.  Does the facility have any equipment which contains any levels
                                                       of PCBs (cither in service or in storage)?
                                                       I   PCB-Conlaining Equipment (Transformers, Circuit Breakers.
                                                          Reactors. Reclosers. etc.)
                                                          1.1  Has this facility ever conducted an inventory of its equip-
                                                             ment \thich contains PCBi (or an inventory of all equipment
                                                             including pieces containing PCBs)?
                                                             I.I.I If yes, when was the inventory conducted? (dale of
                                                             1.1.2 If yes, is the inventory complete? (Check each piece of
                                                                 equipment to see if it corresponds to the inventory.)
                                                             1.1.3 Describe each type of RGB-containing equipment,
                                                                 quantity, status. PCB concentrations and volumes (or
                                                                 include inventory list as working paper):
                                                                                     PCB Cone.    PCB Volume
                                            Yes  No
                                            Yes  No
                                                         1.2  Are all pieces of equipment marked with an appropriate
                                                             PCB ppm level tag and or the appropriate PCB label (ML)?
                                                             1.2.1  If no. list the ones that are not marked.
                                                         1.3  Are any pieces of equipment leaking? (Identify)
                                                             1.3 I  What action is being taken to control these leaks?
                                            Note 1: Action to contain and cleanup leaks in 500+ ppm PCB Transformers must
                                            be initiated within 48 hours of having knowledge of the leak. All leaks from any
                                            (ransformer or other piece of equipment should be contained and cleaned as soon as
                                           Yes  No
                                           Yes  No
                                           Yes  No
                                           Yrs  No
                                      1.4  Are all pieces of equipment inspected?
                                          1.4.1  At » hat frequency?	
                                      1.5  Are inspection records kept?

                                           Yes  No
                 I.S.I Arc operators' inspection records maintained for three
                     (3) years after equipment disposal?
                     Where? . _		
             1.6  Arc pieces of equipment stored for use (spares) listed on the
                 I .ft. 1 Arc these spares inspected regularly as described above?
             1.7  Have any pieces of equipment been serviced or repaired
                 (rebuilt) at the facility?
                 1.7.1 Arc maintenance records for these pieces of equipment
                     kept? Where	

Figure 2 (continued)

Yes  No       1.8  Does the facility ever top-off or refill any existing pieces of
                  1.8.1 What materials are used? (include PCB ppm levels)
Note 3: No new oil will be procured with over 2 ppm PCB level.
Yes  No

Yes  No

Yes  No

Yes  No

  1.9 Have any pieces of equipment been removed from service
      at this facility?
      1.9.1 Are records kept on the disposition of these pieces of
           equipment? Where?	
2. Capacitors
  2.1 Has this facility ever conducted an inventory of capacitors?
      2.1.1 If yes,  when was the inventory conducted? (date  of
      2.1.2 If yes, is the inventory complete?
      2.1.3 Describe each type (or group) of capacitor, quantity,
           status, PCB concentrations and volumes (or include
           inventory list as working paper):
                                          PCB Cone.    PCB Volume
 Yes  No       2.2  Are all capacitors marked with an appropriate PCB label
                       (ML) or in an appropriately marked area?
                  2.2.1  If no, list the ones that are not marked.
Yes  No
Yes  No
Yes  No
Yes  No

Yes  No

Yes  No

Yes  No
Yes  No

Yes  No
Yes  No
Yes  No
                                                                         Yes  No
                                                                           2.4 Is there a designated area for capacitors stored for use?
                                                                               2.4.1 Is this area appropriately marked?
                                                                           2.5 Are all capacitors inspected?
                                                                               2.5.1 At what frequency?	
                                                                                       2.6  Where are inspection records kept?_
      2.6.1 Are operators' inspection records maintained for three
           (3) years after equipment disposal?
  2.7 Are capacitors stored for use (spares) listed on the
      2.7.1 Are these spares inspected regularly as described above?
  2.8 Have any capacitors been removed from service at this
      2.8.1 Are records kept on the disposition of these capacitors?

3. General Procedures
  3.1 Has or is a PCB analysis done before any oil handling?
  3.2 How are solvents, filters or rags disposed of if used on
                 equipment of unknown PCB level?
                                                                                          Does the facility have the SPIFs on handling, inspecting,
                                                                                          storing PCBs and PCB equipment?
                                                                                          3.3.1 Where are they kept?	
                                                                                           3.3.2 List the SPIFS found at the facility:
                                                              Yes  No
           4. Other Equipment Containing PCBs (Light Ballasts, Relays, etc.)
             4.1  Does the facility have any other equipment containing
                 PCBs? List:	
 Yes  No
              2.3  Are any capacitors leaking?
                  2.3.1  How many? 	
                  2.3.2 What action is being taken to control these leaks?
 Note 4: Action to contain and cleanup leaks in PCB capacitors must be initiated
 within 48 hours of having knowledge of the leak.
                                                              Yes  No      4.2 Has any of this other equipment leaked?
                                                              Yes  No      4.3 Was cleanup initiated?
                                                                           4.4 Describe how the material was contained, cleaned up and
                                                                               disposed of:  	
                                                              Yes  No      4.5  Has any of this other equipment been removed from service,
                                                                               but retained for backup at this facility?
phasis and each  is the responsibility of a different organization
(field personnel versus central staff). These three components are
listed below.

Functional Appraisal
   The functional appraisal is an evaluation of the agency's overall
performance based upon a compilation of findings obtained from
the individual  field appraisals. This appraisal is  conducted an-
nually by BPA's Environmental Manager's office. The functional
appraisal focuses its attention on major or  repetitive problems
identified  in the field appraisals. In addition to apprising  top
management of its findings,  the functional appraisal is used as a
means for securing management support for  implementing those
recommendations that require an organizational commitment of

Field Appraisal
   The field appraisal is a documented, on-site appraisal of pro-
gram effectiveness for specific disciplines, namely  hazardous and
toxic waste management,  oil spill  containment,  noise and safe
                                                             drinking water regulations.  Following the field appraisal, a writ-
                                                             ten action plan is prepared. The action plan outlines an agreed
                                                             upon course of action to implement recommendations designed to
                                                             correct problems identified in the field appraisal. The action plan
                                                             is prepared within 60 days after completion of the field appraisal.

                                                             Internal Audit
                                                                A self-audit is a less formal evaluation, conducted by the  line
                                                             organization on a particular part of its environmental program.
                                                             These audits are scheduled  on an as needed basis.
                                                                The three types of appraisals described above permit appraisals
                                                             to be specific to a particular audience or level within the organiza-
                                                             tion and to focus upon issues of greatest concern to that level.

                                                             MEMORANDUM OF AGREEMENT
                                                             WITH THE U.S. EPA
                                                                In February 1985, the BPA and the  U.S. EPA entered into a
                                                             Memorandum of Agreement (MOA)  to clarify each  agency's
                                                             responsibilities and commitments for conducting actions required
                                                             and authorized by TSCA and CERCLA.

   Among other things, this Agreement specifically provides that
 BPA will audit its field installations for compliance with TSCA
 PCB regulations. The Agreement requires that  these audits be
 done on a scheduled basis and that they be administered by cen-
 tral headquarters staff. Any deviations from the PCB regulations
 that are found in the conduct  of these  audits are  to be
 documented, and a compliance plan is to be included in the audit
 report. All reports are forwarded to the U.S. EPA. A copy of the
 PCB audit checklist, which  reflects U.S. EPA Region  10's inter-
 pretation of the TSCA regulations, is included in Figure 2.
   With the Agreement in place, the U.S. EPA agreed to forego
 ad hoc inspections of BPA  facilities. Instead, U.S. EPA inspec-
 tions are limited to verification inspections to be  conducted only
 after the facilities have been audited by BPA.
  The utility of BPA's environmental appraisal and audit pro-
gram was evident even before the Agreement with the U.S. EPA,
but its usefulness to the agency was certainly increased when the
U.S. EPA agreed to use the appraisal program as a substitute for
its ad hoc inspections of BPA facilities.
  From the U.S. EPA standpoint, this ensured not  only that
potential violations would be corrected, but that the underlying
causes would be corrected as well. From BPA's standpoint, rather
than reacting to the U.S. EPA's inspections, this  arrangement
resulted in an opportunity to prioritize corrective actions at BPA
facilities based on environmental and operational considerations.
Consequently, a large measure of predictability and manageabil-
ity was introduced into what could have been a counterproductive
and possibly adversarial relationship with the U.S. EPA.

              Chlorinated  Organics and Hydrochloride  Emissions
              Sampling  from  a Municipal  Solid Waste  Incinerator
                                               Thomas A- Driscoll
                                                  James P. Barta
                                                 Henry J. Krauss
                                              David H. Carmichael
                                                 J. Maxine Jenks
                                            Texas Air Control Board
                                               Waxahachie, Texas
  As the use of solid waste incinerators in Texas increases, so do
the questions regarding the environmental impact of their emis-
sions. Are the incinerators creating or releasing potentially toxic
materials from the waste?
  Currently, there are eight municipal incinerators operating in
Texas, and permit applications are pending for several more. The
purpose of this paper is to report the levels of chlorinated organic,
chlorine and chloride measured during the stack  sampling con-
ducted at the City of Waxahachie municipal waste incinerator.
The sampling was performed in August 1985, by  staff from the
Texas Air Control Board (TACB) Sampling and  Analysis  Divi-
sion. Three sets of 3-hour samples were collected.
  The results of this sampling will be reviewed to determine
whether or not  additional preventive measures are needed to
reduce the emission of toxic chlorinated organic compounds or
hydrochloric acid. The hydrochloric acid emissions generally are
thought to result from the burning of plastics in the refuse.1 The
chlorinated organic compounds (i.e., polychlorinated biphenyls
(PCBs), polychlorinated dibenzodioxins (PCDDs) and polychlor-
inated dibenzofurans (PCDFs) may originate as contaminants in
the refuse or as byproducts of the combustion of precursors in the
incinerator.2 In addition, these data will be used by the Permits
Division of the TACB to develop emission factors to be used in
permitting new incinerators.
  The stack sampling procedure was a modified version of U.S.
EPA Reference Method Five. The sample is collected  by XAD-2,
florisil, alkaline arsenite solution and glass fiber filter media.

  At this time, analysis of all the samples collected is not  com-
plete. Results of the hydrochloric acid testing show the emissions
to be on the order of 2 Ib/hr. A preliminary list of compounds
detected (but not quantified) is presented in this paper. The com-
pounds that were detected are listed below with some information
about their  use, toxicity and exposure limits in air:
• Dichlorobenzene—used as a process solvent and  as an in-
  termediate in the synthesis of dyestuffs and herbicides;3 causes
  hemolytic anemia and liver necrosis in humans;  can also cause
  irritation  of eyes and nose; the OSHA occupational exposure
  limit (OEL) has been set at 50 ppm upper limit.4 The TACB
  Health Effects staff uses  1 % of the OEL as a guideline for per-
  mit and health effects review.
• Hepta and hexachlorodibenzofurans—similar properties  and
  health effects to dioxins of similar molecular weight. The On-
  tario Ministry of Health has set 1 x  10~3 /*g/m3 as a standard.
• Hepta and hexachlorobiphenyls—used in electricity transmis-
  sion transformers; can cause cancer and adverse skin, liver and
  reproductive health effects. The NIOSH standard is 1 jtg/m3.
• Hexachlorobenzene—used as an herbicide, wood preservative
  and also is a byproduct from chlorinated hydrocarbon produc-
  tion;5 can cause death in breast-fed infants, skin  sores, skin
  discoloration and enzyme disruptions.' There are no standards
  set for this pollutant, however, the TACB Health Effects staff
  uses  1 % of the estimated OEL of < 1 ppm as a guideline for
  permit and health effects review.
• Pentachlorobenzene—used as a precursor to fungicide produc-
  tion and as  a flame retardant; also occurs as a product of the
  degradation of Lindane;7 may cause mutagenic and carcino-
  genic effects in mice.' There are no air standards set for this
  compound,  but 1 % of the estimated OEL is used as a guideline
  by the TACB.
• Tetrachlorodlbenzo-p-dioxin—occurs as a byproduct of herbi-
  cide production;' causes liver cirrhosis, spontaneous abortions,
  kidney disease and chloracne.10 The New York Air Pollution
  Control has set a standard of 9.2 x 10-' jtg/m3.
• Trichloroethane—used as a degreaser, "cold" cleaner and dry-
  cleaning agent;" acts as  a narcotic to depress the central ner-
  vous system; can cause dizziness, uncoordination,  drowsiness
  and death systematically. Locally,  it is irritating to eyes and
  causes dermatitis.12 Air  limits include TLV of 10 ppm. The
  short-term exposure limit (STEL) value is 20 ppm (tentative).13
  The OSHA  standard  is 10 ppm.'4
    The compounds detected in the analyses of these samples are
  consistent with compounds  measured  in stack sampling con-
  ducted in other areas.15'16  Stack sampling conducted by the
  state of California and  Scott Environmental (contracted by
  EPA) indicates the small amounts of PCDDs and PCDFs were
  detected in municipal solid waste incinerator emissions.

  The municipal solid waste incinerator stack sampling was con-
ducted from Aug. 13 to 15, 1985. Three separate sets of samples
were collected over a 3-hr period each time. A representative por-
tion of the stack emissions was collected isokinetically using stan-
dard operating procedures described in the TACB Sampling Pro-
cedures Manual.
  The incinerator sampled is located on Singleton Drive in North
Waxahachie, Texas. It  has two dual-chambered incinerators that
operate at approximately 1600°F. The incinerators are permitted
to destroy 2088 Ib/hr of solid waste. The residence time is approx-
imately 45 min to 1 hr in the primary  burn chamber and 1.22 sec
                                                                                        DETECTION OF RELEASES    31

                                                                                      TC  THERMOCOUPLE
                                                          Figure I
                                              Chlorinated Organic*, Sampling Train
for the secondary burn chamber which is often used to burn efflu-
ent gases from the first chamber.
  The California Air Resources Board  recommends  a flue gas
temperature of 1800°F + / -  190° with a residence time of 1 sec
under  well-mixed conditions.  These conditions should destroy
99.9% of the toxic organics compounds."
  The facility is operated 24 hr/day, 6 days/week. There is down
time for maintenance and cleaning each day. The  facility is per-
mitted to emit 12 Ib/hr or 53 tons/yr of particulate matter.

  The sampling train used to collect chlorinated organic  com-
pounds is shown in Figure 1. The sampling apparatus is made en-
tirely of glass except  for the stainless steel nozzle and teflon tape
seals used between the ground glass joints. Following the stan-
dard front-half, the back-half is modified to consist of a water
jacketed condenser followed by two XAD-2 and one Florisil resin
traps in series. The remainder of the train consists of two im-
pingers. The first impinger contains 200 ml of sodium arsenite to
collect chlorides/chlorine and to protect equipment downstream.
The  second impinger contains silica gel to insure the complete
dryness of the sampled gas. The Organic Laboratory Section per-
formed the analyses of  the samples using the Finnigan triple
quadrupole  mass spectrometer (TSQ).
  The hydrochloric acid emissions were captured using the collec-
tion system as described in Chapter 5 of the TACB Sampling Pro-
cedures Manual. The stack  exhaust gas  entered  the  system
isokinetically through a stainless steel probe and traversed a non-
heated probe across a glass fiber filter which was  maintained al
250 to 275 °F. The condenser  train  consisted of a series of im-
pingers containing deionized  water, alkaline arsenite and  silica
gel. The analyses followed the procedures described in the TACB
Laboratory Procedures Manual.
  Leak checks were performed several times during the process to
insure  sample integrity. If acceptable, the  sampling process was
continued.  The sample air was circulated through a condenser
filled with cold water. Due to the large volume (approximately
100 ft3) and  the high  moisture  content of the air drawn from the
stack,  a large  volume of condensate was  collected in the flask
following  the  condenser. As necessary,  the  condensate was
removed from the flask, the volume recorded, transferred to an
amber glass container and returned to the TACB laboratory for
analysis. A check for air leakage in the sampling system preceded
and followed the condensate transfer to insure sample integrity.
  Following the completion of sampling, the three  adsorption
tubes and impingers  were weighed again  and their final  weights
were recorded to determine stack air moisture content. The probe
and glassware in front of the filter were rinsed  with  tetrahydro-
furan (THF). All  rinsate was collected  in  amber bottles.  The
amber bottles were used to prevent rinse solutions from reacting
with ultraviolet  light. The glass fiber filter was returned to its pro-
tective alumi.ium  foil  and  stored   in  plastic  envelopes  until
analysis. The connecting glassware and two impingers  were rinsed
with deionized water and the rinsate was collected for analysis at
the TACB laboratory.

  The XAD-2 and Florisil adsorbers were extracted using  hexane
in a Soxhlet extractor for 8 hr.  The distillate then was  evaporated
down to 2 or 3 ml using a rotary evaporator.  The product  was in-
jected into a gas chromatogram/mass spectrometer (GC/MS).
The pollutant results were interpreted by a  TACB chemist  who
used graphic responses compared to a chemical library to identify
the pollutants.
  The front-half and  back-half rinsates were  concentrated by
evaporation. Separately, the products of the evaporation were in-
jected into the GC/MS. The results were interpreted as discussed
in the previous  paragraph.
  The impinger containing deionized water was used to collect
hydrochloric acid. The samples were analyzed turbidimetrically.
Chlorine was collected in the arsenite solution. The solution also
was analyzed turbidimetrically.
  The glass fiber filter also was analyzed for chlorides and chlor-
inated organics. A 37 mm diameter circle was cut from each of the
three filters. The exposed filters  were analyzed versus the field
blank for chlorine on an x-ray fluorescence spectrometer system.
Duplicates also were run to test the method's precision. In  addi-
tion, the filters were  analyzed for chlorinated organics. They were
extracted separately  using hexane and a Soxhlet extractor. The
resulting solution then was reduced  to 2 or 3 ml with nitrogen.
The product was injected into  the TSQ and analyzed  as described

  More stack sampling of municipal solid waste incinerators in
Texas is needed. During this initial stack sampling exercise, there
were some problems with the plant operation and with the stack
monitoring process due to  malfunctions  in  the  incinerator.
Preliminary determinations of the hydrochloric acid test  have
been made. The emissions of hydrochloric acid are approximately
1 Ib/hr. This value is lower than expected.
  Particulate matter collected during the test was analyzed and
found to contain significant amounts of chloride salts. It is be-
lieved  that the chlorides  exiting this  process  were primarily
chloride salts with relatively low hydrochloric acid emissions. Fur-
ther testing is required to determine if this is the case.
  Also, some of the pollutants detected in preliminary  analysis
are very toxic  and warrant additional attention. PCDDs and
PCDFs have been described by  some  sources as unsafe at any
level. This project has not produced  enough  information  to
prescribe  incinerating  at  higher temperatures, having  longer
residence times in the burn chamber, requiring additional pollu-
tion control devices or sorting out plastics before incineration.
  Whether or  not the Waxahachie municipal  solid waste in-
cinerator  is representative of incinerators in Texas also is not
known. Stack sampling at other sites may be needed to adequately
characterize  incinerator emissions.  Other types  of air  quality
monitoring also may  be warranted.  Perhaps  monitors placed
downwind may help correlate stack emissions levels to levels near-
by residents may eventually breathe. Therefore, developing con-
clusive emission factors is premature.

  The authors would like to acknowledge the Texas Air Control
Board Laboratory, Dallas Regional personnel, and  S. Thomas
Dydek of The Health Effects Section for their assistance with this

  1. Krakower, T.,  Resource Recovery Facilities Overview, Texas Air
    Control Board (Health Effects and Research Division), Sept. 1984,
    Section III.
  2. Ibid., Summary.
  3. Sittig, Handbook of Toxic and Hazardous Chemicals,  Noyes Publi-
    cations, Park Ridge, NJ, 1981, 227-228.
  4. Ibid., 228.
  5. Ibid., 358.
  6. Ibid., 359.
  7. Ibid., 522.
  8. Ibid., 522.
  9. Ibid., 632-633.
 10. Ibid., 632.
 11. Ibid., 669, 671.
 12. Ibid., 670-671.
 13. Ibid., 669, 671.
 14. Ibid., 669.
 15. Sheffield, A., "Sources and Releases of PCDDs and PCDFs to the
    Canadian Environment," Chemosphere, 14,  1985, 811-812.
 16. Nunn,  A.B., III, "Gaseous HC1 and Chlorinated Organic Com-
    pound  Emissions from  Refuse Fired  Waste-to-Energy Systems,"
    Scott Environmental Services (for U.S. EPA Environmental Sciences
    Research Laboratory), Plumsteadville, PA, 1984, 30, 37.
                                                                                                 DETECTION OF RELEASES    33

                CARE—Modeling Hazardous Airborne  Releases

                                                M. Gary  Verholek
                                     Environmental Systems Corporation
                                              Knoxville,  Tennessee
  Emergency preparedness for airborne hazards means being able
to answer such questions as where the material will go, who it will
effect, what areas are safe as evacuation centers, how  long the
chemical will take to disperse and which preplanned response op-
tion would be most effective in the event of an incident. To help
answer such questions,  Environmental  Systems Corporation
(ESC)  has  developed  an assemblage  of  sophisticated
mathematical models into a user-friendly and truly capable com-
puter system called the CARE (Computerized Airborne Release
Evaluation) System.
  CARE  provides the capability to perform a variety  of func-
tions,  such as risk  assessment, hazardous materials  training,
monitoring and emergency response.  In short, CARE  provides
the  tools needed  to prevent an emergency from becoming  a

  Are you ready for a hazardous materials accident? What would
you do if a toxic cloud from a spill  or a ruptured tank threatened
your facility? What if burning chemicals were threatening your
town today? Where would the cloud go? Who would be effected
and who  would be out of danger?  What areas would need to be
evacuated? Where could people  go to be safe? What is the best
way to respond to the emergency? How do you  get answers to
these questions in time to prevent the emergency from becoming a
  To answer such questions. Environmental Systems Corpora-
tion (ESC) has developed a system called CARE, which stands for
Computerized Airborne  Release Evaluation System. CARE gives
you the capability to perform a variety of emergency preparedness
• Risk Assessment
• HazMat Training
• Real-Time Monitoring
• Emergency Response
  CARE   is a  system   that puts  truly  capable  emergency
preparedness on  your desk. In the event of an accident  that
releases a hazardous material into the air, CARE will tell you
where it will go, who it will effect, which areas are safe as evacua-
tion centers, how long  it  will take to disperse and which  pre-
planned response options would be most effective. CARE gives
you the ability to neutralize the  danger, to assess  alternative
defense strategies and to keep abreast of changing situations. In
short,  CARE gives you  the tools you need to prevent  an emer-
gency from becoming a disaster.
Computerized Airborne Release Evaluation System
       SOURCE Mode!
                                      METEOROLOGY Model
      Calculates source
Calculates winds
stability, lemp
over modeling region
                       DISPERSION Model
                       Calcinates pollulani
                       conce^ra'.o^ 'te'ds
                        EFFECTS Model
                       Describes lie Hazards
                       due to ambient
                       RESPONSE Model
                      Provides protective
                      action guidance
                          Figure 1
                   CARE System Diagram
   CARE uses sophisticated mathematical models, assembled by
 ESC's meteorologists and engineers, to provide a realistic depic-
 tion of the hazard due to a  cloud of airborne material. These
 models are grouped into five  modules (Fig. 1):
 •  Source Module
 •  Meteorological Module
 •  Dispersion Module
 •  Effects Module
 •  Response Module

  The Source Module provides one of the two basic mathematical
models required to estimate atmospheric concentrations of air-
borne materials. For chemical spills, it calculates the evaporation
rate of the chemical into the atmosphere. This rate of evapora-
tion, determined through heat and/or mass transfer mechanisms,
it used as the emission rate in the second mathematical model, the
air  dispersion model. The final product is then the calculation of
the  geographical  distribution  of  pollutant  concentration at
selected times  since the spill,  via the atmospheric dispersion
model. Specific source modules are available for various applica-
tions such as chemical spills, fires and transportation accidents.
nuclear plant accidents, fossil plant emissions,  airborne pesticide
releases, military smoke generators, etc.
  The Meteorological Module is a group of several models that,
in combination, characterize the  weather conditions that effect
the dispersion of hazardous materials. This module uses observed
meteorological  data, such  as wind speed,  wind direction, tur-
bulence and temperature and provides the necessary inputs to the
dispersion module.
  The Dispersion Module,  using the source and weather inputs,
provides a realistic depiction of the movement of the  cloud of
hazardous material and calculates its concentration with respect
to time and distance from the release. The model, which is the key
to systems realism, is a variable-trajectory, puff-advection model.
This kind of model treats the airborne release as a series of puffs
which are moved by the air flow that has been calculated by the
meteorological  module. The puffs are diffused along the way ac-
cording to the character of the atmosphere. The resultant disper-
sion of the material is displayed on the computer as a meandering
plume which flows over and around obstacles, much as one would
expect a plume to move once released to the air.
  The Effects Module assesses the critical  concentration values
and provides information on the hazard to health and welfare
posed by the materials in the plume. This module examines the
concentration fields for critical levels based on the application.
For example, the  effects module for chemical accidents deter-
mines exceedances of Threshold  Limiting Values or Lower Ex-
plosive Levels;  for  nuclear releases,  it  calculates dose and dose
rate;  for smoke obscurant  releases, it calculates transmission or
visibility; for ambient air quality, it determines exceedances of the
National Ambient Air  Quality Standards; and  so forth.
  The Response Module provides the user with recommended re-
sponse actions  based on the effects assessment. The  response
module is tailored by ESC to meet user requirements, special con-
ditions and capabilities.
  The CARE system assembles the aforementioned capabilities
into a user-friendly system that is fully documented and tested.
The near real-time display  of the location and concentration of
hazardous airborne materials provides the information needed to
make informed decisions that can save  lives and property.

  The Source Module for toxic releases is based on the SPILLS
model developed by Shell Development Company.' SPILLS is an
unsteady-state  model  which calculates the  evaporation  of a
chemical spill to determine the source strength of the vapor cloud.
Used in conjunction with the atmospheric dispersion model, it is
used to estimate concentrations of the vapors as a function of
time and distance downwind of the spill.
  Three options depending on the nature of the spill have been in-
corporated in the model: (1) continuous spills, such as leaks from
tank cars, tanks or pipelines; (2) instantaneously-formed pools of
liquids or liquified gases; and (3) stacks, where the emission rate is
assumed to be known.  For options 1  and 2, thermophysical pro-
perties (available as a subroutine in the computer program) of 36
potentially hazardous chemicals are  used  to calculate, through
heat and mass transfer mechanisms, the evaporation rate, which
becomes the emission rate for the atmospheric dispersion calcula-
  The computer program was  adapted for use with complex at-
mospheric dispersion models to provide a  realistic prediction of
hazard caused by an airborne toxic cloud. The program now con-
tains  the necessary  properties of  36 potentially  hazardous
materials commonly handled,  used or transported by industry,
but the data easily can be expanded to include the thermophysical
properties of other substances.  The air dispersion model takes in-
to account the atmospheric conditions, type of source and emis-
sion rate  to calculate  downwind  concentrations. The  source
strength can be specified by the user (stacks) or is determined by
the evaporation rate models  if the  chemical is  known. At-
mospheric and soil conditions and chemical properties are utilized
to predict the effects of heat and mass transfer on the vapor cloud

Evaporation Rate Models
  The objective of an evaporation rate model is to predict the
amount of material emitted in the case of a continuous or an in-
stantaneously formed spill from a transport line or a storage tank.
Although the basic theory is the same for both cases, the calcula-
tions are different, and thus they are  presented separately.
  Three evaporation rate models are discussed in this section.
First, the continuous  leak case which corresponds to emissions
generated  from small ruptures  in transport equipment  (e.g.,
pipelines, storage tanks or tank cars).  In this case, the chemical is
assumed to flow at atmospheric conditions through the rupture to
form a quiescent pool on the ground. Since the chemical will be at
ambient temperature, no heat transfer will occur between the
material spilled and the surroundings, and the amount of mass
taken by the wind blowing over the pool will be limited by mass
  Two key  variables  involved in the mass transfer model are
simultaneously unknown: (1) the area of the spill and (2) the emis-
sion rate into the atmosphere. It was assumed that the area of the
pool would be the more difficult variable to predict during the ac-
cidental spill.  Therefore, the  computer program  requests the
value of the emission rate through the rupture,  which the  model
takes  as the emission rate into the atmosphere. The model then
determines the spill area by assuming that this emission occurs by
convective mass transfer. The evaporation then is assumed to occur
indefinitely at  the same rate unless the user specified the elapsed
time from the beginning to the end of the flow through the rupture.
At this point, the program would set the emission rate to zero.
  Evaporation from instantaneously  formed pools, the second
option in SPILLS, corresponds to the spill of a chemical from a
storage tank or transport line. In this case, however, the spill and
the resultant pool at  the ground are  assumed  to occur instan-
taneously. The area of the  pool is assumed to remain constant
during the evaporation process so that its estimation at the scene
of the accident would be easier than the continuous spill case. The
computer program then requests the value of this spill area in ad-
dition to the total amount spilled.
  Chemicals  with  a  normal  boiling  point   below  ambient
temperature will first flash off due to the pressure drop between
the storage pressure (the storage temperature has to be specified
by the user) and atmospheric pressure. The evaporation rate due
to this adiabatic flash calculation is assumed to occur during the
first minute after the spill. The chemical will then form a pool of
liquified gas at its normal boiling point. The difference between
this temperature and the ambient temperature will cause heat to
be transferred from the ground (the soil is  assumed to be  at am-
                                                                                             DETECTION OF RELEASES    35

bient temperature) to the pool by conduction  and from the at-
mosphere to the pool by convection.
   Mass transfer, due to the wind blowing over the pool,  takes
over when the heat transfer evaporation rate becomes equal to the
mass transfer evaporation rate. Note that this occurs because the
heat transfer rate decreases as time increases, whereas the mass
transfer rate is  independent of time.
   Finally, the  emission rate  is equated to  zero when  all  the
chemical  is evaporated. Chemicals with a normal boiling  point
higher than ambient temperature are  assumed to form a liquid
pool at ambient conditions.  Here, the chemical is  assumed to
evaporate only by convective mass transfer.

Continuous Leaks
   Continuous leaks are regarded in the present work as emissions
from a small rupture of a transport line (such as a pipeline) or a
storage tank (such as a tank car). It is assumed that the mass flow
rate through the orifice is known and  that the chemical flows at
atmospheric conditions, i.e., as a gas or as a liquid depending on
the  normal boiling   point  being  below  or  above  ambient
temperature, respectively.
   The chemicals of interest in the present program are all  more
dense than air  so that a gaseous plume or a liquid pool will be
formed at ground level after the chemical exits the original car-
rier. Convective mass transfer is assumed to be the limiting pro-
cess for the chemical to be transported  by the wind. The spill area
then is predicted by equating the mass transfer rate to the flow
rate through the rupture.
   The pool size is determined for both laminar and turbulent flow
cases and is used to predict the corresponding Reynolds numbers.
The value that gives the consistent flow regime is chosen as  the
final pool length.
   The area of the spill is used as a ground level area source in the
air dispersion model.  In addition, the flow rate through the rup-
ture is utilized  as the emission rate in  the air dispersion calcula-
tions. The user has the option of specifying the length of time
from the beginning  to the end  of the flow.  This time  can be
estimated by knowing the amount of material to be emitted. This
case is approached by the air  dispersion model as an unsteady-
state case where the constant emission rate is dropped to zero at
the  corresponding time. The program assumes a  steady-state
operation if no time is specified by the user.

Instantaneously Formed Pools
   The time-dependent evaporation rates of stationary, instantan-
eously formed  pools are calculated by two different procedures
depending upon the chemical spilled: (1) liquified gases or (2) liq-
uids. However, both methods assume that the chemical pool does
not spread on the land as a function of time, such that the spill
area specified by the user remains constant.

Liquified Gases
  An initial adiabatic flash calculation  is performed based on the
pressure drop of the  liquified gas (i.e., from  the pressure of the
chemical  while stored, to atmospheric pressure).
  The original  enthalpy of the chemical is evaluated as the en-
thalpy of the saturated liquid at the user's specified temperature.
For conservative purposes, the amount of vapor which is flashed
off initially is assumed to be emitted in a period of 1  min.
  The rest of the liquified gas remains in the liquid pool specified
by the user at its normal boiling point. From this point on, the
chemical  can undergo two different  transport  processes:  heat
transfer or mass transfer.
  Due to heat transfer, which occurs mainly by conduction  from
the soil assumed at ambient temperature and by convection  from
the atmosphere, the  chemical  evaporates off the  boiling liquid
pool. It is assumed that solar radiation is negligible and that no
thermal resistance exists between the soil and the boiling liquid.
The  mass range at which the chemical evaporates is then deter-
  The air dispersion model uses the time-dependent mass rate
calculation to estimate the source strength during a heat transfer
evaporation mechanism  based on  the total amount  of  mass
evaporated between a time t(0) and any later time t.
  Since all the chemicals  treated in the present work are  more
dense than air, the emission rate into the atmosphere is  mass
transfer limited. This means that only a portion of the evaporated
mass will be transported by the wind. Mass transfer governs this
take-up flux into the atmosphere. The rest of the evaporated
chemical remains as a gas cloud above the liquid pool. For conser-
vative purposes, and due to the fact that the person at the scene of
the accident will report an estimate of the size of the pool, it is
assumed in the present work that all the mass evaporated by heat
transfer will be emitted into the atmosphere.
  Mass transfer occurs  predominantly by forced convection over
the liquid pool due to the wind. The amount of chemical emitted,
used in the air dispersion model, is obtained from integration with
respect to time of the amount of material emitted  by mass transfer
during the specified interval of time.
  Heat  transfer dominates at the beginning of  the process.  As
time increases, the  heat  transfer emission  rate decreases and
reaches a point  in  time where it  becomes equal to the  mass
transfer rate.

  Liquids are  defined as those chemicals with a normal boiling
point,  T(B), higher  than the ambient  temperature, T(0). These
chemicals are transported  at conditions very close to ambient so
that  emission into the  atmosphere after a  spill will occur only
through mass transfer by convection. The risk to health and life
from an atmospheric release of hazardous effluent is a function of
the ambient downwind concentrations and the length of the  ex-
posure period.

  A  vital element in such a consequence assessment is the ability
to estimate the location  and concentration of an airborne effluent
plume. To provide this estimate, the dispersion models employed
must include both the transport and diffusion properties of the at-
mosphere.  Consequently,  the  models must account  for  the
physical, temporal  and  spatial  changes  that  the atmosphere
undergoes during  the course of the release and during transport.
  To estimate the downwind location  and concentration of air-
borne effluents, CARE uses a highly sophisticated atmospheric
dispersion model. The model includes  both transport and diffu-
sion  components  which account for  the  temporal  and spatial
changes that the atmosphere undergoes during the course of the
release and during transport.
  Atmospheric dispersion calculations are  performed usin$  the1
ESC  variable  trajectory, puff-advection  model—TRAJ.' The
model simulates a release as a series of superposed puffs which are
transported by the local atmosphere. Each specification of new
meteorology is considered a new transport interval. During such a
transport interval, the  puffs are diffused according to  standard
Gaussian theory. The dimensions of an individual puff are pro-
portional to its travel distance (or travel time). The model accom-
modates multiple point  sources and includes algorithms for plume
rise and shoreline effects. The model  provides a. rigorous  treat-
ment of the spatial and temporal variations in the atmosphere and
accounts for influences due to local  meteorological effects.
  The advection of the puffs is accomplished through inputs of
local meteorological variables, wind speed and direction. The

model can accommodate several meteorological stations by apply-
ing a weighting to each station variable. For these applications in
areas of complex flow fields (e.g., mountains, valleys, shorelines,
etc.), a three-dimensional wind field model can be added to pro-
vide a rigorous, dynamic treatment of the spatial and temporal
variations in the atmosphere.
  The dispersion model calculates the concentration over a grid
covering the modeling region and also at specified special receptor
locations within the grid. Special receptor locations are stored as
data to facilitate modification should additional stations be needed.
Site-specific run parameters, which are input as data, define  the
various site-specific characteristics unique  to this model applica-
tion. The calculation is performed in a time-wise, step-by-step
analysis of the plume  path from the release point through  the
limits of the calculation grid.

  Dispersion  modeling  requires  the  input  of   critical
meteorological data. Wind speed and direction are needed to
define the transport of airborne material;  stability data are used
to determine the spread of the material as it travels downwind.
The depth of the layer of the atmosphere where mixing occurs is
called Mixing Depth; this parameter defines the maximum height
of  the  plume.  Temperature  data  are  used to determine  the
bouyant rise of the plume and, where applicable, the existence of
a sea breeze.
  At each modeling time  increment  (generally  15 min),  input
meteorological data are used to determine:
• Transport wind (u,v,w) components
• Plume dispersion coefficients (sigma-y; sigma-z)
• Atmospheric temperature
  The meteorological data  can be acquired in real-time from an
on-site meteorological  tower  or,  if data are acquired manually
from alternate sources such as National Weather Service stations
or United States Air Force bases, they can be input from the con-
sole via menu responses. Wind data, input as speed and direction,
are converted to transport wind (u, v and  w) components at  the
grid points. Horizontal and vertical plume  dispersion coefficients
(sigma-y and sigma-z) can be derived from direct measurements
of turbulence  (sigma-theta  and sigma phi) or from routine me-
teorological observations using well-known stability classification
schemes, depending on the available input data.

Stability Classifications
  The most commonly used method to determine dispersion coef-
ficients  is  the indirect method, requiring the determination of a
stability class which  then determines the  dispersion curve. The
Pasquill-Turner stability classification  scheme  or the  NWS-
Turner  method are typically used  with  the Pasquill-Gifford
dispersion curves. Stability categories cannot be measured directly
but must be determined from other meteorological parameters.
Any of the following approaches  can be used:
• Sigma phi (standard  deviation of the vertical wind direction)
• Sigma theta (standard deviation of wind direction)
• Delta temperature (the difference in temperature with height)
• Cloud conditions (i.e., clear,  scattered, broken or overcast,
  and ceiling height)
• Default conditions (e.g.,  F-nighttime; D-daytime)
• Manual override inputs (operator discretion)
  When using sigma theta or sigma phi to determine stability clas-
sifications, and changes in dispersion are expected because of sur-
face roughness, a factor of (Zo/15 cm) A0.2, where Zo is  the
average surface roughness in centimeters, may be applied to  the
tabular values. Suggested Zo values that may be used as a guide to
estimating  surface  roughness  are  given  in  Hogstrom  and
  The  sigma stability classification  schemes  are  adequate for
daytime use,  but during nighttime (1 hour to sunset to 1 hour
after sunrise), adjustments  adapted from Mitchell and Timbre3
are used.
  If cloud cover inputs from NWS or USAF sources are used, the
values  are compared to criteria  published in  the  U.S.  EPA
CRSTR model. Default dispersion classes  (e.g.,  D-daytime;
F-nighttime) can be provided for the operator  in the event no
other data are available.

Dispersion Coefficients
  The  purpose of establishing a stability class is to be able to
calculate  the  horizontal and vertical dispersion  parameters,
sigma-y (Sy) and sigma-Z (Sz), which are used in the calculation
of downwind concentration. Sy and Sz are approximated by a
curve fitting process which describes the PGT  curves with  power
laws  as a function of the total distance  traveled, x:
  Sy(x) = a*vAb
  Sz(x) = c*vAd + e
where the coefficients (a,b,c,d,e) are a function of stability class.
The coefficients are taken from the POLYN routine found in
NUREG/CR-2919, the manual for the XOQDOQ model, written
by Sagendorf, et al.' but have been modified slightly to improve
the continuity of sigma-z from the region of x < 1000m to x < =

Mixing Height
  The  boundary layer height  (H)  is used to limit the  vertical
dispersion parameter (Sz)  to values  < 0.8*H. The mixing depth
will be determined by applying the climatological mixing heights
according to the method  described  in the U.S. EPA CRSTER
manual,' which is the standard Holzworth method.

Sea Breeze
  In regions near  large bodies of water where a  sea breeze
phenomenon may be a significant  meteorological feature, the
Thermal Internal Boundary Layer (TIBL) will be used as the  max-
imum mixing height. The software will determine if a sea breeze
condition is likely to exist, using the following rules:
• The sun is above the horizon (based on the data and time)
• The lower-level air temperature is greater than the sea water
  temperature, or monthly climatological temperature  if  tem-
  perature data are not  available
• The  lower  wind direction  measurement  indicates on-shore
  winds within a 180 degree sector
  If a sea breeze is determined to exist, the mixing height  is set
equal to the height of the Thermal Internal  Boundary Layer
  H = C * sqrt (d)
where H is the TIBL height (in meters) at distance d (in meters)
inland6'7 and C is a constant determined for the local area.
Wind Field Model
  For applications in areas of complex flow (e.g., valleys, moun-
tains, seashores, etc.) a three-dimensional,  mass-consistent flow
field model is used to provide a rigorous, dynamic treatment of
the spatial and temporal variations in the movement of the cloud.
  The ESC windfield model (ESCWIND) was adapted from code
available  in  the  public domain  from several  sources. It  was
developed  to provide  the user with  an economical windfield
                                                                                              DETECTION OF RELEASES    37

predictor.  The  model  produces three-dimensional,  terrain-
dependent,  divergence-free windfields given  observed surface
and/or upper-air data as input. Given the wind components for
each meteorological station and at each level, a 1/r weighted in-
terpolation is used to "fill  out" the  rest  of the windfield along
each horizontal plane.
   The terrain surface boundary is approximated  using obstacle
cells to represent the terrain in a stairstep fashion. For computa-
tional purposes, the irregular  surface boundary is removed by a
coordinate transformation  in  which the  terrain surface also
becomes a coordinate surface.  This "sigma-space"  computa-
tional domain has two main advantages. First,  the bottom boun-
dary condition is defined more accurately; second, the option of
variable vertical  zoning improves  the model's  accuracy and
   ESCWIND follows a procedure  of extrapolating and  inter-
polating  the input  data  to  determine  an  estimated  three-
dimensional  windfield on a specified finite difference grid. This
prospective  windfield then is adjusted to account for  terrain
effects and atmospheric stability considerations constrained  by
the condition that the resulting windfield be nondivergent.
   The basic  set of model  equations was  derived from the ap-
proach that minimizes the squared variation of  the windfield sub-
ject to the constraint that the adjusted field be  nondivergent.8'°

   The CARE Dispersion Modeling System has many applications
such as toxic spills, nuclear dose assessment, pesticide drift, am-
bient air quality assessments  in complex terrain,  chemical fires
and visibility obscurations.  The dispersion model is designed to
provide physically  realistic calculations of the position and con-
centration  of downwind airborne effluents. These calculations
then are available  for application in any scenario that relies upon
atmospheric  dispersion calculations. The type of application for
which the system is to be used  will dictate the type of source term
calculations and the effects  and response modules that would  be
useful in the  CARE system.

Nuclear Accidents
   The system is applicable  for risk assessment and operational
dose  assessments   from  radioactive  releases.  It  has  been im-
plemented  for real-time, operational  dose assessments using ac-
tual meteorological data signals,  for manual and back-up dose
calculations and for training purposes. The model can be used for
dose  projection   estimates based  on  forecase  or  historical
meteorological data. For training exercises,  the operator would
run the model using the  interactive menus where input data and
run parameters have been selected to simulate the desired exercise

Toxic Spills
   For  toxic spills emergency  response, a  chemical source  term
model has  been added which calculates the appropriate  release
rate based on chemical properties, release mode and weather con-
ditions. The  effects model  provides recommended response ac-
tions based on the concentration Threshold Limiting Value or the
Lower Explosive Level.

Visibility Obscurations
  By using a  visibility model to calculate the effects of an aerosol
released to the air, the  system has been  used to calculate the
obscuration effects of various smoke releases for military applica-
tions.  This capability also is useful in determining  visual impacts
on U.S. EPA Class I areas.
Ambient Air Quality
  The system is extremely useful for calculating ambient air qual-
ity  in complex terrain  where traditional straight-line Gaussian
models typically perform unrealistically. The use of the variable-
trajectory calculation capability of the system helps  to account
for  terrain-induced dispersion.  This  capability is  useful  in
transport studies such as odor and  toxics impacts.

  CARE gives one the capability to perform a variety of emergen-
cy preparedness functions:
• Risk Assessment
• Hazmat Training
• Real-Time Monitoring
• Emergency Response
  CARE is a system that puts truly capable emergency prepared-
ness on one's desk. In the event of an accident that releases a
hazardous material into the air, CARE will tell one where it will
go, who it will effect, which areas are safe as evacuation centers,
how long it will take to disperse and which preplanned response
options would be most effective. CARE gives one the ability to
neutralize the danger, to assess alternative defense strategies and
to keep abreast of changing situations. In short, CARE gives one
the tools  needed  to prevent an emergency from becoming a
 I.  Fleischer, M.T., "SPILLS, An Evaporation/Air Dispersion Model
    for Chemical Spills on Land," Shell Development Company, Hous-
    ton, TX, Dec. 1980.
 2.  Hogstrom, A.S. and Hogstrom, U.. "A Practical Method for De-
    termining Wind Frequency Distributions for the Lowest 200m  from
    Routine Meteorological Observations." J. Appl. Meteor., No. 17,
    1978, 942-954.
 3.  Mitchell, A.E. Jr. and Timbre, K.O., "Atmospheric Stability  Class
    from Horizontal Wind Fluctuations," paper presented at the  72nd
    Annual  Meeting of the Air Pollution Control Association, Cincin-
    nati, OH, June 1979.
 4.  Safendorf, J.F., et  a/.,  "XOQDOQ:  Computer Program for the
    Meteorological Evaluation of Routine Effluent Releases at Nuclear
    Power Stations," NUREG/CR-29I9, Sept.  1982.
 5.  U.S. EPA, "User's  Manual for Single-Source (CRSTER) Model,"
    EPA-450/2-77-013, July 1977.
 6.  Raynor,  G.S.,  el  a/.,   "Recommendations  for  Meteorological
    Measurement  Programs  and  Atmospheric Diffusion Prediction
    Methods for  Use at  Coastal Nuclear Reactor Sites," NUREG/CR-
    0936, Oct. 1979.
 7.  Beebe, R.C. and Sorge,  J.M., "Coastal and Overwater Dispersion
    Characteristics  of a  Warm  Water Environment,"  unpublished
    paper, Apr. 1975.
 8.  Patnaik, P.C. and Freeman, B.C., "Improved Simulation of Meso-
    scale  Meteorology Phase I," Science  Applications, Inc.,  Report
    SAI-77-915-LJ, Mar. 1977.
 9.  Sasaki,  Y., "Some  Basic Formalisms in Numerical Variational
    Analysis," Won., H'ea.,  Rev., No. 98, 1978, 312-319.
10.  Sherman, C.A.,  "A  Mass-Consistent Model for Wind Fields  Over
    Complex Terrain," J. Appl. Meteor., 1978, 312-319.

                    The Use  of PC  Spreadsheet-Based  Graphics
            to  Interpret Contamination at CERCLA/RCRA  Sites

                                            George  A. Furst, Ph.D.
                             U.S.  Environmental  Protection Agency, Region 1
                                             Boston, Massachusetts
  During site investigations, large amounts of environmental data
are generated in the form of lists of chemical analyses from differ-
ent sampling locations over various periods of time. These tables
of data are difficult to analyze and are seldom fully interpreted
because of the magnitude of the task and the difficulty of visualiz-
ing trends from long lists of numbers. Electronic spreadsheet soft-
ware for personal computers, originally developed for use by the
business community for financial analysis, may be employed to
store site data and to graphically display contamination patterns
at CERCLA/RCRA sites. This software allows the input  of
numbers, formulas and words into a matrix of rows and columns
and provides an ideal format for listing contaminant concentra-
tions in the air, soil, surface water and groundwater.
  When the above data are entered into a spreadsheet  and in-
tegrated with  graphics software, it is a simple task to rapidly
visualize contamination distribution at a site over space and time.
The following spreadsheet-based graph types are illustrated: the
line, bar, stacked bar, x-y  and pie chart. Each type is well suited
to a particular application.  For example, the pie chart is useful for
showing a connection between a source area (such as  a  lagoon)
and a receptor (such as a downgradient well). The expanded bar
chart can be used to indicate if new sources of contamination are
affecting the downgradient groundwater. Specific applications of
the spreadsheet-based graphics  are illustrated  using data from
Region 1 CERCLA/RCRA sites.

  The  advent  of the personal computer and  its increasing
availability, as well as the increase in user sophistication, has led
to new applications of this  powerful instrument. The spreadsheet
has become the basis for many financial as well as scientific ap-
plications. One recent example of the use of the spreadsheet is
that of Olsthoorm.'  In this case, the spreadsheet was used to
model groundwater flow in two and three dimensions.
  The spreadsheet is used  because it is a general purpose, inex-
pensive  software product  and  one need not  be a  computer
specialist to use it. In the above application,  the spreadsheet
replaced three or more specific programs. This paper applies the
spreadsheet analysis to aid in  understanding the flow of con-
tamination from an uncontrolled hazardous waste site  and from
an active manufacturing site.
  Often there  is  a great deal  of information  known about a
specific site. This information is available as tables of chemical
data collected while sampling the groundwater, surface water,
soils, sediments and air at the site. Without being able to organize
and visualize this data, one often does not utilize the data to the
full potential. The computer-based electronic spreadsheet, com-
bined with graphics, provides the means to more fully interpret
  The spreadsheet,  more than any other software program, has
revolutionized the use of the personal computer. The spreadsheet
consists of a matrix of rows and columns in which numbers, for-
mulas and words are input. It is also a very user-friendly software
tool that is limited only by the imagination of the user. Examples
of popular spreadsheet-based software are: Visicalc, Multiplan,
Lotus 1,2,3 and Symphony.
  The spreadsheet used in this paper is Symphony, which along
with Lotus 1,2,3 represents a new generation of integrated soft-
ware; word processing and graphics are  directly available while
working with the spreadsheet. The Symphony spreadsheet con-
tains 256 columns and 8,192 rows. This capacity allows all known
chemical  information from a specific  site  to be stored on one
spreadsheet. The size of the spreadsheet  is demonstrated by the
fact that  the above matrix may hold 63,000 individual analyses
consisting of 30 chemical elements or compounds. In total, the
spreadsheet contains over 2 million available data points,  limited
only by the computer's Random-Access Memory. The advantage
of storing all the data from one site on one large spreadsheet is the
ability to rapidly compare different sections of data and to con-
duct direct statistical analysis on specific ranges of data. The ap-
plication described in this paper is the ability to focus on sections
of data in the spreadsheet and to make a picture or graph of the
numbers. An important assumption made is that the data are ac-
curate and have passed a quality screening  test.

  Table 1 is a portion of one spreadsheet of data from a Super-
fund site in southern New Hampshire. It  is displayed in the stan-
dard spreadsheet  format: i.e., the  information is listed in rows
and columns. Note that both numbers and words are used. At
this site, there are a sufficient number of monitoring well sam-
pling locations to define the cross-section of a contaminant plume
adjacent  to a water supply. The  horizontal  axis of the table
represents the distance in feet from the assumed center of the con-
taminant  plume to the center of the  sampled monitoring well
screen. All samples were taken in the same period of time.
  An examination of the data shows that  total contamination de-
creases with distance from the center of the plume. Also, the con-
centration of 2 -butanone decreases rapidly, but it is not as ap-
parent that t-l,2-dichloroethene is  decreasing much less  rapidly
across the plume cross-section.
                                                                                       DETECTION OF RELEASES    39

                                                           Table 1
                                    Distribution of Contamination, Plume Center to Water Supply

              Distance  from Plums Center        0         70       130        175      250                 310

              Sample  number             NUS 2-3   NILS  2-2    NIJS 2-1    N'JS  -2   C5X  11     NUS 2BR    WP-4
              1,2-Dichlorde thane
              1.2-D ichloroethene
              Vinyl Chloride
              Total Xylene



























               Note: All concentrations in ppb (or jig'l)
   Figure 1 is an x-y plot of the distribution of specific organic
contaminants versus distance from the center of the plume. Using
this plot, one can see the partition of different organic chemicals
in the plume diffusion gradient perpendicular to a groundwater
flow  direction. This distribution may be due  to differences in
sorption as discussed by Mackay, el al.1 This hypothesis can be
modeled by taking additional data from the spreadsheet (such as
analyses from downgradient wells) and comparing these analyses
to the above theory. Because the data are readily available on the
spreadsheet,  examination  of contaminant  distribution  in the
water supply well and the bedrock  well is facilitated. These two
analyses are listed in the columns headed by  WP-4 and NUS 2BR
on Table 1. The pie chart (Figs. 2 and 3) is jsed because it graph-
                         VOLATILE COKT/umMON NEAB WP «U
                          DISTANCE fDOU PUIME CENTER (IT)
                      TOtUENE        6  H.2-OPCHUWOETHENE
                          Figure I
  X-Y Diagram of Volatile Contamination in a Groundwater Plume
 ically depicts the relative percentages of organics impacting these
 different  receptors and specifically  shows the  persistence  of

  Groundwater monitoring at manufacturing facilities is a critical
 aspect of the U.S. EPA RCRA permitting process. Implementa-
 tion of the 40 CFR Part 265 and 270 groundwater monitoring se-
 quence generates much data related to the permitting process,
 either closure or post-closure. If the detection monitoring wells at
 a land disposal  site detect contamination,  the  site immediately
 goes into assessment monitoring.
  The goal of this monitoring is to clearly identify the rate am ex-
 tent of hazardous waste constituent migration and to establish the
 concentration of individual constituents in the plume. For all of
 the above  assessment  data, the spreadsheet provides an ideal
 storage file for later retrieval and interpretation.
  The following discussion is an example of the  ways in which
 one can use the above data once they have been entered into the
 electronic spreadsheet.  The  site utilized for  illustration  is  a
 northern New England metal machining and plating company
 that has been active since the early 1950s. From  1970 to 1985, an
 unlined waste lagoon was used to manage hazardous waste at the
 site. Detection monitoring wells were installed in 1981 to assess
 the site groundwater flow and  potential contaminant migration
 irom the waste lagoon. During the early 1980s,  about 100,000
 gal/day of waste from the manufacturing process entered the
 lagoon. The waste stream was  composed principally of solvent
 wastes from metal degreasing, waste oils from metal  machining
 and sludge from electroplating operations. In order to assess the
changes in contaminant migration at the site, all data collected
 from  1981  to 1985 were entered onto  the spreadsheet. Table 2
contains the data collected during the fall of 1982. The following
 figures are based upon this table.

                       TOTAL VOLATILES - 93 PPS
    2-Bt/TANOtC (».OX)
                                             1.1-OCHLOROETHANE (J4.4X)
                                   1-1.2-OCHLOROETHENE (19.4%)
                          Figure 2
Pie Diagram of Volatile Distribution in Monitoring Well NUS-2BR
                      WP WATER SUPPLY
                      TOTAL VOLATILCS - « ffB
            TRICHLOROETHE)* (».3»)
                                  1.1.1-TRCHLOROETMANE (8.3*)
                            1-I.Z-DICHlOroCTHeNE (83.W)
                         Figure 3
     Volatile Organic Contamination in WP Water Supply
                  Figure 4 is a stacked bar chart that illustrates the composition
               of the oily waste lagoon. This graph  is used because it gives a
               visual summary of the contaminant profile at different depths in
               the waste lagoon in 1982. Figure 5 shows the distribution of con-
               taminants at mid-level in the lagoon. Figures 6, 7 and 8 illustrate
               the volatile distribution in wells 3, 4 and 7, located 50 ft adjacent
               to and 50 and 1,450 ft downgradient from the lagoon. The main
               difference  in contaminant profiles is the absence of methylene
               chloride in the farthest downgradient well MW-7.  The  other
               organic chemicals are  present at lower concentrations than in the
                               MAJOR COMPOSITION OF OILY WASTE LAGOON
                                            SAMPLES Of 10/5/82
                   35000 -

                   3OOOO -

                   25000 -


                                                                    E7"3 METHYI^NE CHLORIDE

                                                                    [V~l 1.1.1-TOCHLOROETHANE
                                                 IT^ TMCHLOROETHYIENE

                                                 rvTl 1.1-DICHLOROETHANE
                                                                                             Figure 4
                                                                           Volatile Organic Concentrations in Waste Lagoon
                 The pie diagram is an ideal graph to show the hydrologic con-
               nection between source area, lagoon and receptor which in this
               case is the downgradient well. At sites where there are multiple
               source areas,  such  as  separate lagoons used for waste oil and
               plating and degreasing wastes, the pie chart represents a signature
               that is useful  for indicating which downgradient wells are con-
               nected hydrologically with a specific lagoon.
                                                           Table 2
                               Lagoon Composition at Surface, Mid-Depth, Lower and Sludge (10/82)
                                        Monitoring Well Groundwater Composition (10/81)
DATt:                      10/5/82
SAMPLE NUMBER           Surface

1,1-Dichloroethane           630
Methylene Chloride
1,1,1-Tr ichloroethane

Total:  (PPB)
                                     10/5/82   1-1/5/8 ^
                                     M id-Dap th   lower
                                            0       0


22775   43731

10/16/131 10/81
   MW-4 MW-7
  50    1450














                4617    3230
Note: All concentrations in ppb 0
                     OILY WASTE LAGOON. MID-DEPTH
                          IOIAL voutus - 2277s m
             HWCHUXlOCTmVCW (I0.2X)
                                                  I.I-OCW-OHOCTHWC (31 !»)
             urnmrxc cmonct <«3 M)
                               Figure 5
     Volatile Organic Concentrations at  Mid-Depth in Waste Lagoon
                    TOT*. KXATUI - «M4 m
                                                                                                                            I.I-OCMU*OCTNW*OCT»*M*t (07X)
                                                                                                                    TTTHMXJCMOCTMtUM (OJX?
                        Figure 7
Volatile Organic Contamination of Groundwater at MW-4
                                  i - 3M t ^0
                  I0t«t MX*rU3 - 3241 »»•
                mO«.OHOCTHTU>C (
                           SO MONITORING WELL 3
                            PERIOD SAMPLED 10/81 - 10/85
                         SO MONITORING WELL 3
              1.1 -OICHLOROETHANE
                              t. 1.1 -TR1CHLOROETHANE
                           BATE SAMPLES ANALYSED

                    U7~* 2/U/82       r^J 10/5/82
                                                   r\7\ 8/11/&J
                    E53 5/25/84
                                      | 10/85
                           Figure 9
         Volatile Organics Present in MW-3 (10/81 - 10/85)
  The spreadsheet also provides the basis for comparing changes
in organic contaminants at a specific well over a number of years.
This comparison is facilitated since data can be moved rapidly
from one section of the spreadsheet to another. Table 3 contains
data from monitoring well 3 which were collected from October
1981 to October 1985. Figure 9 is a bar graph of the major organic
contaminants, which,  with the exception of methylene chloride,
are contained in the groundwater. The graph illustrates that the
highest organic  contamination at well 3 occurred in late 1982.
Since then, there has been a decrease in the concentration of the
three compounds, as shown in this figure. Figure 10 shows the
concentration of methylene chloride over the same time period.
The highest concentrations of this  species were observed in Oc-
tober 1982 and  October  1985. The later date coincides with the
final closure of the lagoon and may be related to it. This line chart
is particularly useful for showing changes in contamination over a
period of time.
  In summary, spreadsheet-based graphics allow one to visualize
the  changes  in  concentration of  organics in the source  area,
lagoon and downgradient monitoring  wells and changes over a
4-year  period in  well number 3.
                                                                                     O DATES OF SAMPLING
                                                                                            Figure 10
                                                                      Concentration of Methylene Chloride at MW-3 (10/81   10/85)
  The spreadsheet is the fundamental way to store data in the
computer. Once data from a hazardous waste site or manufactur-
ing site are entered into the spreadsheet, one can readily combine
data, move data and, when the spreadsheet is  combined with
graphics software, one can generate graphs from any portion of
the spreadsheet. In this paper, the spreadsheet has formed the
basis for quickly assessing the contamination at two sites. The
software also has been used by the author to  prepare graphs for
overhead projection at a meeting, to prepare tables and graphs
for reports and to forecast when increasing contamination in a
well will cause the water to exceed Safe No Action  Response
Levels (SNARLS). With the increasingly available portable com-
puters, the above site-specific information can be taken  to the
manufacturing  company and used as a resource during discus-
sions regarding contaminant distribution at an RCRA site. As
stated earlier, the application of this new generation of software is
limited only by the imagination of the user.

1.  Olsthoorm, T.N., "The Power of the Electronic Worksheet:  Model-
   ing Without Special Programming," Groundwater, 23, 1985, 381-390.
2.  Mackay,  D.M., Roberts, P.V. and Cherry, J.A., "Transport of Or-
   ganic Contaminants in Groundwater," Environ. Sci.  Technol., 19,
   1985, 384-391.
                                                                                              DETECTION OF RELEASES    43

         Emergency Response to  Toxic Fumes  and  Contaminated
                          Groundwater  in Karst Topography:
                                             A  Case Study

                                              P. Clyde  Johnston
                                               Mark J. Rigatti
                                             Roy F. Weston, Inc.
                                      Jacobs Engineering Group,  Inc.
                                               Atlanta,  Georgia
                                               Fred B.  Stroud
                            U.S. Environmental Protection Agency   Region 4
                                               Atlanta,  Georgia
  This paper presents a preliminary report on the on-going con-
trol of contaminated aquifers in karst topography. Over the past
several years, there has been a growing concern over the presence
of volatile hazardous chemicals in the karst formation cave sys-
tems in the Bowling Green, Kentucky area. By April 1985, con-
tamination and related chemical fumes warranted an "emergency
response action" by the U.S. EPA, Region 4. The management
of migration procedures under these conditions are discussed and
summarized. Dye tracing, geophysical studies, speleological car-
tography, exploratory drilling, down hole cameras and sampling
techniques have  been utib'zed to locate the primary  flow char-
acteristics of the  "Lost River" and its tributaries and to identify
the source(s) of contamination and mitigate its effects.
  Karst formations  present peculiar  and distinct management
difficulties. The  impact of leaking underground storage tanks
and illegal dumping  may be magnified by the motility of spilled
substances. Pollution source determination is especially proble-
  Many methods used in this emergency response action were
relatively  experimental with few  strong  precedents.   These
methods are comparatively analyzed for the benefit of  future
cleanup activities.

  Karst topography, evident to some degree almost anywhere
carbonates are found at or near the surface of the earth and sub-
jected to  humidity and water movement, presents  unique and
perplexing difficulties to the professional involved in hazardous
waste management. By definition, karstic terrains are character-
ized by closed  depressions, i.e., sinkholes or dolines and uvalas,
caves and the prevalence of underground drainage. Although
these conditions can  be beneficial, as in the case of high capacity
groundwater accessibility (it is the closest thing in  nature to a
fresh water pipe), extreme liabilities also may be present. For ex-
ample,  the impact of  leaking underground storage tanks and
illegal dumping may be greatly magnified by the motility of lost
substances. While permeability in non-karst areas may be meas-
ured in ft/day, permeability in karst can be measured in miles/hr.
Discrete sources of groundwater recharge are numerous and var-
ied and the velocity of recharge so high that pollution source de-
termination can be especially complex.

           1 LOST RIVER UVALA I
             CAVE ENTRANCE

           3 LOST RIVER RISE
                       Figure 1
                 Bowling Green Location
  Located on the textbook karstic terrain of south central Ken-
tucky, Bowling Green is the only city of such size (pop. 41,000)
in the United States to rest almost entirely over a cave system.
The Lost River, which is exposed in a uvala in southwest Bowling
Green, flows into the Lost River Cave  entrance and  has been
mapped for 4 miles. At this point, the river sumps and takes an
undetermined flow path through the St. Genevieve and  St. Louis
limestones underlying Bowling Green until it exits north of the
city at the Lost River Rise.

  Fixed  facility  and  transportation related  spills,  deliberate
dumpings and undetected leaks, all associated with oil and haz-
ardous materials, have become daily occurrences in every major
city across the nation. Bowling Green is no exception. Its loca-
tion in a sinkhole plain and over an extensive cave system, makes
the city highly susceptible to groundwater pollution and to the re-
lated problems associated with karst areas.
  Many  contaminants can immediately  affect the  watershed
either through unobstructed flow into the cave system or through
rapid percolation through thin cover soils. In Nicholas Craw-
ford's earlier research, he showed that heavy metals and road
salts from urban detritus quickly entered the  groundwater sys-
tem,1 but volatile materials pose special problems.  As the sub-
stances flow through the subterranean labyrinth, they can volatil-
ize in the cave air space and ultimately find their way to the sur-
face.  The control of gases and vapors in the atmosphere usually
is  not attempted because natural dispersion dilutes  vapors and
gases rapidly. However, fumes rising from  cave systems may con-
centrate in areas of concern. If these fumes collect in a basement,
crawl space or simply in or near public areas, a potentially explo-
sive situation exists.

 Bowling Green's Problem
   The toxic and explosive fume problem in Bowling Green dates
 back to 1969, when explosive fumes were detected and believed
 to have caused  a residential  explosion.  In 1981, the residents of
 five homes in the Riverwood area  were evacuated  when fumes,
 rising into their basements and crawl  spaces, reached explosive
   In 1982, benzene and  methylene chloride were detected in the
 Lost River Cave. Subsequently, the U.S. EPA, Region 4, initiated
 an emergency response action in March of 1983. Preliminary in-
 vestigations showed alarmingly high levels of volatile chlorinated
 hydrocarbons2  flowing  into the spring-fed Keith  Farm pond,
 which discharges into the Lost River drainage system. Dye traces
 pinpointed the source of the  contamination to be a series of leak-
 ing industrial underground storage  tanks.  The U.S. EPA treated
 1.9 million gal of contaminated pond water, and the responsible
 party removed the leaking tanks and excavated the contaminated
   Throughout 1984, toxic and explosive fumes plagued the homes
 in the Forest Park and Parker-Bennett School areas. In January
 1985, toxic and explosive  fumes were  detected in  the Dishman-
 McGinnis Elementary School. To date, fumes of this type have
 been investigated in over  100 homes and commercial buildings,
 two schools, one church, 18 drainage wells and 24 sinkholes.
   After an extensive public  health  evaluation of  the fumes by
 the Center for Disease Control, a health advisory was issued on
 the grounds that explosive levels existed and that the detectable
 fume levels of benzene,  toluene and xylene exceeded standards
 for  "non-occupational settings,"  At that time, the U.S.  EPA
 Region 4 Emergency Response Branch initiated a  "Superfund"
 emergency response to address the problems designated in the
 public health advisory.  Mitigation efforts were initiated to  re-
 duce the threat, identify the  source or sources of contamination
 and  define the  underground  drainage system controlling the
 transport of chemicals in the Bowling Green area.
upon the assumption that  continuous releases or intermittent
dumpings into the Lost River drainage system, i.e.,  sinkholes,
solution joints and fractures, have resulted in a chronic flow of
chemicals within the caves.
  The first  hypothesis is that, once in the  cave, the chemicals
volatilize; a volatilization that is accelerated by natural struc-
tures. Waterfalls and riffle areas provide an indigenous turbu-
lence which increases the rate of volatilization, increasing, in
turn, the  fume concentration within the cave atmosphere.  Dur-
ing high water flow,  the air space within the cave is decreased,
forcing the gases to the surface. This process  not only occurs dur-
ing high water flow, but also during normal cave breathing due
to changes in barometric pressure.
  The second hypothesis is that water-filled  passages,  forming a
natural underflow dam, trap floating chemicals  and restrict them
from flowing downstream. This flow restriction, in turn, causes
low density  and volatilize chemicals  to become concentrated in
these areas. This  frequently occurs  during  increased river  dis-
charge when heavy rainfall causes the river within the cave to rise.
Consequently, the floating chemicals and the fumes are pushed
  As the water recedes, the floating chemicals adhere to the ceil-
ing and walls forming a "bath tub ring" within the system. This
bath tub ring phenomenon coats the surface of upstream perched
pools with floating chemicals.  The rise and fall of the  river level
results in a  repetitive deposition of  chemicals  within the cave,
leaving a stratified layer of aliphatic hydrocarbons and mud.
  Finally, it is proposed that contaminants have been dumped or
spilled in soils and areas underlain by perched water tables or
water pockets and are flushed into the Lost River Cave  system
when meteoric water  percolates through the soil or overflows a
perched water table. Perched water tables and water pockets are
associated with the pinnacle  nature of the limestone at the lime-
stone/soil interface (Fig. 2).
                                                 (Terra Rossa)
                                                PERCHED WATER
                                                (VARIABLE SIZE)
                           Figure 2
    Generalized Illustration of Pinnacle Nature of Limestone Showing
          Possible Perched Water Table and Water Pockets
   Although not fully documented, there is an apparent direct re-
 lationship between increased reports of fumes and wet weather;4
 similarly, there have  been  fewer fume incidents  during  dry
 weather. Thus, the fumes appear increasingly manifest in periods
 of heavy precipitation. The  hypotheses which follow are based
  These hypotheses are not mutually exclusive. Investigative tech-
niques in verifying them and solving the resulting problems asso-
ciated with them have been varied. In this verification, geological
techniques were used in conjunction with sound environmental
studies and common hazardous waste site cleanup practices.
                                                                            CONTAMINATED GROUNDWATER CONTROL   45

  The logical place to start this investigation  was to locate the
contaminated medium, i.e., the Lost River. In  this search, an in-
tegrated team approach based on multidisciplinary scientific sup-
port was utilized to address the diverse problems associated with
the fumes rising from the cave system.
  Excavation of sinkholes and crevasses initiated the physical ex-
ploration of the project. In the search for possible entrances into
cave passages,  sinkhole breakdown was removed with a track-
hoe or a backhoe and by hand. Excavations were made adjacent
to homes and buildings at identified channels from which fumes
were  reaching the surface. Excavations also were made within
sinkholes along the hypothesized Lost River flow path and where
historical accounts indicated the closing of cave entrances.
  Concurrent  with these  excavations, exploratory  drilling was
initiated.  However, with  few means of locating drilling sites for
cave passages other than topographic features,  the production of
exploratory wells at 10-ft intervals appeared to  be a  costly and
time-consuming method of cave discovery. Clearly, a method was
needed to locate drilling sites more precisely and economically.
This need led to the implementation of gravity survey techniques.

Gravity Survey
  Gravity surveys are utilized by, and in fact originated in, the oil
industry,  where there is great interest in  detecting  large under-
ground structures.  In microgravity surveys, which  include cor-
rections  for tidal effects, more  sensitive instrumentation and
more precise techniques,  it is possible to  detect very small and
finely localized variations  in  geologic structure.  In Bowling
Green, it was  anticipated that microgravity surveys  would de-
tect the presence of void space (or absence of solid rock) along a
given transect.
  Using a Lacoste and  Romberg  Model D  Microgal  Gravity
Meter, 10-ft stations  were measured  over thousands of  feet of
transects in the study area. The readings were plotted and large
anomalies were investigated by exploratory drilling.  It often was
found that anomalous  readings represented  unusually deep
troughs of soil between pinnacles of limestone  or  break down
zones  of possible cave collapse  (Fig. 3). Nevertheless,  micro-
gravity surveys proved to  be the most reliable  means of locating
void space or cave passages when no historic  or surface indica-
tions were available.
     2000   IBOOcoo  1*001200   ioootoo   «oo   400   zoo

                           Figure 3
   Representative Comparison Between Topographic (Solid Line) and
   Microgravity (Dotted Line) Surveys Showing Probable Breakdown
               Zone (A) and Small Solution Void (B)
  Exploratory and core drilling, utilizing rotary water and forced
air hammer drill rigs, continued throughout the investigation to
verify  the  microgravity data  and to correlate  localized strati-
graphic columns. Many of the exploratory and core wells later
became monitoring wells used to record a wide variety of rele-
vant variables. In certain instances, monitoring wells were placed
by survey and drilled into the stream thalweg of cave passages,
eliminating  the  need for  lengthy  and  labor  intensive  cave
sampling and/or monitoring trips.
  Standard hydrologic methods, including dye tracing and piezo-
metric surface contouring, were used to locate primary flow char-
acteristics, while air monitoring was used to locate and character-
ize areas  of  fume contamination.  A 1.5-in.  downhole camera
with both downhole  viewing  (Lighted  Actual Viewing  Attach-
ment,  LAVA)  and sideviewing (Right  Angle Viewing  Attach-
ment, RAVA) capabilities also were used  to investigate the sub-
terranean environment.
  The investigation team occasionally experienced difficulty  in
distinguishing between solution voids and breakdown zones and
between true void space  and mud-filled cavities.  The downhole
camera allowed visual inspection of these features, ensuring cost-
effective and time-saving verification  of the presence of  cave
passages. Also, the downhole  camera confirmed the indications
of microgravity data and historic  accounts  that a large  cave
existed in  the vicinity of Creason Street and  Robinson Avenue
  This discovery prompted drilling  of  a  new caliber. A 30-in.
core barrel drill rig  was employed to drill entrance wells  into
Robinson's Cave and Creason Cave. These caves were actually
one prior  to  a  1960s  road construction blasting accident which
collapsed  100 ft of road, effectively severing the original  cave
  Utilizing a team of experienced  spelunkers, detailed speleo-
logical investigations and complete surveys of the caves were in-
itiated. From these explorations it was determined that the Robin-
son's and  Creason Cave  system represented  a paleo-channel of
the Lost River,  allowing the  realization that the river was not
confined by the resistant Lost River chart,' but that it had broken
through into what Crawford calls a "Karst Sandwich" (Fig. 4).
                                                                             GRANULAR LIMESTONE
                                                                             (FOSSIL FRAGMENTS t COLITESI
                          Figure 4
  Partial Section Showing Generalized Flow of Lost River into "Karst
               Sandwich" of Chert and Limestone

  Unfortunately, the Lost River proper has yet to be found by
way of Robinson's Cave. The cave does, however, represent a
tributary conduit into the Lost River system contributing con-
taminants from perched water tables and/or soil percolation.
  Another cave  investigation was prompted by fumes rising to
the surface. In one residential crawl space and in an adjacent
storm sewer and drainage well, fumes reached explosive levels
frequently. The investigation team began hand excavation accom-
panied by continuous air monitoring in the drainage well. During
this apparently minor excavation, fumes peaked at levels  above
300 ppm and 100% LEL, prompting a more comprehensive ex-
cavation. Using spark-proof tools, an entrance to a new, hitherto
unknown,  cave was discovered.  This cave, Napier's Cave, be-
came the focal point of further study and investigation. Physical
exploration of this cave required positive pressure flushing of con-
taminant fumes, utilizing two high volume smoke  exhaust fans
blowing into the cave.  Exploration revealed  floating contam-
inants, wall staining and contaminated perched pools, all indi-
cating contamination transport at high and low water flows.
  The ultimate success of determining locations where investiga-
tions could  be efficiently performed, of determining mitigation
procedures  and of initiating possible enforcement actions de-
pended on accurate and valid analysis of the chemicals involved
and of the vulnerability of the environment to their effects. This
information was obtained by extensive sampling and monitoring
of soil, air and water matrices characterizing the contaminants
and affected medias. Potential  and real effects on the environ-
ment, immediate and long-term risk to public health and to the
investigating team were established  and monitored.  Several
methods were employed.
  Air monitoring was performed with the use of real-time, direct-
reading, portable  instrumentation, -photoionization  detectors,
combustible gas indicators and oxygen meters. This highly mobile
instrumentation proved very convenient in screening cave atmos-
pheres and monitoring stations  over the large study area. How-
ever, instrumentation limitations were revealed. The HNU photo-
ionizer, Model PI  101, gave unreliable readings within the cave
environment and in the monitoring wells. This problem was due
to the cave  humidity and subsequent condensation on the ultra
violet light source.
  The GasTech, Hydrocarbon Super Surveyor, Model 1314, was
the most practical  portable instrumentation for use in cave en-
vironments.  It  enabled monitoring  for the parameters of total
organics, combustible gas and  oxygen with one instrument.  Its
light weight  and size, ease in calibration and relative insensitivity
to moisture  made  it the  preferred instrument for  speleological
  Once initial contaminant parameters were identified,  species-
specific instrumentation provided accurate quantification. Specif-
ically, monitoring  was  aided by the use of a Photovac, a pro-
grammed photoionization gas chromatograph. The Photovac was
placed within a crawl space in the Parker-Bennett Elementary
School. Scheduled samples were  automatically collected and
immediately  analyzed before and  during building occupation.
Generated chromatographs provided the specific qualitative and
quantitative  data utilized for public health considerations.
  The U.S.  EPA  Environmental  Response Team's mobile en-
vironmental unit (SCIEX) also was utilized in the study. The com-
puter-coupled tandem mass spectrometer, mounted  in a specially
adapted bus, provided real-time measurements of contaminants
in ambient air from the atmosphere, monitoring wells and within
buildings and homes. This versatile field laboratory was not lim-
ited  geographically and provided  quality assured  quantitative
and qualitative analysis in minutes. It proved to be the most effic-
ient  and accurate means of screening air samples and identify-
ing compounds.
  Tedlar  air sample bags  were used  for capillary soil  gas
sampling and occasionally for atmospheric sampling when sub-
sequent qualitative  laboratory analysis was indicated. The soil
gas technique was employed at industrial sites to establish a corre-
lation between volatile chemicals associated with those sites and
those identified  rising from the Lost River Cave system,  and
where  fumes were detected  in the absence of any surface  evi-
dence of escape paths. In addition to real time air monitoring and
sampling, Tenax collector  tubes were employed, while subse-
quent desorption and analysis provided time weighted average
concentrations for public health considerations.
  Soil,  sediments and water samplings completed  the charac-
terization of chemicals involved and the assessment of this com-
plex problem. Surface and cave sediment samples were collected
from drainage ditches, cave walls and cave  stream beds to iden-
tify contaminants and to establish correlations between industrial
releases of chemicals and cave contamination. Soil samples also
were collected, specifically as a means of documentation during
underground storage tank excavation.
  Due to the scarcity of surface streams in the Bowling Green
area, surface water samples were limited to the Lost River Cave
Entrance and Rise. However, groundwater samples were collected
directly from the caves, monitoring wells and potable wells. Cave
stream and  monitoring  well samples  helped to verify contam-
inant flow paths, while potable well samples were useful in pub-
lic health determination and helped to  locate  contaminated
perched water tables.

  A  priority throughout the  investigation was mitigation  of
immediate and potential threats to health and safety. Prompted
by the  1983 U.S.  EPA identification of leaking underground
storage tanks as a source of Lost River contamination, an exten-
sive  search was performed  to  locate  all storage  tanks in study
areas. Data were compiled to indicate the integrity of the tanks.
Several underground storage tanks were tested and removed or re-
paired if leaks were detected.
  Similarly,  the discovery of one residence that relied entirely on
wellwater and the subsequent determination that this water was
contaminated, led to an extensive data search and public survey
in an effort to identify all water wells in the study area. Residents
on contaminated potable water supplies were connected into the
city water system.
  The desired culmination of mitigation with regard to hazardous
fumes is the  ultimate elimination and removal of those fumes and
their sources. However, prior to the identification of the chem-
ical  nature  of the fumes and their  precise sources, the most
immediate and obvious mitigation procedure was the venting of
homes and public buildings. Using exhaust fans in crawl spaces,
basements and  cave entrances, fumes were dispersed into the
atmosphere instead of being allowed to concentrate in closed and
inhabited areas.

  As a part of an on-going study, the following  conclusions on
particular aspects of the investigation and mitigation of contam-
inated aquifers in karstic terrains are tentative. Based on informa-
tion  derived from this study, the body of this paper might pro-
vide  guidelines for future investigations of  hazardous waste gen-
eration and handling in karst areas.
                                                                           CONTAMINATED GROUNDWATER CONTROL    47

 Ventilation of fumes
   Venting provides a convenient short-term means of lowering
 or eliminating fume  concentrations  in  living  areas  and public
 buildings.  Venting is, of course,  a short-term solution to a long-
 term problem.

   Clear-cut advantages were noted between instruments applied
 to various monitoring tasks. The GasTech proved  superior in
 speleological monitoring due to its relative insensitivity to mois-
 ture and multiple function capacity. Single function instruments
 used in conjunction  with multi-function instruments provided
 data checks.  The  HNU with its increased sensitivity was ideal
 for surficial monitoring of organic  vapors. Similarily, the ex-
 plosimeter's specialized function provided a comparison  and
 verification of data accuracy.

 Microgravity Survey
   Not overlooking topographic features and historic information
 of an area,  microgravity surveys were  the most effective tech-
 nique utilized in this  study for the remote sensing of karstic  fea-

 Smoke Testing
   Although smoke testing met with little success in this investiga-
 tion, we believe that a very dense paniculate and aromatic smoke
 with an efficient injection system could be an effective tool in
 locating small or hidden cave openings.

   Although  rotary water drilling initially was used  as a safety
 precaution to prevent spark ignition of cave  fumes, it is now
 thought that forced air rotary and hammer drilling are acceptable
 methods in all but the most dangerous areas; the forced air would
 flush fumes away from any spark generation.

 Tank Testing
   As national attention to the  Leaking Underground  Storage
 Tank Program and hazardous material handling grows, special
 attention should  be  paid  to karst  areas.  Precision  testing of
 underground  storage  tanks identified  leaking tanks. Their re-
 moval or repair possibly may have eliminated the source of con-
 tamination and the subsequent  fume problem in a  Lost  River

 Chemical Assessment
   The only substance releases whose consequences might be min-
 imized in karstic terrains could be those of acids which rapidly
 neutralize  in contact  with limestone.  In all other instances, oil
 and hazardous substances  releases  in karst areas  can rapidly
 migrate and generally pose more complicated and variable prob-
 lems than in non-karst regions.
  These studies further concluded that all of the water wells in
the study area, with one exception on its boundry, were contam-
inated with a variety  of  hazardous substances. Some of these
contaminants were not present in the Lost River; hence it was in-
ferred that  these contaminants were trapped  in perched water
tables, migrating down well into the phreatic zone.
  As general methods and specific techniques of hazardous waste
management continue to develop and improve across the country,
special attention should be paid to environmental influences on
hazardous  waste dispersion. Karst areas,  in particular,  require
specific guidelines  and possibly  regulative  measures governing
hazardous  material generators, transporters and emergency re-

I. Crawford, N.C., "Sinkhole Flooding Associates with Urban Develop-
  ment upon Karst Terrain: Bowling Green, Kentucky," in B.F. Beck,
  ed., Sinkholes: Their  Geology, Engineering, and Environmental Im-
  pact, A.A. Backema,  Rotterdam,  1984, 283-292.
2. U.S. EPA, "Superfund Cleanup,  Keith Farm Pond, Bowling Green,
  Kentucky," Vols.  I &  II,  1983.
3. Ibid.
4. This section owes much to the formal work  of, and  informal dis-
  cussions with, Nicholas C. Crawford; see footnote I, above.
5. The relative position of the Lost River chart in the stratigraphic col-
  umn is subject to  some disagreement. The Geological Survey places
  the Lost River chart at the base of the St. Genevieve formation; how-
  ever, authorities in the field place the Lost River chart in the Horse
  cave member of the St. Louis formation. Cf. Palmer, N., A Geologi-
  cal Guide lo Mammoth Cave National  Park,  Zephyrus Press, Tea-
  neck, NJ,  1981, 196; Pohl, E.R.,  "Upper Mississippian Deposits in
  South Central Kentucky," Kentucky Acad. of Sciences, Trans., 31,
  1970, 1-15. This placement has  been further confirmed through per-
  sonal communication with Nicholas Crawford.

1. Beck, B.F., ed.. Sinkholes: Their Geology, Engineering and Environ-
  mental Impact, A.A. Backema, Rotterdam, 1984.
2.  "Bowling Green Toxics," U.S. EPA Report, Vol. I., 1985.
3.  Crawford, N.C., "Toxic  and Explosive Fumes Resulting from Con-
  taminated Groundwater Flow through Caves Under Bowling Green,
   Kentucky," unpublished  research proposal submitted  tothe City of
   Bowling Green, KY, April 1985.
4.  Palmer, A.N., A Geological Guide to Mammoth Cave National Park,
  Zephyrus Press, Teaneck, NJ, 1981.
5. Pohl, E.R., "Upper  Mississippian Deposits  in South Central  Ken-
  tucky," Kentucky Acad. of Sciences, Trans.,  31, 1970, 1-15.
6. U.S. EPA, "Superfund Cleanup,  Keith Farm Pond, Bowling Green,
  Kentucky," Vols.  I &  II,  1983.
7. Tejada, S., "EPA Goes Underground at Kentucky Superfund Site,"
  EPA J. 2. July/Aug.  1985, 26-27.

                        Use  of  Low Flow Interdiction Wells  to
                  Control Hydrocarbon Plumes  in Groundwater

                                          John H.  Sammons, Ph.D.
                                       Ninth U.S. Coast Guard District
                                                Cleveland,  Ohio
                                         John M. Armstrong, Ph.D.
                                           The Traverse Group, Inc.
                                             Ann Arbor, Michigan
  This paper presents a case history and discussion of a situation
where the initial hydrocarbon contamination occurred in 1969, re-
mained undiscovered until 1980 and since has resisted massive
efforts to  even adequately define the situation. The  paper de-
tails the installation of low-flow wells across the plume to block
flow of contaminants. This interdiction resulted in a serendipi-
tous activation of the naturally occurring soil microbes in the
sediments down-gradient from the well field. The net result of the
interdiction and microbiological activation was to  effect a rapid
and dramatic reduction in the hydrocarbon contaminants in the
plume down-gradient from the interdiction well field.

  Hydrocarbon contamination of groundwater is rapidly becom-
ing one of the most troublesome and costly problems to be con-
fronted  by both the regulators and the regulated community.
Virtually every day, an article is  written announcing another
underground water supply that has become  contaminated  and
that it appears virtually impossible and prohibitively expensive
to contain and completely decontaminate the aquifer.

  In July 1942, the United States Navy established an Air Sta-
tion at Traverse City, Michigan, a small and isolated community
located in the northwestern section of  the Lower  Peninsula
(Fig. 1). The purpose of the Air Station was to conduct highly
classified research and  development of pilotless  drone aircraft
that could be remotely guided  by television to the target from
chase aircraft. This research effort was continued until 1944 when
it was suspended. When the war  ended, the Air Station  was
turned over to the United States Coast Guard (USCG) to serve as
a major Search and Rescue base for Lakes Superior, Huron and
the upper portion of Lake Michigan.

Coast Guard Operation
  Coast Guard operations at the site commenced in  1944 and
have continued until the present with a mix of rotary-wing and
fixed-wing aircraft. A  review of operations revealed nothing
extraordinary or unusual had been  reported although historical
records of the station are sketchy.

History of Problem
  In 1979, during the removal of two fuel farms preparatory to
the  installation of a new system, there was some indication that
leaks had occurred. The only soil contamination was in the Jet
Fuel (JP-4) storage area. These soils were removed and disposed
of under the direction of the State of Michigan Department of
                         Figure 1
  Location of East Bay Township and U.S. Coast Air Station in Michigan
 Natural Resources (MDNR). This area was located some 1,500 ft
 upgradient and to the north of the area that was ultimately impli-
 cated as the "geographical origin" of the plume. The Aviation
 Fuel (115/145) fuel farm, located  immediately  adjacent to the
 "geographical origin" of the plume,  was excavated at the same
 time with little indication that any unusual occurrences had taken
 place. There was some odor noted in the soil,  but laboratory
 analyses did not confirm gross contamination, so no soil was re-
  In 1979 and  1980, residents in the Avenue E area of the Pine
 Grove Subdivision of East Bay Township complained to the local
 health department that  their  wells were producing  discolored
-water, were foaming and had bad tastes and foul odors. The first
                                                                     CONTAMINATED GROUNDWATER CONTROL   49

 residence reporting the problem is located 1,200 ft to the north-
 east of the Coast Guard Air Station (CGAS). At that time there
 was no explanation for the contaminated wells and, unfortunate-
 ly, the health department did not test for  any of the possible
 organic contaminants. Later in 1980, the MDNR  did a limited
 hydrogeologic study in the area and concluded that the source of
 the contamination was from some unspecified site on the CGAS.'
 In May 1982, the government was notified of the findings and be-
 gan an attempt to elucidate the problem.

   The U.S. Geological Survey (USGS) was retained in June of
 1982 to do a thorough hydrogeological study of the area to define
 the plume of contaminants and to find the source. By April 1983,
 the USGS had determined the direction and velocity of ground-
water flow through the area and  had tentatively identified  the
boundaries of the plume (Figs. 2 to 5). The USGS was not able to
determine the source of the contaminants. They did, however,
conclude that the majority of the contaminants identified were re-
lated to components in fuels with  some chlorinated compounds
also present.1

Detailed Studies
  In November 1983, the Coast Guard contracted with the Uni-
versity of Michigan (UM)  for  a scientific study and  feasibility
study of the site. Objectives for these studies  included a tem-
poral analysis of the plume  and a positive identification and loca-
tion of the contaminant origin(s). The final  report' included a
description of the time variation of contaminants and adsorptive
characteristics  of soils,  including  side-by-side  comparisons of
                                                           n  n w  n to w
                             WEST ARM GRAND
                              TRAVERSE BAY
                                                                             EAST ARM GRAND
                                                                               TRAVERSE BAY
                                                                       study area
                                                                -roo-LINE OF EQUAL ALTITUDE OF
                                                                       LAND SURFACE--Interval
                                                                       20 feet NGVO of 1929
                                                                  10-WATER-TABLE  CONTOUR--
                                                                       Shows altitude of water table.
                                                                       Contour interval 10 feel.
                                                               	»- GROUND -WATER FLOW--Arrow
                                                                       indicates direction of flow
                                                               	GROUND-WATER DIVIDE
                      Base from U S Geological
                      Survey I 62.SOO quadrangles
                                                         Figure 2
                                             General Direction of Groundwatcr Flow

water and soil concentrations found by MDNR, USGS and UM;
a description of the quality assurance program for collection and
analysis of groundwater samples; a preliminary risk assessment
for chemicals found in the groundwater and a description of the
effort to numerically model the behavior of  the groundwater
plume, a description of the groundwater sampling program and a
preliminary  remediation analysis that identified several cleanup
  During this time, the USCG was pursuing an internal investiga-
tion which resulted in the discovery of an Aviation fuel spill inci-
dent. In November or December of 1969, a flange in an under-
ground pipe line under a 115/145 high octane Aviation Gasoline
fueling station failed, resulting in the loss of approximately 2,000
gal of product over a 12-hr period.
  The Traverse Group Inc.  (TGI), an Ann Arbor-based multi-
disciplinary consulting firm, was contracted by the USCG to do a
detailed feasibility study that included  the design,  construction
and operation of groundwater treatment plants (involving carbon
adsorption and air stripping techniques) and to continue a sam-
pling program for selected wells to monitor the plume. This part
of the investigation was completed in February  1985." TGI was
then asked to continue the project to include the following tasks:
• Selection of a specific cleanup technology
• Design, construction  and operation  of an interdiction well
• Design and installation of monitoring wells
• Design, construction,  operation and evaluation of a full-scale
  carbon adsorption system
• Assembly,  installation and evaluation of an advanced rotary
  air stripping and vapor incineration system
                                                          Figure 3
                                  Water Table and Direction of Groundwater Flow on Apr. 5-7, 1983
                                                                          CONTAMINATED GROUNDWATER CONTROL    51

  In February 1985,  new hydrocarbon  contamination  was dis-
covered at the JP-4 fuel farm south of the Hangar/Administra-
tion building (HAB). The four fiberglass underground storage
tanks at the station were tested; three were found to be leaking
and were removed. In March 1985,  routine sampling indicated
the presence of high levels of benzene, toluene and lighter hydro-
carbons in  monitoring wells  along the south boundary fence of
the USCG  property upgradient  from the  JP-4 Fuel Farm. The
origin of this contamination  was not then, nor is it now, under-
stood since these wells are now  clean. One possible explanation
for the contamination in these wells was a spill of approximately
2,000 gal of Jet-A at the Republic Air Lines fuel farm some  1,800
ft directly  upgradient from  the south boundary of the USCG
property.  What  cannot be explained  is the high contaminant
levels when the wells were sampled in March 1985, and the current
absence of contaminants in the wells.

Plume Characterization
  Much of the investigative effort  focused on the location and
characterization of the groundwater plume. All agencies involved
installed and monitored observation wells for this purpose. Due
to temporal variations of the  contaminants and differences in
analytical  techniques, not all results were in complete agreement;
however, there is sufficient similarity to delineate locations and
concentrations of the plume's major constituents.
                                                                            \    EXPLANATION
                                                                            *l\   • «VU iOUtiOi
                                                          Figure 4
                                  Water Table and Direction of Groundwater Flow on July 19-20, 1983

                                                                              LOCATION  OF
                                                                              INTERDICTION WELLS
                                                                                                — 590— WATER-TABLE CONTOUR--Showi
                                                                                                     olliludc ot limuloled -aw table
                                                                                                     Aoril 1383 Contour interval I loot.
                                                                                                     NCVO ol 1929
                                                                                               	»-GROUND-WATER FLDW--Arrow
                                                                                                     indtcotci direction of fto*
                                                                                               	  AIR STATION BOUNDARY

                                                                                                :•:•:•:•  PLUME OF CONTAMINATION

                                                                                                     O      500    ICOOFEET
              Bole odapfrrd Irom U.S Cootl Guard mop and
              mopi bv Gojrdic-Frotcr and Atiocialtf. Inc
                                                                                                        IOO   200 METERS
                                                            Figure 5
                                    Water Table for Apr. 1983, Simulated by Groundwater Flow Model
  The UM study3 identified benzene and toluene as the compon-
ents in the plume presenting the greatest health risk.  The largest
concentrations occurred in the vicinity of the Hangar/Adminis-
tration Building (HAB) at the geographical  head of the plume
(Figs. 6 and 7) although significant amounts of some  compounds
such  as  benzene were found at some distance  downgradient.
Other chemicals also were found in the plume but at smaller con-
centrations and reduced distributions.3
  The USGS provided a  detailed map of the plume's size  and
location (Fig. 1). From the HAB, the plume follows groundwater
flow to the northeast and off the base, passes under an industrial
park and turns slightly north,  narrowing as it passes  underneath
Parsons Road and widening out again  under Avenue E (Fig. 5).
The plume is approximately 4,300 ft long and varies from 180 to
400 ft in width. Its vertical dimension ranges from 25 to 80 ft.
  Small  concentrations of benzene and toluene  have been de-
tected in the water of East Bay. The USGS2  reported maximum
values of 20 /ig/1 benzene  and  3.1 /tg/1 toluene aproximately
330 ft from shore. The vast majority of subsequent measurements
by the TGI have been less than those found by the USGS with the
majority being below the detectable limit.

Soil Contamination
  Both the UM study and the USGS study reported numerous
measurements of organics in the soils  at the Air Station, again
with discrepancies due to analytical, temporal and  areal varia-
tions. The UM study found maximum concentrations of 25.4 jtg/1
benzene, 27.6 /jg/1  toluene and 229  /itg/1 xylene.  Analyses were
done for seven other hydrocarbons with negative results. Analysis
of soil borings indicates that much of the organic material is re-
tained in the soil in  a 6 to 12 in. thick layer in the capillary zone
immediately above  the water table.  It  has been suggested that
this zone is slowly  leaking organics  into the groundwater over
time and is thus serving as a contaminant source for the plume.3

Proposed Cleanup
  After consideration  of the various  long-term treatment  or
cleanup options available, it became clear that the most logical
first step in any remediation action program would be to decrease
or stop  the further  movement of contaminants off U.S. Coast
Guard property.
  This option was judged to have several advantages:
• Reduce any possible increase in risk to human populations that
  may have been related to fuel based contaminants present in
  the groundwater.
• Promote reductions  in  contaminant  concentrations  in  the
  groundwater either by dilution or possible biodegradation of
  fuels by indigenous microbial consortia present in  the subsur-
  face soil-water  system.
• Provide a better  opportunity (e.g., more time) to efficiently
  select and design appropriate method(s) for dealing with the
  contaminants present in the geographical origin of the plume.
                                                                             CONTAMINATED GROUNDWATER CONTROL    53

   /r-   »S20   »SI7
system to withdraw the minimum amount of water yet still cap-
ture the contaminants passing through the  aquifer so that no
contaminants,  or at least a significantly reduced amount, were
leaving government property.
  After reviewing several of the USOS groundwater model runs,
the team selected a well field configuration shown in Figure 8.
Seven wells were located laterally across the plume in the east-
northeast area  of the Air Station. These wells were 6 in. auger-
drilled wells with full 10-slot stainless  steel screens running from
the top of the  aquifer to the clay confining layer at the bottom
(Figs. 5 and 8).
  The system initially was constructed  with  all elements  above
ground. In 1985, the interdiction system was modified by install-
ing pitless adapters and providing a heated building to protect the
manifold and control systems from the extreme cold experienced
at this site. The water produced from  the interdiction wells is
piped to a carbon treatment system consisting of four 20,000-
Ib carbon reactors. The carbon reactors  were specified to reduce
the levels of benzene and toluene in the water to less than  1 /ig/1
as  measured using  headspace technique on  a HP5710A  Gas
  The USGS model gave clear indications that these flow rates
would produce closure of the equipotential lines at the interdic-
tion well line. One difficulty was that the two-dimensional nature
of the model did not yield any prediction on the vertical move-
ment of water at or  near the individual well locations. However,
because of the uniform nature of the saturated zone and the rapid
mean field velocity of the aquifer, which was measured at 5 ft/
day, it was believed that vertical movement in a fully screened
well at 15 gal/min would be sufficient to capture contaminants
flowing in the lower areas of the aquifer.
                           Table 1
         Specifications for Groundwater Interdiction Systems
North Interdiction Field
Date Wells Installed:
  Wells ID 1-ID6      Feb. 26-Feb .28,1985
  WellID-7          Mar. 22, 1985
Date of Water collection
  system construction: Mar. 19-Apr. 18,1985
Date Pumping
  commenced:        Apr.  19, 1985
Well and Screen Data:

                       North Interdiction
                        (6 in. diameter)
                              Table 2
                Monitoring Wells for North Interdiction
We) 1 *
Flow cpm
Top of
uel 1

[ bottom
of pump
Below ;
ground ',
level) 1
.' ID-1
; is




10-2 i 10-3
; is : is

13 ' 13

53 : 43

43 ; 33













[0-7 \
25 ;

13 :


48 :

  Five full screen wells installed—Mar. 20-27, 1985
  Downgradient of North Interdiction Field on Mar. 20-Mar. 27, 1985
  Four-inch diameter wells extending to clay liner beneath the aquifer
Uell «
Depth below
well screen to
bottom of well
screen { top of
water at 13'
below ground
Vertical loca-
tion of sample
Dumps in we 1 I s .
feet below
Sample pump HI


25 '8"
35 '8"
54 '5"


66'9" ;


14 '6"


4 '


i?' :
33' :
43' :
 TGI-2 sample well located 50' downgradient from ID-3 on a line parallel to the flow lines of the
 plume drawn NE from ID-3. TGI-3 located 50' downgradient from TGI-2 on a line parallel to
 the flow lines through ID-3 and TGI-2. TGI-4 located 90' northwest of TGI-3 on a line through
 TGI-3 perpendicular to the flow lines. TGI-1 located 180' northwest of TGI-3 on a line through
 TGI-3 and TGI-4. TGI-5 located at the end of a line drawn perpendicular to a line through
 TGI-3 and TG1-4 starting at a point 90' SE of TGI-3 running 50' NE.
   In most situations, contaminants such as benzene and toluene
 are found more predominantly in the upper areas of the aquifer.
 However, in this  situation the original spill had been  in  the
 ground so long that visible product was no longer present and
 the product not contained in the interstitial area of the capillary
 zone had solubilized and had mixed downgradient so that it was
 present throughout the vertical cross-section of the aquifer. Thus,
 placement of the pump in the well had to be in conjunction with
 a full screen configuration to capture contaminants from all ver-
 tical locations.
  A second consideration in the design was the monitoring of the
 interdiction system to  obtain data on  the effectiveness of  the
 blocking  of contaminant flow. Since the area of the  aquifer
 downgradient from the interdiction system had been contam-
 inated for several years, the monitoring wells must yield  data in
 sufficient detail to prove any decrease in contaminant levels after
 the "valve" had been closed.
  A sampling network of five  wells was located downgradient
 from the interdiction well system (Fig. 9). The wells were  located
 outside the zone of influence of the interdiction wells.  Sampling
 was done using a dedicated "Well Wizard" sample pump system.
 Groundwater samples were collected following standard field pro-
 cedures for the collection of volatiles and analysis was done with-
 in a few  hours of collection, always on the same day that  the
 sample  was collected. Five sample pumps were located in each
 monitoring  well at the top  and bottom of the saturated zone
 and at three equidistant points (Table 1).

  The results downgradient from the interdiction  system have
been satisfying to date. Toluene and benzene levels in the down-
gradient monitoring wells were monitored in one of the  five wells
(M2/TGI2)  on a daily basis for the first 8 weeks and on a bi-
weekly basis thereafter. The  other four wells are monitored once
a week.
                                                                             CONTAMINATED GROUNDWATER CONTROL     55

  Since benzene is the more mobile of the two constituents of in-
terest and has  the higher solubility, it should be slower than
toluene to respond to the interdiction. In addition, there was
speculation that some biodegradation by selective demethyoxyla-
tion might be seen and be reflected in  the benzene and toluene
concentrations  found in the downgradient wells. A considerable
amount of data have been collected which allow trend  observa-
tions to be made regarding the fate of benzene and toluene in the
downgradient area (Figs. 8 to 10).
  Figure 8 shows the variations and trend of benzene concentra-
tions as a water column average by month since the interdiction
field was activated on Apr. 15,  1985. Figure 9 provides  the same
information for toluene. These  data are from Well M-2 (TGI-2)
located immediately downgradient and outside the zone  of influ-
ence of the  interdiction well  field. This well logically would be
the first monitoring well to provide data on  the effectiveness of
interdiction. Although microbiological activity would  not be sep-
arated from hydraulic blocking, toluene levels decreased from a
baseline level of 10329 jig/I to less than  10 /ig/1 in  approximately
 100 days.
  Figure 10  shows benzene  and  toluene levels plotted against
time from April to December 1985. Since this figure shows the
variation of the two compounds, it could substantiate the micro-
biological activity. The decrease in toluene levels, together  with
the concomittant rise in benzene  between May and July, could
be attributed to the demethoxylation of toluene to  benzene. Both
compounds then began a steep downward trend with  the toluene
being the first  to approach non-detectable levels  in August. As
can be seen, the benzene reached its lowest level during October.
  The increase  in benzene levels since October is attributed to the
appearance of a red slime that plugged  pump screens  and coated
the inside of the piping systems,  resulting in a reduction of the
hydraulic capacity of the system. The groundwater in  this aquifer
is high in iron  content, and  the  substance has been tentatively
identified as oxidized iron compounds  and  complexes, mineral
deposits and biomass. Microbial action is believed to play a signif-
icant role in the dissolution  of these solids in the aquifer and
subsequent precipitation within the system. The slime is complete-
ly soluble in  an acidic solution, and gamma  logging of the wells
does not indicate the presence  of any clay layers that  could be
contributory. Obviously this situation must be better  understood
and controlled  or the field will lose its effectiveness as may be in-
dicated by the gradually increasing benzene levels.
  The USCG has an active project with  the U.S. EPA Robert S.
Kerr Environmental Research  Laboratory (RSKERL)  in Ada,
Oklahoma to evaluate the soil and water chemistry as well as the
microbial consortia present at the site. The results of these studies
should provide a mechanism to control the slime and  elucidate
the biodegradation phenomenon.
  While monitoring of the interdiction system continues, there al-
ready is substantial evidence that  the desired effect was achieved
in terms of contaminant containment. Toluene has decreased to
virtually non-detectable levels and, while it continues to be vari-
able, benzene concentrations clearly demonstrate a downward
trend.  The benzene variability may be due in part  to some, all or
a portion of the slime problem  discussed earlier. Biodegrada-
tion variability, the effect  of recharge and the higher  mobility of
benzene  in the groundwater  system also may contribute to the
variability of the benzene.

  Based  on the data collected over the past 9 months, it appears
that, for well defined plumes,  low flow interdiction is a viable
alternative to intercept and block contaminant flow.
                                                                                         HELL M-2
      Hit   a*   ui  JJ<   4J.I  U  OT'   Kl  •>•   OK

                           Figure 8
           Benzene Concentration as a Function of Time
      iiv   *•   wt  JLM  juf   »4   ari  XT   «*   ore

                          Figure 9
           Toluene Concentration as a Function of Time
     ST   **»    ui   JU<   all  14*   apt   XT   4*   OCC

                          Figure 10
      Benzene and Toluene Concentrations as a Function of Time
  Careful attention must be paid to the  mechanical as well as
contaminant systems. The appearance of  the biomass is a good
example of this since no homeowners in the area using wells have
reported a problem with slime.
  The possibility of triggering increased microbial activity is a
potential factor to consider.
  Particular attention must be given to the vertical  distribution
of contaminants since withdrawal volumes may be significantly
different for new plumes as compared to old plumes.
  Minimum volume requirements as dictated by available process
water disposal  options  are a major  determinant  in  interdiction
system design.

  The opinions or assertions contained herein are the private
ones of the writers and are not to be construed as official or re-
flecting the  views of the Commandant or the Coast  Guard at

1.  Sibo, K., "Groundwater Investigation of East Bay Township," Un-
   published, Mar. 3, 1982.
Twenter, F.R., Cummings, T.R. and Grannemann, N.G., "Ground-
water Contamination in East Bay Township, Michigan," U.S. Geo-
logical Survey Water—Resources Investigations Report 85-4064, 1985.
Rossman, R., Rice, C.P., Hartung, R., Simmons, M.S., Armstrong,
J.M. and  Wright, S.J.,  "Scientific  Study and Feasibility Study
Groundwater Contamination at Traverse City, Michigan,"  Unpub-
lished, The University of Michigan, Ann Arbor, MI, Mar. 1985.
"TGI Final Report, Phase I," Unpublished, Apr. 1985.
                                                                            CONTAMINATED GROUNDWATER CONTROL    57

                 Computer  Groundwater Restoration  Simulation
                                 at  a  Contaminated  Well Field

                                            Shih-Huang Chieh, Ph.D.
                                        Ecology and Environment, Inc.
                                                Buffalo, New York
                                            Jeffrey E. Brandow,  P.E.
                               State Department of Environmental Conservation
                                                Albany, New York
  The most obvious and direct  means of contaminated aquifer
rehabilitation  is to remove the contaminated water by pumping
from recovery wells. Factors affecting the technical and economic
feasibility of this procedure include the number of recovery wells,
the rate and duration of pumping and the method used to dispose
of  the contaminated  water.  Since the aquifer rehabilitation  is
assumed  to be the reverse of the contaminant spreading process,
the theory and models that have been developed for predicting the
spread of a contaminant should provide a basis for predicting
contaminant withdrawal.
  The drinking water of western  Vestal, New York, is supplied by
Water District 1, which consists of wells  1-1,  1-2 and 1-3. Well
1-1, located on the bank of the Susquehanna River about 400 ft
from the Endicott-Vestal Bridge, is contaminated with volatile
organic compounds. Currently, a remedial investigation/feasibil-
ity  study is being conducted at the  site. The purpose of the
remedial  investigation/feasibility study is  to locate the contami-
nant source, to define  the extent of the contaminant plume and to
evaluate the remedial alternatives. As a practical tool to assist in
the feasibility study, a  two-dimensional finite element model is be-
ing applied to this site. First, the model is calibrated against field
data. Then the model  is utilized to simulate the movement of the
contaminant plume under the operation of a recovery well. The
model results  will be  used to evaluate and refine the  remedial

  A groundwater contamination problem due to toxic organic
chemicals exists in Water District 1 of Vestal, New York. In early
1980, organic chemicals were detected in a municipal well in
Vestal. As a result of  this finding, all wells in Vestal were tested
for synthetic organic compounds. Well 1-1 of Water District 1
was found to contain large quantities of organic compounds. This
triggered  an assessment study of the potential of the pollutant to
contaminate other water supply wells in the district. Currently, a
remedial investigation/feasibility study is being conducted at this
contaminated  well field site. As a result of the remedial investiga-
tion, the contaminant  source areas were identified and the extent
of the contaminant plume was defined.1'2
  A feasible alternative to cleanup of the contaminated well field
is to remove the contaminated groundwater by pumping from
recovery wells. Since the aquifer rehabilitation is assumed  to be
the  reverse of the contaminant spreading process, the theory and
models that have been developed for predicting the spread of a
contaminant  provide a  basis  for  predicting contaminant
  A two-dimensional finite element model is applied to this con-
taminated well field site to simulate the transport of the pollutant
under a recovery well is being operated.  The model utilized a
quadrilateral element and the Crank-Nicholson  method  for
calculation of the time marching scheme. The model result can be
used to evaluate the effect of the recovery well and to refine the
remedial alternatives.

  The drinking water for most of western Vestal is supplied by
Water District 1, which consists of well I-I and water supply welk
1-2 and 1-3. Well 1-1, located on the south bank of the Susque-
hanna River about 400 ft from the Endicott-Vestal Bridge, is con-
taminated with volatile organic compounds.  Contamination was
detected first in early 1980; since spring 1980. the well has been
pumped directly into the Susquehanna River. At present, well 1-2
is supplying the Vestal drinking  water;  well 1-3  is serving as a
backup supply. Because the water in well  1-3 is highly corrosive
and well 1-2 has a limited capacity, it is important that well 1-1 be
restored, especially  to meet anticipated future peak demands for
Water  District 1 and other interconnected districts.  Figure 1
shows the site of the contaminated well field and the groundwater
table contours  under wells 1-1 and 1-2 while in operation. The
groundwater table contours were developed from field measure-
  The area immediately surrounding  well 1-1 is within the Sus-
quehanna River floodplam and consists of marshland, wooded
areas and commercial and residential land use areas. The river oc-
cupies a bedrock  valley filled with glacial deposits ranging in size
from clay to gravel. The bedrock surface, situated between 130
and 160 ft below ground level, is overlain by a productive aquifer
consisting  of  about 40  ft of permeable sand and gravel. The
aquifer is overlain by 120 ft of poorly permeable clays and silts.
Within these deposits are locally permeable pockets of material
which apparently have acted as contaminant pathways between
the ground surface and the producing aquifer.
  The high rate of pumping from this aquifer (approximately 1.5
million gallons daily, including the pumping of well 1-1 to waste)
has induced organic solvent movement toward well 1-1 from the
source area.  The source  area is an industrial park containing a
variety of industries. Volatile organic levels exceed 10 mg/1 in one
monitoring well located  within the area, but no information is
available regarding spills or disposal or organic solvents.

  The governing equation of pollutant transport in porous media
is the following:

                                                           Figure 1
                       Map of Vestal Well Field and Groundwater Table Contours with Wells 1-1 and 1-2 in Operation
   3c    f 3  /    3c      3c \   3 /    oc       3c\l
   — = n\ — \D,,, — + £>xv— +— £v,- — + DVV —
   3t    lax \    3x    y3y/  3y\  yx dx    yy 3y/J
         r 3         3
       - —  (Vxc) + —
         Idx         3y
in which:

    R = 1  + esA://z
      Dxv, D
            •yx>   yy
             V   V
             'x>  'y
                      = concentration of the pollutant
                      = retardation factor
                      = bulk density of the porous medium
                      = distribution coefficient of the
                      = porosity of the  porous medium
                      = component of dispersion tensor
                      = Darcian velocity component
                      = 1st order decay constant
                      = concentration of the source fluid
                      = flow rate of the source fluid
Let L denote the operator on c in equation (1), thus:
Lc = 0.
For the finite element method:
                 {N}T{c}                                 (3)

where £/V} is the shape function. The residue becomes Lc. The
principle of Galerkin method requires that:
           = 0 .
  Equation  (4)  states  that  the  inner  product of L? and the
weighting function Wj over the solution domain B vanishes. The
integrations  of Equation (4) are now carried out, and the results
are expressed in matrix and vector notation as follows:4
                                                                 {c}T[Af{c} +(K-A+E) {c} -{p}] =0
                                                                 in which:

               —  — •}  +nDxy  — •
               dx   3         x
K   -//  [n£)x
                 + nDyy  {	\l	
        E   =//   n\R{W}{N}'* dxdy

        M   -ff   nR{W}{N}'T dxdy
                                                                 {p} =      Qc*{A'}dxd;y+
                                                                                                   3c        3c
                                                                                                 xs — + "-D^y	VjcC Inx
                                                                                                   ax        ay       '
                                                                                   (3c       3c      \   "1
                                                                              nDyx — + nDyy -- VyC rty   { W } dS .
                                                                                   dx       3y      /   J







The advantage of the finite element method over the finite dif-
ference method is its ability to handle complex boundaries and
normal derivatives. In the time dimension, these advantages are
not present, so the finite difference method will be used for the
time derivative term of the solute transport Equation.5
  With finite difference  in time, Equation (5) becomes
                                                                            CONTAMINATED GROUNDWATER CONTROL    59

                                                                                              Table 1
                                                                     1,1-Dichloroelhane Concentration Measurements at Vestal Well Site
 in which ^ =  1  for backward-differences and /i  = 0.5 for the
 Crank-Nicholson differencing scheme. Along the boundary of the
 study area, Dirchlet boundary conditions were specified.
   For the preparation of input files for the computer simulation
 study, it was necessary to set up a grid system for the discretized
 site and determine several model parameters. According to the
 site history, production wells 1-1 and  1-2 are constantly in opera-
 tion. Therefore, a non-uniform but steady-flow field  was as-
 sumed from the measured ground water table under wells 1-1 and
 1-2 while in operation'  as presented in  Figure  I. As shown in
 Figure 2, the Vestal well site is discretized into 103 quadrilateral
 elements with 125 nodals.
   Among  the dissolved organics  in groundwater,  1,1-dichloro-
 ethane was chosen to model  the groundwater solute  transport in
 the study  area.  The  measurements  from the  samples  at  the
 monitoring wells are summarized in Table I. The locations of the
 monitoring wells are shown in Figure 1.  The 1- series wells were
 installed during the assessment study. The s-  series wells were in-
 stalled during the remedial investigation  by E & E.
   Since  no field measurements are available for the dispersion
 coefficient, this value was selected based on the literature. The
 molecular  diffusion is assumed to be  small compared  to  the
 hydrodynamic  dispersion. Values of 205  ft  for  longitudinal
 dispersivity and  12.9  ft for  transverse  dispersivity were used.
 1,1-dichloroethane could be  adsorbed on the solid phase of the
 aquifer.  However, the retardation effect is believed to be small
 and will not seriously affect  the qualitative prediction of solute
 transport.  Therefore, a retardation factor of 1 was used for this
 study. The values of model parameters are listed in Table 2.



                            Table 2
           Model Parameters Used in Computer Program
  Number of elements
  Number of nodes
  Number of we)It
103       Longitudinal dliperslvlty
125       Transversal dlsperstvlty
7        Retardation factor
  Hydraulic conductivity  47 ft/day  First order decay constant
205 ft
12.9 ft
  In order to verify the dispersion coefficient chosen for  this
study,  a simulation of the dispersion of 1,1-dichloroethane  was
conducted. Two constant-strength sources located at nodals 84
and  94 were assumed in this computer run. The contaminant
sources were found during the remedial investigation. The simula-
tion  time was 6 years. The assumptions were based on our best
knowledge of the history of the site. The computed contaminant
plume  and comparison with  field  measurement  are presented in
Figure  3. The results show that  the simulated  concentration is
very  close to the field measurements. Other values of longitudinal
and lateral dispersivity also were selected in simulating the disper-
sion  of 1,1-dichloroethane. No satisfactory' results were obtained
from these tests. Therefore,  the longitudinal and lateral disper-
sivities selected for the current study represent a rational choice
for this site.
                                                            Figure 2
                                         Finite-Element Grid System for the Vestal Well Field


1597 Slmul.i.d Cone.
11300) M.I,
ltd Concin
                                                          Figure 3
                                         Computed Contaminant Plume and Comparison
                                                   with Field Measurement
  Since spring 1980, well 1-1  has been pumped and discharged
directly into the Susquehanna River. A possible remedial action
would be the removal of the contaminant source and would con-
tinue use of well 1-1 as a recovery well to withdraw the pollutant
from the aquifer.  Computer simulation runs were conducted to
simulate  the movement of the  contaminant  plume after the
removal of the contaminant source and while well 1-1 was in
operation. Figure  4 presents the simulated concentration at the
end of 5 years. A comparison of Figure 4 with Figure 3 shows that
the 1000-/tg/l iso-concentration contour is moving toward well
1-1, and the iso-concentration area is shrinking. Figures 5, 6 and 7
present the simulated concentration at  the end of 10,  15 and 20
years, respectively. It can be seen  clearly in these figures that the
contaminant plume is moving toward recovery well 1-1 and that
the high concentration region is shrinking.
                                                          Figure 4
                                           Computed Contaminant Plume, 5 Years After
                                                Removal of Contaminant Source
                                                                          CONTAMINATED GROUNDWATER CONTROL    61

                                                          Figure 5
                                          Computed Contaminant Plume, 10 Years After
                                                Removal of Contaminant Source
   The model predicted the movement of the contaminant plume
 under the operation of well 1-1 and the removal of the contami-
 nant sources. The prediction was made for the time periods of 5,
 15 and 20 years. The study provided a basis to evaluate the effec-
 tiveness  and the cost of using well 1-1  as a  recovery well. This
 summary paper demonstrates  the usefulness of  the computer
 model in assisting the feasibility  study at uncontrolled hazardous
waste sites. Further  applications  of computer  groundwater
restoration  simulation can be conducted by varying the pumping
rates, locations and numbers of recovery' wells.

  The authors would like to express their appreciation to U.S.
EPA and  the  New  York Slate  Department  of  Environmental
Conservation for their support in this study.
                                                         figure 6
                                          Computed Contaminant Plume, 15 Years After
                                                Removal ol Contaminant Source

                                                              Figure 7
                                             Computed Contaminant Plume, 20 Years After
                                                   Removal of Contaminant Source
1.  Ecology  and Environment,  Inc., "Town of Vestal  Water District
   No. 1 Focused Feasibility Study," May 1985.
2.  Chieh, S.H., Cook, D., Hwang, J.C. and Brandow,  J., "Computer
   Model Study of a Contaminated Well Field—Groundwater Pollution
   Source Identification and Contaminant Plume Simulation," Proc. of
   the 1985 ASCE Hydraulics Division Specialty Conference, Orlando,
   FL, 1985, 330-335.
3.  Martin, R.J., Coates, D.R. and Timofeefe,  N.P., "Well Field Con-
   tamination Investigation for Town of Vestal Water District No. 1,"
   Report for New York State Department of Environmental Conserva-
   tion, 1983.
4.  Yeh, G.T.  and  Ward, D.A., "FEMWASTE:  A  Finite-element
   Model of Waste Transport  Through Saturated-unsaturated Porous
   Media," Oak Ridge National Laboratory Report No.  5601, 1981.
5.  Hwang, J.C. and Koerner,  R.M.,  "Groundwater Pollution Source
   Identification from Limited  Monitoring Well Data—Part 1, Theory
   and Feasibility," J. of Haz.  Mat,, 8, 1983, 105-119.
6.  Bear, J., Hydraulics of Groundwater, McGraw-Hill, Inc., New York,
   NY, 1979.
7.  Gray,  W.G.  and Hoffman, J.L.,  "A  Numerical Model Study of
   Ground-Water Contamination  from Price's Landfill, New Jersey-II.
   Sensitivity Analysis and  Contaminant  Plume Simulation," Ground
   Water, 21,  1983, 15-21.
8.  Finder, G.F., "A Galerkin-Finite  Element Simulation of Ground-
   water Contamination on  Long  Island, New York," Water Resources
   Res., 9, 1973, 1657-1669.
                                                                               CONTAMINATED GROUNDWATER CONTROL    63

                          Enhancement of  Site Assessments by
                                      Groundwater Modelling

                                            Joseph R. Kolmer,  P.E.
                                            John B. Robertson, P.G.
                                               Roy F.  Weston, Inc.
                                           West Chester, Pennsylvania
  It is a commonly  held misconception that extensive ground-
water monitoring and  other geologic and groundwater related
data must  be  available before a groundwater model  can  be
selected, calibrated and implemented. The primary use of models
in this mode of operation has been to predict long-term ground-
water flow  and groundwater quality due to implementation of
engineered groundwater supply systems and/or remedial cleanup
  Advances in computer technology have dramatically revised the
utilization of groundwater models and expanded their utility with
respect  to  groundwater investigation  programs. Groundwater
flow models are now  readily adaptable to personal computers. To
use these models in a site  investigation, only a very  limited
amount of information is needed initially, and assumed data can
be used to supplement this initial  information.
  Given the limited data, the model is responding primarily to the
general equations of groundwater flow and not to unique site con-
ditions.  This uncalibrated flow model, however,  provides the
technical skeleton to which additional  groundwater monitoring
information can be input and assessed. As additional site data are
added to the model,  it is refined and calibrated. During the data
gathering process, the model is used not only to identify data
gaps,  but  also to determine when  sufficient  data  have  been
gathered to  quantify  site conditions.
  The groundwater model is not limited to utilization as a predic-
tive model based upon exhaustive site information. Rather, the
groundwater model is now most properly  and  cost effectively
used in the groundwater monitoring program to assess data gaps,
quantify  site  conditions  and evaluate engineered remedial

  The purpose of this paper is to discuss and present a method of
site  investigation using groundwater  computer  models to guide
the  investigation as well as to indicate when sufficient data have
been collected. The methods used to implement this approach are
neither unique nor  technically complex, but  they do  require
utilization of an integrated geoscience and engineering staff. This
technical integration,  plus the general availability of groundwater
models on microcomputers, makes this approach to site investiga-
tion technically feasible and advantageous.
  The need for quantitative groundwater  resource  assessment
tools is growing at an  accelerating rate. Water resources investiga-
tions for domestic, agricultural and industrial uses are increasing,
and the investigation work associated  with definition of con-
taminated groundwater due to hazardous materials is expanding
daily. These latter investigations are particularly difficult because
they require quantitative analysis methods to define the contami-
nant migration conditions as well as the remedial solutions. The
use of groundwater flow models to quantify the site investigation
work and standardize the basis upon which remedial systems are
comparatively evaluated makes use of the inherent attributes of
these methods.  A brief discussion of the current practices in
groundwater modelling is presented below followed by a detailed
discussion of how models can be used in site assessments.

  The predominant practice for projects involving groundwater
quality and water resource investigations is illustrated in Figure 1.
The objective or the field investigation work is to obtain sufficient
information to describe site conditions, specifically the geologic
conditions. After these data are gathered, they are assembled and
assessed to determine their adequacy.
                         Figure 1
       Silc Investigation Followed by Groundwater Modeling

  This data evaluation exercise might reveal the need for addi-
tional information to describe an anomalous or vague site condi-
tion, and a second phase of field investigation would commence.
This interactive process would continue until the  objective of
understanding  the geology, ground water hydrology  and  water
quality conditions is accomplished. A report describing the in-
vestigative procedures  and the defined site conditions is finally
  After  the  field investigation is complete,  the  engineer or
engineering  group receiving the final report develops a set of
remedial action alternatives  designed to mitigate  any adverse
chemical quality conditions identified during the field investiga-
tion. To conduct a comparative analysis of the identified remedial
alternatives, a groundwater flow model is usually employed.  The
flow model is  calibrated based  upon the findings of the field
report and run (exercised) to satisfy the identified  needs of the
engineer. If, during the conduct of the engineering  analyses, the
need for additional data is identified, it usually is assumed as  part
of the model input. The mathematical operations inherent to the
flow model check these assumed inputs against measured site con-
ditions. Thus, assumptions made at this stage of a site investiga-
tion are not totally unchecked, but they usually do  not have the
advantage of being field verified.
  After the site model has been developed and calibrated, it is ex-
ercised to simulate  the  various remedial solutions being  con-
sidered for the site. For example, a groundwater barrier wall, such
as a soil bentonite slurry cut-off wall, could be simulated by the
insertion of no-flow boundaries around the perimeter of the  site.
Guswa1 noted that implementation of a remedial action plan can
be very expensive, and it is desirable to have the preimplementa-
tion assurances that the proposed plan can be effective. While no
type of modelling provides guarantees that any particular plan
will be effective, they do provide a comparative basis against
which all remedial alternatives can be evaluated.

  The use of the groundwater model is a good tool to evaluate
various remedial solutions. This use of the model is obviously ad-
vantageous, but additional information can be obtained. That is,
the groundwater model can be used not only to evaluate the facts
gathered after conclusion of the field investigation, but also can
be used to better define what information needs to be gathered
and where it can be found.
  Figure 2 illustrates how the groundwater  flow model can be
used during the site investigation phase of work and in the subse-
quent analysis of remedial alternatives. Comparison of this figure
to Figure 1 shows that the steps leading to the final report have
been reduced. This reduction primarily centers around integration
of  the groundwater modelling work into  the analysis of data
obtained during the field investigation exercise.
  To initiate use of  groundwater modelling during the field in-
vestigation phase of any project, a basic amount of site informa-
tion must be known.  General site geology,  basic  groundwater
flow conditions and surface water conditions should be known to
establish  the initial flow parameters and boundary conditions for
the model. Published literature is replete with geologic definition
and groundwater ami surface water factors for a large portion of
the United States.  If an assessment of existing information on a
site area is conducted in a fairly exhaustive manner, the general
information needed initially to establish the groundwater model
usually can be accumulated.
  One of the primary questions raised during conduct of a field
investigation is how does one know when  sufficient data are
available to satisfy the objectives of the project? The  groundwater
flow model  can provide significant assistance in answering  this
question. As indicated in Figure  3, in order  to  assemble or
                          Figure 2
                  Integrated Site Investigation

establish the groundwater model, existing data as well as input
from numerous technical disciplines must be obtained. Questions
as to the area to be covered by the model, the geologic conditions
to be  considered and the  overall engineering requirements
associated with the model must all be discussed and incorporated.
Thus, at this early stage of the project, the flow model has in ef-
fect served as a catalyst for integration of the technical disciplines
that eventually will be needed to resolve the identified problem.
This integration,  properly carried to fruition,  will result in the
establishment of a set of objective criteria for the site investiga-
tion and the overall remedial action study. Specific data needed to
satisfy these criteria would be identified for acquisition during the
field investigation.
  As field investigation work is conducted, new site information
is developed and  can be used  with the established groundwater
model. As these data are added, the aquifer characteristics are ad-
justed so that the model representations are consistent with iden-
tified field conditions. As more information is gathered in the site
area, the adjustments that need to be made to the model will
       Calibration •
                          Figure 3
                Groundwater Model Development
                                                                            CONTAMINATED GROUNDWATER CONTROL   65

become relatively minimal. As this point is approached, acquisi-
tion of additional field information is not needed to refine the site
model, and field investigation work is complete. Site conditions
are now defined consistent with the set of objective criteria iden-
tified for the project.  Assessment of remedial alternatives can
   The model assembly, identification of additional  data  needs
and model revision work tasks identified in Figure 3 are in  effect
the data analysis  and  additional data input loop identified  in
Figure 2. The exercise  of  identification of additional data  needs
and acquisition and  input of these data to the  flow model con-
stitute calibration of the flow model. An important distinction be-
tween this approach to use of groundwater modelling and the
approach discussed  in the previous subsection  is that  model
calibration is integral to the field investigation exercise.
   Finally, no (or at  least very little) assumed data are needed  to
calibrate the groundwater model.  Thus,  the flow model that
ultimately will be  used to  compare various remedial alternatives
already  will  have undergone the  rigors  of development and
calibration placed upon it not only by the simulation of known
site conditions, but also by the technical requirements of the disci-
plines involved in  overall problem resolution.

   The principal concerns surrounding the use of computer model-
ling  have  been  its  availability  to  the  project's principal
investigator and the cost associated with development and opera-
tion of these models. Over the past few years, both of these con-
cerns have been significantly diminished by  the development  of
the desk top personal computer. Thus, the principal investigator
or project manager has  a significant  library  of groundwater
models readily available to him that can be developed and exer-
cised at a  relatively  inexpensive cost.  In  fact,  the graphics
capability as well as the data base management  capabilities  of
these machines can reduce the overall project costs compared  to
hand manipulation and graphics development for a  medium  to
large scale site investigation.
   In  the past, groundwater flow models have been misused by
misrepresenting their capabilities  and  application. One of the
common problems associated with groundwater modelling  is the
tendency to use the most sophisticated model available for  prob-
lem resolution. This  temptation is difficult for the technical com-
munity to resist, but can be managed if integration of technical
skills is accomplished in the early stages of the project. This in-
tegration  of  the  groundwater  modelling  expertise with the
pragmatism of the engineer usually results in  model development
consistent with the identified objectives for  the site study.  This
management of technical  expertise must be  coordinated by the
project's principal investigator.
   Another technique that can be used to help prevent  over exten-
sion and misrepresentation of groundwater  model usage is the
phased implementation of  a particular model's capability. For ex-
ample, throughout this paper, only groundwater flow  models
have been discussed  and solute transport models have not been
strongly  considered. This analysis  does  not  say  that  solute
transport models are not appropriate for the problems discussed
herein; it does, however,  say that in the early  stages of model
development  and utilization in site investigations, the first step
should be a  characterization  of  groundwater flow  hydrology.
Once   groundwater  flow  characteristics  (especially velocity
parameters) have been identified, solute transport models can be
included. One must remember that the model only need to detailed
enough to satisfy the  stated objectives of the project.
  The following brief example shows how various data elements
can be obtained and input for site investigation modelling. The
advantage of a phased field investigation program is highlighted
in the example.

Existing Data  Review and Model Development
   During the review of existing information for a site area, data
concerning the history of operation of the site, the geologic and
geohydrologic conditions of the area, precipitation, surface water
hydrology and so forth, are compiled. The completeness of this
information depends upon the location of the site, the notoriety it
has received, site management and numerous  other factors. For
the purposes of this paper, only that  available data pertinent to
development of the groundwater model will  be discussed. The
lack of consideration of any other data does not mean it is not
needed for a site investigation nor does it mean it may not be per-
tinent to model development at some  sites.
   Figure 4 presents  a map of the example site area and contains a
significant amount of information. The example site is located in
a river floodplain below an upland area. Thus, the aquifer of con-
cern is  more than likely alluvial in nature, probably containing a
wide distribution of grain-size materials. In addition, channeliza-
tion of sediments due to alluvial action is very possible. Finally,
the upland area probably  forms a recharge  boundary for the
alluvial  aquifer,  and  the  river probably  forms a  discharge
                          Figure 4
                      Example Site Area

  There are four existing monitoring wells identified at the cor-
ners of the site. These wells could provide information on ground-
water  levels as well as the  type  of  sediments  contained  in the
saturated zone. Figure 5 presents preliminary geotechnical find-
ings showing a gravel seam through the center of the site flanked
by  finer grained  sediments. If no aquifer characteristic data is
available, the existing monitoring wells could be used to conduct
abbreviated aquifer tests, such  as  slug  tests. These tests are

described by Cooper, Bredehoeft and Papadopulos2 and Bouwer
and  Rice.3  Thus,  existing site information as well as readily
obtainable information can provide the needed data.
                           Upland Area
                           Figure 5
                Preliminary Geotechnical Findings

   In some site areas, no existing monitoring wells are available. In
such  cases,  geologic and hydrogeologic reports concerning the
aquifers of  the area as well as the  estimated permeabilities and
transmissabilities would  have to be utilized. In these cases, it
might be advantageous to postpone initial model development un-
til specific site information could be obtained from installation of
groundwater monitoring  wells. Generally speaking, sites being in-
vestigated today have a limited number of monitoring wells or soil
borings available for use in the initial stages of  the projects.

   Based upon the  findings of the data review, it is often advan-
tageous to conduct the field investigation work in a phased man-
ner. That is, early acquisition of general site area  information
(such  as aquifer characteristics determined by slug  tests  and
geophysical  data) would  be advantageous to more detailed plan-
ning of the later stages of field investigation.
   Figure  6  shows a  probable set of conditions related  to
geophysical  findings for  the example site area.  The geophysical
methods employed might include determination of terrain  con-
ductivity by electromagnetic techniques.  The results of this type
of investigation  usually are mapped so that  anomolous conduc-
tivity  patterns (inferred to be indicative of contaminant transport
conditions) can be identified. In our example problem, it appears
that the gravel seam through the site area is a preferred avenue of
groundwater movement  and contaminant transport.  In a pre-
ferred groundwater movement area, it is likely that water levels in
the gravel will be slightly lower than in the immediately adjacent
less permeable zones.
  Given the information obtained from the data review as well as
the preliminary work conducted in the initial stages of the site in-
vestigation, a plan of overall investigation effort can  be for-
                         Figure 6
                 Geophysical Survey Findings

mulated. Figure 7 displays how additional data with respect to
groundwater flow conditions and contaminant levels might be ob-
tained by monitoring well installation.  At the same  time, the
available data can be used to establish the initial conditions of the
groundwater flow model.  The groundwater elevations obtained
from the existing monitoring  wells coupled with  the assumed
groundwater levels associated with the gravel seam can be used to
predict the findings of the investigative work. The actual data
from each well would be put into the model as each well was in-
stalled and the model would be rerun to refine predictions for the
remaining wells scheduled for installation. Based upon these find-
ings,  monitoring well  locations would be modified to  maximize
data acquisition efforts. The use of the model to guide and refine
the field investigation exercise can result in the elimination of
monitoring wells and the early cessation of the field investigation

Simulation of Remedial Work
  After the  field  investigation  exercise has been completed and
groundwater flow calibration has been accomplished, the impacts
on groundwater movement due to various remedial scenarios can
be simulated. In the case of the example problem,  groundwater
flow  barrier walls intersecting the gravel zone,  coupled with
groundwater pumping wells, could be simulated.  Groundwater
pumping without barrier walls also could be modelled and the
results could be compared to estimate the savings in  treatment
costs that could be realized due to the difference in groundwater
pumpage. These simulations, as well  as numerous other simula-
tions, can be made on a consistent comparative basis through the
                                                                           CONTAMINATED GROUNDWATER CONTROL    67

 developed groundwater  flow  model. This consistency of com-
 parison is essential to the quantitative evaluation of alternative
 remedial measures.
     I  ILJLl_U-i_LiJ.
                          Figure 7
                Additional Data Needs Definition

   Not all projects may be appropriate for application of ground-
 water flow models as described in this presentation. Certain site
 problems may be so small that a conceptual model may be suffi-
 cient and a computer model may not be warranted. The ground-
 water model is simply a tool that can be used to help organize and
 streamline the site investigation work and the process of remedial
 system selection.
   Whenever a model is being developed during a site investigation
 and is intended to be used to compare and predict remedial action
 impacts, care must be taken to properly apply historical factors
 specific  to  the  site  area.  For  example, long-term  historical
 precipitation  patterns should  be  considered  when  designing
 groundwater   pumpage,  treatment  and  recharge  networks.
 Anomolously high rainfall conditions could cause an increase in
 the volume of water requiring treatment or in other ways impair
 operation of the remedial system. Anomolously low rainfall could
 result in  a general lowering of the groundwater table and cause
 flow conditions to be significantly different than those considered
 by the predictive model. In either of these cases, if the predictive
 model did not consider the long-term fluctuations in precipitation
and the resultant impact on infiltration quantities, the predictive
analyses  compiled by computer modelling might not  be valid.
Mercer and Faust* recommend that groundwater flow models be
 calibrated with known  historical site information and that the
 period of prediction not exceed twice the period of historical data
 used in the calibration.
    When evaluating the overall factors of a site investigation and
 development of remedial alternatives, a general review  of the
 work should be made to insure that it is consistent with sound
 technical "intuition." One of the worst mistakes that can  be
 made in groundwater modelling is unchecked acceptance of the
 model results. Data cross-checking and integration are advisable
 with  or without  a  groundwater  computer  model,  but must
 definitely be utilized when  employing models to protect against
 blind acceptance of computer output data.
    A good conceptual understanding of the site conditions and
 general influences in  the site area is a prerequisite to initiation of
 the site investigation  work and utilization of groundwater model-
 ling. Without such understanding and cross-checking of technical
 information, little or  no integration of technical disciplines will be
 accomplished,  the  overall   program  effectiveness  will  be
 significantly reduced  and the investigation may, in fact, become a
 waste of time and money.

   Until now, groundwater models have not been extensively used
 as tools in the conduct of site investigations. Generally, models
 have been applied after completion of  field  investigations to
 predict the impact of  various remedial alternatives. The prolifera-
 tion of sophisticated  computer equipment and the adaptation of
 groundwater model codes to this equipment have literally brought
 the groundwater flow modelling capability to the desk top of each
 project  engineer and scientist. Thus, groundwater modelling is
 now a readily available tool  that can be used to identify data col-
 lection requirements,  evaluate these data, conduct data sensitivity
 analyses with respect  to groundwater flow impacts and assess the
 long-term impacts of various remedial actions on the groundwater
 flow conditions.
   When employing groundwater flow models, a good conceptual
 understanding of the  aquifer and other pertinent subsurface con-
 ditions  must be available  as  the  underlying  basis for model
 development. The extent to which these conceptual understand-
 ings backed by actual observed field conditions can be simulated
 by the model is  the  quantitative measure of  the ability of the
 model not only to adequately represent existing field conditions,
 but also to predict future conditions. The ultimate responsibility
 to ensure both proper integration  of technical talent in model
 development and calibration, plus proper technical usage of the
 model in simulating remedial activities, rests with the principal in-
 vestigator or overall  project manager.  Proper  integration  of
 technical talent in model development and usage is key to suc-
 cessful site investigation and remedial action alternative analysis.

 I. Guswa, J.H., "Numerical Models, Practical Tools in Groundwater
   Investigations," The  Weston Way. Fall 1984, 10-13.
 2. Cooper, H.H., Bredehoeft,  J.D. and Papadopulos,  I.S.. "Response
   of A Finite-Diameter Well To An Instantaneous Change of Water."
   Water Resources Res.. 3, 1967,  263-269.
 3. Bouwer, H. and Rice,  R.C., "A Slug Test for Determining  Hy-
   draulic Conductivity of Unconfined Aquifers with Completely or
   Partially Penetrating Wells," Water Resources Res.. 12.  1976.423-428.
4.  Mercer, J.W. and Faust,  C.R., "Groundwater Modelling," National
   Water Well Association, 1981.

         Alternative Treatment  Techniques for  Removal  of  Trace
             Concentrations  of Volatile  Organics  in Groundwater
                                                Mark E. Wagner
                                                Brian V. Moran
                                            Geraghty & Miller, Inc.
                                             Annapolis, Maryland
  In response to the detection of volatile organics in wells serving
the process and drinking water needs of an industrial manufactur-
ing facility, a groundwater quality investigation was initiated for
an industrial manufacturer in Maryland. The objectives of the
investigation were to define the extent of groundwater contamina-
tion, to identify probable  sources and  to  develop remedial
measures to rehabilitate the  affected aquifer while ensuring a
potable water supply.
  Field studies confirmed the sources of volatile organics to be
burial trenches located 350 ft southeast of one of the affected
wells.  A hydraulic-control recovery  system was recommended,
using the affected well as a pumping center.
  Several treatment techniques were identified which could treat
the low-levels of tetrachloroethylene and trichloroethylene. The
alternatives included aerated  lagoon volatilization, spray irriga-
tion/land application and packed-tower air stripping. Cost com-
parisons and  treatment  efficiencies for  each  process were
evaluated. Air stripping was selected as the preferred alternative
based on efficiency, cost and reuse of the treated water to satisfy
the facility's water needs.
  An industrial manufacturer in Maryland depended on wells
on the plant property for process cooling and drinking water.
When low levels of tetrachloroethylene (PCE) and trichloroethy-
lene (TCE) were detected in several of the supply wells, the facility
installed individual granular activated carbon (GAC) cartridges at
various points in  the plant. In response to increasing levels of
volatile organics found in the water supply, a groundwater quality
investigation was initiated to determine the probable sources of
the contamination and the degree of aquifer degradation, and to
recommend  remedial measures to rehabilitate  the aquifer  while
ensuring a continuous potable water supply for the plant.

  The plant site is located within the eastern division of the Pied-
mont physiographic province. The region is  characterized by
moderate relief, gentle  slopes and  rounded hills. Locally, the
bedrock geology consists  of folded biotite and chloritic-albite
schist of the Wissahickon formation. The Wissahickon is overlain
by saprolitic material maintaining parental bedding-plane struc-
tures. The saprolite consists of silty clay varying in thickness from
                                                       X STUDY AREA  _. :^°'
                                                    WOODED AREA
                            I WATER SUPPLY WELL
                                                     Figure 1
                                        Site Plan of the Groundwater Investigation
                                                                    CONTAMINATED GROUNDWATER CONTROL   69

             SURFRCE  MRGNETIC SURVEY  (Gammas+50K)
                            Figure 2
       Geophysical Survey Results Using Magnetics Instrumentation
                                                  "'     SURFRCE CONDUCTIVITY SURVEY  (mmhos'm)

                                                                       Figure 3
                                                 Geophysical Survey Results Using Conductivity Instrumentation

                                                     "/ .N
            f~.   Scare* *!•••
    .„    .5"Ts~r
\          S^b\NS-^f $
                                           100      200
                                                                                                        3 ^
                                                           Figure 4
                                         Well Locations and Source Areas of Contamination

                           Figure 5
        Geologic Cross-Section and Groundwater Flow Patterns

 25 to 125 ft. Groundwater exists in both bedrock and the overly-
 ing  unconsolidated  deposits  predominantly  in  fractures  of
 bedding-plane   microfractures.1  The secondary  permeability
 results in a wide areal variation in well yields dependent on prox-
 imity of the well to fracture zones. The facility and surrounding
local populace depend on the Wissahickon formation for water
supply.  Wells in the area typically penetrate the bedrock from 50
to 450 ft, with well yields varying from 5 to 250 gal/min.

  The manufacturing facility is located on top of a hill with a
small stream located along the northwestern and western property
boundaries.   This  stream locally  represents  the  ground water
divide. On-site water-supply wells for the plant are located along
the northwest property line. At the southwest corner of the facil-
ity, a lagoon collects and treats stormwater runoff from the boun-
dary swaleway  and treated wastewater effluent from the plant
(Fig.  1). Water-supply well 7  (WS-7), located at the northwest
property boundary, was found  to contain the highest concentra-
tions of volatile organic compounds (VOCs) with 1300 /ig/1 tetra-
  Facility personnel tentatively identified possible VOC  source
areas  as waste pits  previously excavated in a clearing  350 ft
southeast of Well WS-7. Wastes, which were suspected of being
present in these pits, included process waste drums, construction
debris and defective products. A surface geophysical investigation
was  conducted  in  this area using a proton precession mag-
netometer and terrain electromagnetic conductance instrumenta-
                                                                         FRACTURED ROCK
                                                                                                    MOVEABLE PACKER /PUMP STRIN6
                                                                                                    INFLATABLE PACKER
                                                                                                   INFLATABLE PACKER
                          Figure 6
        Straddle Packer Apparatus Used for Stressing Wells

 tion to locate buried metallic objects. Both surveys were run on
 common transects with 20-ft grid spacings. The proton precession
 method delineated three anomalous areas of low and high signals.
 Terrain conductance delineated three areas of anomalous signals
 located between the proton precession  highs and lows  (Figs.  2
 and 3).
  Subsequent to the surface geophysical work, numerous nested
well clusters were installed around the major anomalies to: (1) de-
fine horizontal and vertical flow patterns and (2) determine local
groundwater quality. Seventeen monitor wells were installed at
discrete water-bearing intervals in  the saprolite material. Wells
were designated by series P, W or S, representing the first water-
bearing zone (20 ft), intermediate water table (40 ft) and bedrock/
overburden interface (60 ft), respectively. The well locations are
shown in Figure 4.
  Based on water-level measurements from the installed monitor
wells, groundwater flow is directed to the west toward the small
stream. In  addition, three significant  seeps  were  discovered
downgradient of the area under investigation.  These seeps are
believed to represent surface discharge of groundwater from the
first water-bearing  zone. Vertically, a strong downward com-
ponent  of groundwater flow also  exists. Groundwater samples
from  the wells and  seeps were  collected and  analyzed for
suspected VOCs. Based on the sampling, Anomaly B was iden-
tified as the major contributor of VOCs  to  the groundwater
regime (Fig. 5). Wells located between Anomaly B and Well WS-7
exhibited  tetrachloroethylene  concentrations  equal  to levels
discovered in the water-supply well.

  In an effort  to better define the groundwater regime in the
vicinity of Well WS-7,  several aquifer  test programs were in-
itiated.  A downhole pneumatic straddle packer/pump assembly
was used in Well WS-7 to stress discrete intervals of the openhole
well (Fig. 6). Short-term pumping was performed at various inter-
                                                                            CONTAMINATED GROUNDWATER CONTROL    71

                                                                Table 1
                                                     Lagoon Disposal Cosl Estimates
                                                          Table 2
                                         Slow Infiltration (Spray Irrigation) Cost Estimates

-. I
' U)



                                                           Table 3
                                                 Stripping Tower Cost Estimates
S t r i p|i l nq Towe i (with b 1 owo r )
P t p i nq

French Drat n
s ?s,ooo
$ 10,000
S 3,000
$ 15,000
$ 19,000
C4M, T«Bpa
Mean* C CD
Means CCD
G*« I
Means CCD
                                            CoetB Un Iquc to
                                            Stripping  IUC)
                                           Total Capital
                                           Costa (TCC)
Subcontracting Fees
Eng 1 nee r ing and De» ign

Start-up/Shake Down

Operating Labor
Electricity for Pumps
$ 7,600 10% o( TCC
S $, 100 9% of TCC

S 3,300 M of TCC

$ 32,700
$ 11,000 2 nan-hri/day 1-30
S 3,000 6.25 h.p. * 1-30
                               Electricity for Blower
                                                                            3 h.p. •
                                                                            Not  Included
                                                               S  1,000
                               Maintenance Contingency Reserve
                                                               S  1,000
 evaluated with and without a liner for the swaleway. Mass balance
 calculations were made, treating the lagoon as a continuous reac-
 tor with first-order kinetics for volatilization of the PCE and
 TCE. Evaluation of flow  conditions in the lagoon at both max-
 imum and average flow rates showed that there  would be suf-
 ficient retention time for volatilization of PCE to concentrations
 below 10 /ig/1 throughout  the year.
   Total capital costs were  estimated to be $41,000 for the unlined
 swaleway option and $79,000 for the lined swaleway. Operational
 costs would be  about $16,000 yearly. Cost estimates  for both op-
 tions are presented in Table 1.

 Spray Irrigation
   Treatment of the volatile organics in the groundwater also was
 evaluated  using  a spray  irrigation method.  With  this option,
 groundwater would be pumped through a network of sprayers
 located throughout part of the wooded  land located above the
 swaleway. It had been  estimated that approximately two acres of
 land would be required to  allow infiltration of the  100 gal/min of
 water. Approximately 60% of the PCE and TCE would volatilize
 during spraying. Water containing residual PCE and TCE would
 enter the swaleway and flow to the lagoon; however,  we deter-
 mined that some could infiltrate to deeper sediments.
   Since spray irrigation for this site  represented a combination of
 slow rate and overland flow methods (for water to the swale), the
method was evaluated using both design procedures. Both design
procedures confirmed that sufficient land was available at the site
for this treatment option. Volatilization rates of the contaminants
were evaluated using a mass transfer approach as in the lagoon
option; volatilization  rates compared  favorably to values in the
literature for similar studies by others.'
   Estimated capital costs for the spray irrigation/recovery system
were $108,200; ytarly operational costs  would be approximately
$15,000 (Table 2). One problem with this option was that the
spray irrigation system would  not be  operational during the
months of January and February due to cold weather conditions.
An additional cost consideration for spray irrigation would be the
need for  a pilot study to determine actual PCE removal rates and
the ability of the  soils to accept the sprayed fluids.

Air Stripping and Reuse of Groundwater
   The last  option  considered was pumping and treating the
groundwater using  a  packed  column tower air stripper.  Using
mass transfer  calculations, we calculated a 99<% removal of the
volatile organics  (PCE and  TCE). This option  would involve
pumping the water to be treated to the top of the aeration tower
and allowing it to flow downward  against a countercurrent flow
of blown air.  The system would be designed so that the treated
water  could achieve volatile  organic concentrations  as low as
1  /ig/1 which,  in turn, could be used as  the water supply for the

                                                             Table 4
                                            Summary of Alternatives, Costs and Evaluation

1 .

Lagoon Disposal


could be

desirable on
costs basis
                               2.  Spray Irrigation
                                  Air Stripping/Reuse
                                                           S  88,600   S 15,000
                                                           $108,200   S 18,000
                                                                    (S  5,060)
                                                                                    Cold weather operation
                                                                                    problems makes this option
                                                                                    less attractive
                 Option would allow for
                 S13,000/yr. savings by
                 eliminating in-plant GAC
plant. Locating the tower adjacent to the plant buildings would
make it convenient to use excess steam, if available, to ensure
operation of the treatment system throughout the winter.
  Total capital costs for the completed stripping-tower/recovery
system were estimated to be  about $108,200; yearly operational
costs would  be $18,000. The cost  estimiates  are presented in
Table 3. These costs assumed that the volatile organic air emis-
sions would be permissible and that no carbon recovery system
would be required.

  The three treatment options were evaluated for effectiveness
and cost (Table 4). Assuming that the only contaminants of con-
cern are trichlorethylene and tetrachloroethylene, both the strip-
ping tower and the lagoon systems can effectively reduce VOC
concentrations in the groundwater to acceptable levels; the spray
irrigation  option  would remove approximately  60%  of  the
volatiles. The capital costs of the lagoon system would either be as
expensive as the stripping tower or 50% less expensive,  depending
upon  whether or  not lining  the  swaleway would be required.
However, considering the operating cost of the in-plant activated-
carbon filters currently used for drinking water (which  would still
be necessary if the lagoon system or the spray irrigation system
were used), the $18,000 per year operating cost of the stripping
tower would be offset by a $13,000 savings by eliminating the ex-
isting GAC system currently  installed. This option would also
allow reuse of the treated groundwater in the plant, since the con-
taminants could be reduced to a "safe" level.
  The air stripping option was recommended to the client and will
be  constructed for  treatment and  reuse  of the contaminated
groundwater. The lagoon treatment system, although competitive
from a capital cost perspective, was eliminated from further con-
sideration  because it did  not eliminate the  high  costs of the
presently installed in-plant activated carbon system nor did it pro-
vide the water-supply advantages that resulted from the packed
tower air stripping treatment system.
1.  Meyer, G. and Beall, R.M., "The Water Resources of Carroll and
   Frederick Counties, State of Maryland," Department of Geology,
   Mines and Water Resources, Bull. 22, 1958, 355.
2.  Jenkins, T.F., Leggett, D.C., Martel, C.J., and Hare, H.E., "Over-
   land Flow: Removal of Toxic Volatile Organics," Cold Regions Re-
   search and Engineering Laboratory, Special Report 81-1, Feb.  1981.
                                                                             CONTAMINATED GROUNDWATER CONTROL    75

             Cost-Effective  Soil  Sampling  Strategies  to  Determine
                        Amount  of Soils  Requiring  Remediation

                                          Gregory J. Gensheimer, Ph.D.
                                            William A. Tucker, Ph.D.
                                                Steven A. Denahan
                                 Environmental Science and Engineering, Inc.
                                                Gainesville, Florida
   Designing a sampling strategy to determine the extent of con-
tamination in uncontrolled hazardous waste sites is not as simple
as putting out some transects or an evenly spaced grid. Similarly,
contouring the data should not be considered a simple exercise in
running software. The spatial distribution of  the data plus the
statistical distribution will determine the type of sampling scheme
and the type of contouring (interpolation scheme) that will repre-
sent field conditions with greatest certainty.
   Sampling schemes  and contouring programs  are most easily
designed and applied when the spatial distribution  of con-
taminants is continuous across the field and when the statistical
distribution tends toward normality or lognormality. If these two
conditions are met, the certainty about the boundaries separating
clean from contaminated  soils will be highest.  This high level of
certainty also can be achieved  by using  a  minimal sampling
   Such ideal conditions of spatial and statistical distributions are
rarely met on  hazardous waste sites.  Hazardous  waste sites
typically are very heterogenous. The statistical distributions most
often are dominated by outliers or exist as multiple populations,
and the spatial distributions usually are discontinuous across the
landscape.  Such non-ideal conditions reduce the efficacy of sim-
ple sampling  schemes or  simple contouring programs. Non-ideal
conditions  also increase the  chances of non-optimal remediation
costs; costs that are much higher  than necessary considering the
"true" extent of contamination.
   A sampling strategy designed to overcome the non-ideal condi-
tions will be  expensive to accomplish, but the extra expense of
sampling can be compensated by reducing the amount of material
removed during remediation. An example of sampling scheme/
remediation optimization is described in detail later in this paper.
   Designing the appropriate soil  sampling strategy and choosing
the methods of interpolation depends on a variety of factors in-
volving the spatial and statistical distributions of the data set as
well as the expected remedial measures. These factors include: (1)
the method of disposal, spatial scale of disposal areas, site history
and previous sampling  results;  (2)  chemical and  physical
characteristics of the contaminants and site including mobility of
contaminants, soil organic carbon, water table and groundwater
flow direction; (3) present spatial extent of the contamination tak-
ing into account the homogenizing effects of the environment; (4)
appropriate sampling, analysis, methods and their costs consider-
ing site conditions, depth of contaminant penetration into the
subsurface  and anticipated  compounds; and (5)  most  likely
remedial alternatives and their costs.
Extrinsic vs Intrinsic Variability
  The firsi three factors above can be broken into  two major
groups relative to the variability contributed by each  factor. Ex-
trinsic variability can be thought of as the variability applied to
the system by non-natural forces (humans).
  Methods of disposal,  spatial  scale and distribution of the
disposal areas are all factors directly imposed onto the system by
humans. There is no way to predict the locations  or size of
disposal areas from viewing the natural system. A series of aerial
photos may  indicate  the presence  of roads, process buildings,
burning grounds or landfills, the location of which can be used in
designing the appropriate sampling scheme. Human  actions set
up the heterogeneity for natural factors to homogenize.
  On the other hand, intrinsic variability can be thought of as the
variability  applied  to the system by the natural environment. In-
trinsic factors  such as temperature,  precipitation and  physico-
chemical characteristics of the media act  in conjunction with the
characteristics  of the contaminant  to homogenize the  extrinsic
  When extrinsic factors are more evident than intrinsic factors,
the spatial and statistical distributions tend to be non-ideal (i.e.,
very  heterogenous).  As the effects of intrinsic factors become
more apparent, the distributions tend to be closer to ideal. This
certainly is not always the case as evidenced by the extreme
variability  of contaminant distribution  in naturally occurring
fractured  bedrock.
  For an example of the intrinsic/extrinsic effects of a hazardous
waste site, consider a drum storage area  with some leaky drums
interspersed through the area. At time zero, numerous small areas
of heavily  contaminated  soil  would  exist   in  an  area of
predominantly clean  soils. A gridded sampling scheme would
probably  miss most  of the contamination, but the statistical
distribution would be dominated by a few extreme outliers. The
extrinsic factors dominate in this situation.
  If the same site  is studied 10-20  years  later, the contaminants
will have begun migrating vertically and horizontally through the
soil. The homogenizing effects of the intrinsic factors are evident.
The same gridded sampling scheme will identify more of the con-
taminated  areas, especially at depth. The statistical distribution
will not be dominated by the extreme  outliers as before. The
statistical distribution will tend toward lognormality (more ideal),
exhibiting  some outliers but not as extreme as at time zero. The
spatial distribution will be more continuous than at time zero.
  Factors 4 and 5 consider the economics of the total remedial ef-
fort including sampling and analysis costs as well as cleanup costs.
The importance of designing an optimum sampling strategy is

minimized if the costs of the expected remedial alternatives are
minimal. If  one of these feasible remedial alternatives is to ex-
cavate, the number of samples to be taken can be optimized by
considering the estimated volumes of soils needed to be removed
considering sampling density.

  When the  pattern of contamination is relatively simple (a con-
tiguous contaminated area with little or no contamination outside
that area) an optimal sampling density can be defined to minimize
total remediation costs,  including both  the  investigation and
cleanup phases. This  concept  is illustrated best by considering
some extreme examples.  Assume that  100 m2 of soil are con-
taminated. Next assume that samples  were taken to define this
contamination: one sample at its center and then samples on a
uniform rectangular grid with a sampling interval of 100 m2 for a
total of 9 samples as in Figure 1. As has been discussed, soil con-
tamination is extremely heterogeneous, and interpolation between
data points  is risky at best.  It is prudent, based  on the  data
available from this hypothetical site, to remediate all soils within
the region defined by the clean samples, approximately 40,000 m2.
                                      CONSERVATIVE ESTIMATE OF
                                      CONTAMINATED AREA
                             CONTAMINATED AREA
                          Figure 1
            Example of Inappropriate Sampling Density

   On the other hand, one might have started by taking a sample
every meter along a similar rectangular grid. Since phased sam-
pling frequently is not feasible, this approach would have  re-
quired 40,000 samples to define the contamination distribution
over the same area. Even if we had been clairvoyant, or had some
independent  information on  the soil contaminant distribution,
sampling at a density of 1 m would have required 121 samples to
define a contaminated area of only 10 m2.
   Obviously these are ridiculous extremes, but they clarify an im-
portant problem. How does one define an appropriate sampling
density?  If we sample intensively,  sampling and analysis  costs
could exceed the cleanup costs. If the sampling grid is too coarse,
we may not accurately define the contaminant distribution. Based
on a reasonable requirement to ensure the public health and safe-
ty, remediation of large quantities of clean soils may be indicated,
wasting valuable resources.
  This problem is resolved as a simple linear optimization prob-
lem where the objective is to minimize total remediation costs, in-
cluding the wasteful remediation of clean soils and the cost of
sampling and  analysis.  The cost  of remediating contaminated
soils must occur in any case and can be ignored in the analysis.
  The cost of wasteful remediation is calculated as follows. It can
be assumed that two rings of sample locations will define the con-
taminated perimeter. The perimeter of most simple shapes is 3 to
4 times the width (a circle is 3.14 times; a square is 4 times). It can
be assumed approximately that the perimeter, P, is given by P =
3.6 W,  where  W is a length scale, such  as the width. The area
where wasteful remediation  can occur is the distance between
sampling points, L, times the perimeter of the contaminated area.
  The sample density, D, is the number of samples per unit area.
Thus, L = D - '/2. Combining these definitions, the area in which
wasteful remediation is likely to occur is AR  = 3.6 WD- '/*. The
cost of wasted remediation is  given by a unit  cost, R$,  the
remedial cost per unit area times the area of wasteful remediation.
  The sampling and analysis costs will be simply the S&A costs
per sample, SA$, times the number of samples, N. The number of
samples is given approximately by N =  DW2 (sample density
times area). The total costs thus are given by:
  $ = SA$ (W2) D + R$ (3.6 W)D -
                                                                   The solution of the optimization problem is derived by differ-
                                                                 entiating with respect to sample density and setting the derivative
                                                                 to zero. This will define the sample density at which total costs are
                                                                 minimized. The result is given by:
Consider the following example:
    SA$  = $1,000 per sample
     R$  = $900/m2
      W  = 75 m
Then D = 0.069 sample/m2.
  The sampling interval, L = D -l/z = 4 m, and one sample every
4 m would optimize overall cost. Approximately 400 samples
would be indicated with a sampling program  cost of $400,000.
Contrast  this  with the cost of wasted remediation, estimated in
this example to be $810,000, even though a relatively dense sam-
pling grid is employed. If the sampling interval were 20 m,  the
sampling  costs would be reduced to  $15,000, but the cost of
wasted remediation would increase to $4,000,000. Total costs
would  be approximately $3,000,000 higher than the optimal
  By inspection, Equation 2 indicates that a lower sampling den-
sity (greater spacing between samples) is optimal if the con-
taminated area is large (large W) or if remedial costs are less than
in the example. If remediation is inexpensive, then overdoing it is
less costly than sampling so there  is less need to define the con-
taminated area precisely.

  Geostatistics are best  applied to normal populations that are
spatially  continuous. As normality  and  continuity  criteria
deteriorate, geostatistics  become less optimum in ability to inter-
polate. Somewhere between ideal  and non-ideal data sets, inter-
polation schemes such as method of triangles, inverse distance
weighting and kriging become useless.
  Kriging is a valuable interpolator because, unlike other geosta-
tistical methods,  kriging produces an error value for each inter-
polated value. By paying close attention to the errors, the user can
                                                                                    CONTAMINATED SOIL TREATMENT   77

determine when kriging becomes a liability in determining extent
of contamination. There  is no easy way to tell when other inter-
polators fail to produce accurate estimates.
  The Semi-Variogram (SV) is the tool within  kriging that iden-
tifies the expected ability of this interpolator to work. A combina-
tion of SV characteristics and goodness-of-fit values helps iden-
tify the usefulness of  kriging before interpolation even begins.
Goodness-of-fit values are typically called jackknifing errors' and
have been the standard method of determining the usefulness of
the kriging system and its SV.
  Characteristics of the  modeled SV  are as valuable or more
valuable than jackknifing errors as indicators of the ability of
kriging to  produce  optimum  estimates.  The  important  SV
characteristics include the nugget value (y-intercept), sill value
and shape (value of gamma where the  SV levels  out) and noise.
The ideal SV has a nugget value of zero, a well defined flat sill and
minimal noise. The noise can be defined as the variability of the
SV data points about the fitted model.

Pitfalls of Using Geostatistics
  User knowledge of SV  characteristics and the Cumulative Den-
sity Function (CDF)  is  vital  to  be  confident that the kriging
algorithm actually is producing optimum estimates.
  Using jackknifing  errors alone may be misleading.  Certain
hazardous waste site characteristics tend to result in problems if
kriging is  to  be used as the interpolator. The problems  are
evidenced by SV or CDF characteristics.
  Waste sites consisting of numerous  small waste spills  inter-
spersed through generally uncontaminated  soils  most likely will
result in a statistical population dominated by outliers. The CDF
would be heavily skewed. The effects of the outliers also will be
noticeable as significant  noise and/or a significant nugget value
on  the SV. Sometimes the SV is so noisy that a model cannot be
fit to it.
  Log transformation is the simplest remedy  for this situation
but, depending on the extremity of the outliers, it  does not always
work. The transformation will usually smooth the SV, enough to
be modeled, but rarely affects the nugget value.1 Transformation
removes the extreme variance caused by the outliers but does  not
remove the baseline variance between the majority of points. In
one case,  log transformation reduced the noise but resulted in a
pure nugget (horizontal SV) indicating that  the sampling scheme
was not of sufficient density to identify spatial variability.
  Kriging this system  will tend to greatly overestimate the extent
of contamination. The peak true contamination values will be re-
duced  while the true extent  of contamination will be overesti-
mated. The interpolation scheme will not generate any new peaks
in the field but will incorrectly increase the extent of low level con-
tamination especially around the measured outliers. This becomes
significant when the cleanup criterion is near background.
  Another waste site  characteristic that causes problems is one
that contains several lagoons or larger areas of contamination in-
terspersed through relatively uncontaminated soil. A grid system
of points  covering this site most likely would show a bimodal
distribution, one peak of near background values and  another
peak of high values. The SV may show a poorly defined sill or
possibly even two sills representing the effects of the two popula-
tions of values. If the true SV model cannot be well estimated,
one of the populations will be overemphasized causing a severe
over or under estimate of the true extent of contamination. One
solution to this problem  is to work with the two populations in-
dividually to calculate individual SVs and then use both SVs as a
comparison for interpolation.
  Waste sites characterized by basin environments are conducive
to kriging. In a basin, the extrinsic variability is decreased by the
intrinsic factors involved with seasonal flooding. Each time the
system floods,  the  contaminants disperse somewhat creating a
more continuous spatial distribution and decreasing  the extreme
outliers to create a more lognormal population. These spatial and
statistical distributions are more ideal than those found on other
types of hazardous  waste sites.

Case Histories
  The  following example shows how information about extrinsic
and intrinsic factors can be used to optimize the utility of kriging.
The site is a basin environment, so the values are dominated by in-
trinsic  variability.  Boundaries of  expected contamination  were
identified from   previous investigations (extrinsic variability) and
used to help locate  the sampling grid (Fig. 2). The results from
this first phase are not available yet, so we've described one of the
probable outcomes. We assume that the eastern third of the site
was uncontaminated while the samples up to the western boun-
dary were contaminated. Our confidence along the western half
boundary is bolstered by the previous information on  the location
of the perimeter (extrinsic factors). There is less need  of sampling
along this boundary since it is more certain. The dashed line, P2,
represents the line of proposed phase  2 sample  locations based
only on seeking a precision of 35%  in estimation of contaminated
soil areas. This line would encompass 1.35 times the area cir-
cumscribed by P|. This precision of estimation of contaminated
soil  lies  in  the requirement under recently  issued CERCLA
Guidance' that feasibility study cost  estimates  be  accurate to
within  + 50% or - 30% of the actual costs. Where the perimeter
extends beyond line Pi, an additional line of samples is proposed
(P3) which again would encompass an area 1.35 times larger than
the previous area.
                                                      SE I

                           Figure 2
   Sampling Scheme for Hazardous Waste in a Basin Environment

  We have less confidence in the boundary along the eastern half
of the site since it is completely estimated. A parameter,  C,  has
been used to help locate subsequent sample locations.3
  Parameter C is defined as follows:
  C =  [(Criterion) - (Est + (Conf)(KE)]/Standard Deviation       (3)
  Criterion =  Cleanup criterion
        Est =  Interpolated value
      Conf =  Confidence interval (1.28 for 90% confidence)
        KE =  Kriging error associated with interpolated point.
  The 75 and 90% confidence intervals about the boundary of
minimum extent of contamination were drawn  using parameter
C. It is obvious that many more samples are required to identify
this boundary than the western boundary due to increased uncer-
tainty in this area.
  Parameter C is unique since it combines the cleanup criterion
with the kriging error to arrive at the most likely Phase 2 sampling
locations. Areas that are significantly above or below the criterion
are eliminated  from further  consideration even if uncertainty is
large  in those areas. The use of kriging error alone to indicate
future sampling  areas discriminates based  only on uncertainty
rather than boundary locations which can be misleading.
  Kriging  was  used in  conjunction  with a phased  sampling
scheme on another site and reported by Tucker et al.3 Parameter
C was used here to identify confidence intervals for the whole site
since the expected extent of contamination was not well known on
this site. This site was not a basin area as described previously, so
it had less than ideal spatial and statistical  distributions. Log
transformation  of the  data eliminated  SV noise resulting in a
smoother curve.  The uncertainty of boundary locations was  still
significant however, and caused a large portion of the site (1/3 of
the area) to  be considered  for Phase  2 sampling. This level of
uncertainty was caused mostly by a few outliers in the data set  and
resulted in the  need  of a significant Phase 2 effort.

  With the advent of user-friendly computer contouring pack-
ages,  there will be an  increased tendency to misuse kriging or
overemphasize the positive benefits of kriging when used on inap-
propriate data.  Kriging along with other geostatistic methods
should not be used as standard procedure at all hazardous waste
sites.  The data must be  examined  for  statistical  and  spatial
distribution patterns before initiating geostatistics. Geostatistics
become more valuable as the spatial distribution tends to be con-
tinuous and the statistical  distribution tends to  be normal  or
lognormal. Both patterns are relatively rare on hazardous waste
sites since extrinsic factors usually dominate intrinsic  factors in
controlling  site  variability. Extrinsic factors  tend to result  in
discontinuous spatial data exhibiting statistical outliers.
  Kriging is more useful than other interpolators since it produces
an error term with each estimated value. This error can be used to
plan a second phase of sampling using Parameter C when the data
from Phase 1 sampling are less than ideal. This error also may in-
dicate that kriging or other interpolators are not useful on that
particular site.
  Finally, sampling schemes can be designed to optimize the total
costs of remediation (sampling costs + remediation costs) by op-
timizing the relationship between the number of samples taken
and analyzed versus the costs of soil remediation. By examining
this relationship mathematically, it appears that costs of extra
sampling can be easily overcome by reducing the final costs of soil

1.  Gambolati, G.  and Volpi, G., "Groundwater Contour Mapping in
   Venice  by Stochastic Interpolators. 1.  Theory,"  Water Resources
   Res., 155,  1979,281-290.
2.  Gensheimer, G.J., Flaig, E.G.  and Jessup,  R.E., "Effects of Log
   Transformation on Semi-Variogram Modeling Using Kriging Error,"
   Agronomy Abstracts, 1985, Chicago, IL.
3.  Tucker, W.A., Gensheimer, G.J.  and Dickinson,  R.F., "Coping
   with Uncertainty in Evaluating Alternative Remedial Actions," Proc.
   Fifth  National Conference on Management of Uncontrolled Haz-
   ardous  Waste Sites, Nov. 1984, Washington, DC, 306-312.
4.  U.S. EPA, Memorandum from William Hedeman, Director of Office
   of Emergency and Remedial Response, and Gene Lucero, Director of
   Office of Waste Programs Enforcement, to Waste Management Di-
   vision Directors,  Regions 1-10;  Subject: Preparation of Records  of
   Decision for Fund Financed and Responsible  Party Remedial Ac-
   tions, Mar. 1984.
                                                                                     CONTAMINATED SOIL TREATMENT    79

                   Land  Treatment  of Wood Preserving  Wastes
                                                   John R. Ryan
                                            Remediation Technologies
                                               Ft. Collins, Colorado
                                                     John  Smith
                                                    Koppers  Inc.
                                             Pittsburgh, Pennsylvania
  Numerous wood preserving plants currently are evaluating dis-
posal alternatives for sludges and contaminated soils generated
at these plants. Land treatment is a potentially environmentally
sound and cost-effective means of treating these wastes on-site.
Results of four pilot-scale studies completed at wood preserving
plants in California, Montana,  Minnesota and the northeastern
United States are summarized. The land treatment studies in-
volved a detailed evaluation of  the degradation, adsorption and
toxicity  reduction of the hazardous constituents present in these

  This  paper presents a description of on-going industry spon-
sored studies concerning the land treatability of sludges and con-
taminated soils resulting from wood preserving operations. The
objectives of this paper are to:
• Identify the types of field  and  laboratory studies being con-
  ducted in this area
• Identify the methods being used in the treatability studies
• Present preliminary results of the on-going studies

  Of the estimated 400 wood treating plants in the United States,
over 80% use creosote and/or pentachlorophenol as wood pre-
servatives. These plants may generate sludges and contaminated
soils containing the above preservatives as a result of spills, leaks
and  settled material from wastewater  impoundments. Many of
the wood preserving plants are regulated  under RCRA or  are on
the  National Priority List  and  must identify  alternative  means
for disposal of sludges and contaminated soils.
  This  paper briefly discusses four current laboratory and  field
studies   concerning the land  treatability of  wood  preserving
wastes. Table 1  summarizes the location at each  study and the
types of wastes involved. All the studies have been designed to
meet the objectives of  a  land  treatment demonstration  under
RCRA as defined in 40 CFR 270.20. These objectives are to:

• Accurately simulate proposed full scale conditions including
  soils, climate and design and operating conditions
• Demonstrate that the wastes can be degraded, transformed and
  immobilized within the treatment zone of the land  treatment

  The four studies  described  in this paper will provide valuable
data in determining whether land  treatment of wood preserving
wastes can successfully meet RCRA standards. It should be em-
phasized that additional studies concerning  the land treatability
of wood preserving wastes have been completed. These include
but are not limited to studies by Sims', Sims and Overcash2, Um-
fleet el al.\ Umfleet,4  Edgehill and Finn,' Baker and Mayfield'
and  McGiniss.' Laboratory  studies also  are  being conducted
under the sponsorship  of the U.S.  EPA Robert S. Ken Labora-
tory in Ada, Oklahoma.

  The objective of this study was to develop design and operat-
ing criteria for full scale on-site treatment of  creosote-contam-
inated soils at a wood  treatment plant in Minnesota. The study
included both bench scale and pilot scale evaluation of several
performance, operating, and design parameters. These param-
eters include:

  Soil characteristics
  Treatment supplements
  Reduction of organics, phenolics and PAH compounds
  Toxicity reduction
  Effect of initial loading rate
  Effect of reapplication

Soil Characteristics and Climate
  The study soil is a fine sand which comprises the  upper 2 ft of
the bottom of a RCRA impoundment  previously  used for the
storage of wastewater from a tie treatment plant. The soil is con-
taminated with creosote constituents which primarily consist of
PAH and phenolic compounds. Total benzene extractable hydro-
carbons in the contaminated soil range from approximately 8 to
12% by weight.
  The soil has a natural pH of approximately 6.5 and, therefore,
no pH adjustments were necessary prior to initiating the studies.
Specific conductance values and  metal  contents of the soil are
generally within the range of values reported  for  natural soils.
The carbon to nitrogen (C/N) ratio of the contaminated soil is
high, and nutrient additions were necessary to reduce the C/N
ratio to a range which promotes microbial growth.
  Because the natural  soils are fine sands and  extremely perme-
able, it was decided that the full scale system would include a liner
and leachate collection system to prevent leachate break through.
To replicate the proposed  full scale conditions, the pilot studies
consisted of five lined, 50 ft1 test  plots  with leachate collection.
Table 2 shows the experimental conditions for each of the test
  The field studies were conducted from July 25  through Oct.
30, 1984. Average ambient temperatures generally exceeded 50° F
during this period. July and August were exceedingly dry months,

                                                            Table 1
                                                Summary of Land Treatment Studies


3C 11 i ty


Loca tion


s 1 i f or nia


Type(s) of waste
Creoso te
con tamina ted
and pentachlorophenol

Con taminants
PAH, T. Phenol

PAH, clorinated phenol

contaminated soi 1




North East



Cr eoj
con tf
imi n;
contamina ted
and pentachlorophenol
ated soil
PAH, T. Phenol

PAH, chlorinated phenolic s

L— Laboratory Study
F— Field Study
and the field  plots required weekly watering to maintain them
near field capacity. October was a very wet month, and the plots
were near saturation during this  period. The active degradation
period appears to extend through the month of October at this

Treatment Supplements
  The studies were designed to maintain soil conditions which
promote the  degradation of hydrocarbons. These conditions
• Maintain a pH of 6.0 to 7.0 in the soil treatment zone
• Maintain soil carbon to nitrogen ratios of 25 to 1
• Maintain soil moisture near field capacity
  In addition, the studies evaluated the effect of seeding the soil
with commercially available microbes adapted to hydrocarbon.
  The soil pH was within the desired range at startup and no lime
additions were necessary. Soil pH decreased below 6.0 in most  of
the plots after 2 months of operation. Subsequent additions  of
1.0 to 1.5 tons/acre of agricultural grade limestone raised the pH
of the test plots to the desired range and no subsequent drops in
pH were observed during the test period.
  The equivalent of 10-20 tons/acre of 10:10:10 fertilizer (10 Ib
nitrogen, 10 Ib phosphorous and 10 Ib potassium per 100 Ib fertil-
izer) were added to all the plots at the  beginning of the study
based on the estimated carbon content of each  plot. The fertil-
izer additions  successfully reduced the carbon to nitrogen ratios
of each plot below 25:1. Residual ammonia  levels following the
fertilizer application were fairly high, however,  creating a con-
cern about possible ammonia toxicity problems.
  All the test plots were watered based on the results of daily rain-
fall and pan evaporation data as  well as daily soil  tensiometer
reading. Soil irrigation was necessary on a weekly basis during
most  of the study. The monitoring and irrigation program  suc-
cessfully maintained the soil near field capacity. No significant re-
ductions in organics were observed in the plot which was not
watered. These data illustrate that maintaining the soil  near field
capacities is a critical  operational  parameter for obtaining  suc-
cessful degradation. Field capacity of the fine sands appears to be
approximately 10% moisture by weight.
  No significant advantages in organic degradation can be attrib-
uted to seeding the plots  with the adapted microbial culture used
in the study. It is hypothesized that an active natural soil micro-
bial population exists in the contaminated soil. Due to the length
of time the soil contamination existed, the existing soil microbes
are well adapted to the contaminants present in the soil.

Reduction of Organics, Phenolics and PAH
  Reductions of benzene extractable hydrocarbons were fairly
similar between all the field plots.  Similar reduction trends were
observed in the laboratory reactors, however, at lower rates.  Per-
cent removal efficiencies and first order kinetic rate  constants
were fairly similar for all the field plots. Average removals for all
field plots over the 4 months were approximately 40% with a cor-
responding first order kinetic constant (K) of 0.004.
  The greatest mass of benzene extractable removals  was asso-
ciated with the highest initial loading rates. Total  removals  over
the  4-month test period  on the basis of  pounds of  organics re-
moved/ft3 of soil/degradation month are presented in Table 3.
                                                            Table 2
                                              Experimental Specifications for Test Plots





Soil Initial Oil1
Specifications Content (%)
Con t aminated
Hydrobac CL
Hydrobac CL
no microbial
Hydrobac CL
Visibly non-
Hydrobac CL
soil and clean sand; 10

soil and clean sand: 5

soil and clean sand, 5
soil and peat moss, 5

contaminated soil, 3

Fertilizer Cultivation
(Ib) Frequency
50 Weekly

24 Weekly

25 Weekly

25 Weekly

15 Weekly

                 'Benzene Extractable Hydrocarbons.
                 •10-10-10 Fertilizer.
                                                                                     CONTAMINATED SOIL TREATMENT    81

                           Table 3
         Total Benzene Extractable Removals In Field Plots

Extractable Content
Plot Number (percent by weight)
1 8.83
2 4.00
3 4.06
4 10.59
5 1.73
Total B
( Ibs/f t of
e n z e ne
e Removals
. 56
. 34

   These removal rates can be compared to data from petroleum
 refinery land treatment facilities.' Typical  removal rates at well
 operated refinery facilities range  from 0.1  to  1.0  Ib of oil  re-
 moved/ft5 of soil degradation month with  the lower end corres-
 ponding to a "low" initial concentration (such  as in Plot 5) and
 the high end corresponding to "high" initial concentrations such
 as Plots 1 and 4. The data, therefore, are in good agreement with
 removals documented at refinery land treatment  facilities.
   Phenol removals did not exhibit the same linear relationship
 over time as the  benzene  extractable contents.  Phenol removals
 were observed over the first three  months followed by a market
 increase in the phenol content of  the soil in the fourth month.
 The trends observed may  be  partially a result of analytical vari-
 ability but also may be affected by the breakdown of complex
 organic compounds releasing phenolic constituents.
   First order rate kinetic  constants for PAH were fairly similar
 between the field and laboratory studies with the exception of 4 +
 ring compounds in test runs 1 and 4.  The laboratory tests had
 higher first order rate constants than the field plots for these com-
   Table 4 compares the  first order  rate constants for benzene
 extractables and  PAH compounds between the laboratory and
 field studies which had an initial benzene extractable content of
 4 to 5% and the laboratory and field studies which had an initial
 benzene extractable content of 8 to  10%.
   The median kinetic values (of the 4 to 5% studies and the 8 to
 10% studies) are approximately equal for all the parameters. The
 most notable variance is for the 4 + ring PAH compounds. The 4
 to 5% initial loading rates resulted in slightly higher kinetic rates
 for these compounds as compared  to the 8 to 10% initial loading
 rate. The kinetic  rates appear independent  of the initial loading
 rate within the range of loading rates tested.
                          Table 4
             Comparison of the Range and Median
           Kinetic Values at Two Initial Loading Rales
                     4 to 5 percent
                      (K. day-')
                                                to 10 percent
Range       Median
     (K,  day-')
Ben zene

2 Ring
3 Ring
4 + Ring
Total PAH







.01 3-0.



                        Toxicity Reduction
                          A battery of toxicity assays was completed on 3 test plots (1,2
                        and 5) at the completion of the 4-month study as well as on creo-
                        sote sludge and contaminated soils which approximately repre-
                        sented the start up conditions in plots 1 and 2 at the beginning of
                        the study. The battery of assays included:
                                                                  • Microtox
                                                                  • Ames mammalian mulagenicity
                                                                  • 96-hour static acute fish bioassay
   The battery of assays resulted in the following observations:
 • Test plots 2 and 5 were non-toxic after 4 months
 • Test plot 1 showed intermediate toxicity after 4 months
 • Both contaminated soils and the creosote sludge were toxic to
   all the bioassays

   The microbial assays (microtox and Ames test) on the contam-
 inated soils would  have suggested that  the  initial concentrations
 in test plots 1 and 2 would have resulted in a toxic response to the
 soil microorganisms  and  therefore no significant  degradation
 would occur. The results from these plots, however, have shown
 significant degradation.  This indicates that  an acclimated micro-
 bial population existed in the contaminated soil prior to the initia-
 tion of the studies.  The creosote constituents (within the range of
 initial concentrations tested) do not  result in toxic effects  to the
 acclimated soil microorganisms used in the toxicity tests.
   From an environmental toxicity standpoint, however, the toxic-
 ity tests show  that the toxicity of the contaminated soil has been
 decreased through  the treatment process.  Complete toxicity re-
 duction appears to fall between 2.5 and 5% benzene extractable
 content. Complete toxicity reduction was achieved after 4 months
 in plot 2.

 Effect of Initial Loading Rate
   First order rate constants were fairly similar between all  the
 loading rates.  The  intermediate loading rate (4 to 5% benzene
 extractionables)  may demonstrate a slightly higher removal of
 high molecular weight PAH compounds. Toxicity reduction also
 appears to occur at a faster rate for a 4 to  5% initial loading rate
 than at the higher initial loading rates.
  The higher loading rates, however, showed the greatest mass
 removals  and  a  clear detoxification trend. The studies suggest
 that all the loading rates tested are feasible. The  selection  of an
 initial loading rate  should balance additional land area require-
 ments against  time requirements for completing the treatment
 process. Moderate loading  rates (5%) will result  in  a faster  de-
 toxification  whereas higher loading rates will decrease  land area

 Effect of Reapplication
  Greater kinetic rates  were observed  after waste reapplication
 to a treated soil. The results suggest that average loading of 1 Ib of
 benzene extractable/ft' of soil/degration month to a treated soil
can be treated effectively when the application does not exceed
3.0 Ib/ft.1 The  first application should occur after the initial
application has been shown  to be successfully treated.

Operating and Design Criteria
  The studies  have successfully developed operating and design
criteria for a full-scale system at this site. These criteria include:
• Treatment period can be extended through October
• Soil moisture should be maintained near field capacity
• Soil pH should be maintained between 6.0 to 7.0
• Soil carbon:nitrogen ratios should be maintained at 25:1

• Fertilizer applications should be completed in small frequent
• Initial benzene extractable hydrocarbon contents of 5 to 10%
  are feasible
• Bioassays can be used to determine treatment effectiveness
• Waste reapplication should occur after initial soil concentra-
  tions have been effectively degraded
• Waste reapplication rates of 2 to 3 Ib of benzene extractables/
  ft3 of soil/2 degradation months can be effectively degraded
  The  studies completed  have shown that the visibly contam-
inated  soil at the site can be effectively degraded and detoxified.
The data indicate that the organic content of  the visibly contam-
inated  soil present in the upper two feet of the RCRA impound-
ment can  be degraded to levels similar to the organic content of
the soils beneath the visibly contaminated soils. The data also
indicate that the treatment is effective in detoxifying the waste.

  This study involves a field pilot program to evaluate the feas-
ibility  of land  treatment of contaminated soils from a wood pre-
serving plant in California.  Principal hazardous constituents  in-
clude polynuclear aromatic hydrocarbons (PAH), pentachloro-
phenol, dioxins and furans. The studies were initiated in the sum-
mer of 1985 and may be continued through the fall of 1986.

Treatment Medium Soils
  The test plot soils  are located  east of a spray irrigation field
which is proposed as the full-scale treatment area. The general
characteristics of the surface soils at the  spray field and the test
plots nearby are characterized as clayey silts which grade to silty
clays with depth.  Depth to groundwater in the area is generally
30  to  40  ft beneath the ground surface.  Specific soil character-
istics of the test plots based on core samples  to greater than 7 ft
show  the soils to be relatively homogeneous with depth. The
soils through  the test plots are characteristic of a brown silty
clay, tight to very tight, with small gravel clasts to 1 in.

Field Plot Experiments
  Three experimental field plots were constructed at the facility
 near the  spray irrigation field. The plots were graded to a  1%
 slope and earth berms were constructed between the plots. Run-
 off is completely collected at the low slope position of each plot.
The collection areas for the plots are sized to collect the run-off
 from the 25-year, 24-hour storm (i.e.: approximately 4.5 in).
  Also included in the plot design was  the  installation of soil-
 pore liquid samplers. The samplers are used to collect water from
 the unsaturated zone for the purpose of determining if hazardous
 constituents are migrating out of the treatment zone.  The plots
 were fitted  with three types of lysimeters per plot: a commer-
 cially  available vacuum lysimeter, a fabricated  glass block lysi-
 meter and a gravity flow trench lysimeter. During the course of
 the study, the vacuum and glass block lysimeters have proven to
 be  severely limited in extracting  water for analysis. The trench
 lysimeter has worked effectively  in  collecting macropore flow
 through the treatment zone which is the most critical factor in
 determining if hazardous constituents are moving through  the
 treatment zone.
  The plot design also included the installation  of tensiometers.
 Tensiometers  are important in monitoring  soil suction  or  the
 force that determines which way moisture will move in soil. For
 the study, tensiometers are used to measure moisture in the zone
 of  incorporation and at 18 in. to determine irrigation scheduling
 and at 5 ft to  monitor soil moisture movement through the treat-
 ment zone of the three plots.
  After the plots were constructed and the treatment area was
readied, the contaminated soil was spread and  mixed in place
with native soil to achieve an approximate benzene extractable
content of 2% in plot 1  and 3000 ppm of pentachlorphenol in
plot 2. Sewage sludge was applied at a rate of 20 tons/acre to
Plots 1 and 2 to increase the organic matter of the contaminated
soil as well as reduce the  carbon:nitrogen:phosphorous ratio to
a range to promote microbial growth. Also, agricultural lime was
spread as a step to raise soil pH to 6.0 to 7.0. The final step of
startup activities was the mixing of contaminated soil, fertilizer
and lime with a tractor mounted rototiller to a depth of 6 to 8 in.
  Table 5 summarizes the operational activities conducted dur-
ing the pilot scale studies. These activities included test plot culti-
vation, irrigation, sampling and on-site climatological and sub-
surface monitoring.
                           Table 5
                  List of Operational Activities
Test Plot Cultivation
Lime Additions
Fertilizer Additions
Moisture Additions
On site monitoring
-soil temperature
-pan e va poration
-ra in fa 11
-soi 1 suction
-soil moisture
Freq uenc y
As necessary to maintain pH 6-7 '
As necessary
As necessary

 'Determined from monitoring data.
Soil and Water Sampling
  Soil samples were collected from Plots 1 and 2 at 2-week inter-
vals for the first 6 weeks of the study and then monthly until the
end of the study. Each plot was subdivided into three subplots for
sampling purposes. The subplots were each 12 ft by 16 ft  with
1-ft grid spacings for a total of  192 grid squares per plot. Four
subsamples were collected in each subplot during each  sampling
event. These four subsamples were then composited into one sam-
ple for each subplot. Each  subsample was collected from one of
the grid squares. The four grid squares sampled in any one event
were selected from a random number table. Each grid square was
sampled only once during the course of the study. The random-
ized sampling was recommended as part of quality assurance pro-
cedure to account for field variability.
  Also, during the study, nine grab samples were taken from one
subplot during one sampling event as well as the normal com-
posite sampling. This sampling was done to compare composite
sampling versus grab sampling.
  Soil core sampling was conducted prior to waste application,
60 days and 90 days after waste application. Cores were taken in
each  subplot including the  control plot. The cores were collected
every 1 ft to a depth of 7.5 ft.
  Two water samples were taken during the study. Runoff sam-
ples were  collected  from  each  plot  in  September,  and trench
lysimeter samples were collected from each plot in December. All
samples were shipped in coolers packed  with ice under standard
chain-of-custody procedures. Sampling equipment  was decon-
taminated with deionized  water and methanol  after each sam-
pling round.
                                                                                     CONTAMINATED SOIL TREATMENT    83

                           Table 6
                 Range of PAH Kinetic Constants






2 Ring





3 Ring




4 Ring






Tola 1





(1) not delected at startup.
Organic Reduction Rates
   Complete analytical results for  the study were not available
during the preparation of this paper. Preliminary results, how-
ever, indicate that benzene extractable removals in  each of the
subplots ranged from 34-66% for the 4-month time period with
a  corresponding  first order kinetic constant (K) ranging from
0.003 to 0.009.
   Table 6 summarizes the first order kinetic constants computed
from the preliminary analytical results for various ring classes of
PAH compounds. The following compounds are listed under the
noted ring classes:
2 Rings           3 Rings         4+ Rings
Naphthalene      Fluorene        Fluoranthene
Acenaphthylene    Phenanthrene    Pyrene
Acenaphthene     Anthracene      Benzo (a) anthracene
                                Benzo (a) pyrene
                                Benzo (b&k) flouranlhene
                                Benzo (ghi) perylene
                                Indeno (1,2,3-cd) pyrene
                                Dibenzo (a,h)anthracene
   Considerable variability was observed in the pentachloraphenol
(PCP) data. Table 7  summarizes the measured values over the
course of the study  and the corresponding  calculated removal
rates. The  PCP data are non-conclusive and continuations of
these studies into 1986 are necessary  to draw any  treatability
  As a byproduct of PCP production, furans and dioxins may be
present at variable concentrations. Tetra and 2378 furans and
dioxins were not detected at startup in Plot 1  and  were below
5 ppb and 1 ppb for the Tetra and 2378 furans and dioxins re-
spectively in Plot 2. The remaining furans and  dioxins  (Penla,
Hexa,  Hepta) show removals after 4 months  between  75 and
95%.  The octa furans and dioxins have the least percentage re-
movals with values ranging from 0-81%.
  In summary, significant removals of furans and dioxins were
observed over the 4 month period, and initial startup concentra-
tions were relatively low in comparison to PCP and PAH startup
  The studies may be extended  through 1986 to evaluate  long-
term performance of the plots and evaluate the immobilization
of the contaminants.

  This study is an  on-going  RCRA land  treatment  demonstra-
tion for a proposed full scale land treatment facility in Montana.
The wastes being evaluated are creosote contaminated soils and
sludges resulting from the closure of a RCRA surface impound-
ment. The proposed full scale land treatment facility will be used
to treat and detoxify the wastes removed from the impoundment
at closure. Design parameters for the field studies were developed
based on results of Case Study A.

Soil and Climate
  The treatment area soils consist of loamy silts  and silty sands.
The caution exchange capacity of the surface soil is approximate-
ly 20 milliequivalents/100 g of soil and the organic matter content
of the surface soil is approximately 1.5%.
  The climate of the area is  semi-arid. Annual  evapotranspira-
tion exceeds annual precipitation and there is limited recharge to
the  groundwater  resulting  from  precipitation and run-off.
Monthly ambient  temperatures  exceed 50° F approximately 6
months of the year.

Description of Study
  The soil and climatic conditions at the site are considered high-
ly favorable for attenuation of the creosote contaminants. There-
fore, unlike  Case Study A, this system is  not lined.  The results
                                                           Table 7
                                       Summary of Pentachlorophenol Removals in Plots 1 and 2
                                   Concentration  (ppm)

                                   at Time (t)  Daya
                                                               Percent of  Removal
                                                               at Ti«e(l)  Days(l)
                                      Da to
                          7/12/85 8/16/85 8/27/85 10/10/85  11/11/85  8/16/85 8/27/85  10/10/85  11/11/85
                     Plot   t-0     t-35    t-46     t.90      1-122    t-35    t-46     t-90      1-122
1 10
1 20
1 10
1 1
1 5
                    (I) Percent Removal at Time   (C(O)-C'(T)/C(O))IOO.

of Case Study A, however, were used to select the design param-
eters  for the study including waste application rate and treat-
ment supplements.

Waste Application Rate
  Waste application rates are based on gross organic loading as
defined by benzene extractable hydrocarbons. The loading were
selected on the basis of results from Case Study A and involved
initial loadings of 2% and 5% benzene extractables. The 2% plot
includes subsequent bi-monthly applications of 2% benzene ex-
tractables to evaluate the effect of waste re-applications. The 5%
plot does not include waste re-application in order to evaluate the
effect of one batch loading.

Treatment Supplements
  To minimize  the need for replicate plots and minimize the
analytical costs of the study, certain management practices were
selected as "good management practices" for the study and were
applied consistently to all the plots.  These "good management
practices" are anticipated to optimize treatment (based on liter-
ature data) and include:
• Maintain surface soil pH between 6.0 and 7.0
• Maintain a soil carbon:nitrogen ratio of 25:1
• Maintain soil moisture near field capacity through irrigation
• Apply manure or organic  amendments  to provide  nutrients
  and increase the organic carbon content of the surface soil
• Till the soil once every 2 weeks
• Prevent storm water run-on from entering the plots and capture

Construction of the Plots
  The field plots were constructed in June  1985. One plot is a
control where no  wastes have been  applied but  all the  "good
management practices" are implemented. The other two plots are
for the two waste loading rates being evaluated.
  Each plot is 12 ft x 50 ft with a defined treatment depth of 5 ft.
The plots are separated by earthen berms which prevent run-on
from entering the site. Each plot is graded to a  1 % slope,  and a
lined run-off collection sump capable of handling two 25-year,
24-hr storm events is present  at the toe of the slope. The  entire
study area is fenced:
  Three types of soil-pore water collection devices were installed
in each plot. These include:
• A vacuum lysimeter capable of placing a suction on the sur-
  rounding soil
• A pan lysimeter constructed from a hollow glass block approx-
  imately 1 ft2
• A trench  lysimeter consisting of a 10-ft long 6-in.  diameter
  PVC drain, placed at the bottom of the treatment zone
  In addition soil tensiometers were placed at 3 ft and 5 ft. These
tensiometers in conjunction with a hand tensiometer used to take
soil suction readings at the  surface  are used  to evaluate  the
hydraulic flux through the soil.

  Monitoring at the site includes monthly sampling of the zone of
incorporation to evaluate the degradation of specific organic con-
stituents and bi-annual sampling of soil cores, soil pore water and
groundwater to evaluate contaminant immobilization.
  Table 8 presents the analytical  parameters evaluated in each
medium. The principal hazardous constituents  being evaluated
are polynuclear aramatic hydrocarbons PAH. These constituents
are monitored on a monthly basis along with indicator  para-
meters such as benzene extractable hydrocarbons and total phen-
                          Table 8
                 Analytical Parameters, Site C











-Benzene Extractables/Soxhlet extraction


-Total kjeldahl nitrogen/automated titrimetric

-Ammonia nitrogen/automated titrimetric

-pH/combination electrode

-Conductivity/conductivity bridge

-Total organic carbon/combustion or oxidation
-PAH/Method 8270

-Base neutrals/Acid extractables

-Microtox/Toxicity Analyzer System

-Fish bioassay
Z   zone of incorporation soil samples.
C = core soil samples
L   lysimeter water samples.
ols. In addition, nutrients,  soil electrical conductivity and soil
pH are monitored on a monthly basis to evaluate the need for
treatment supplements. Soil moisture is checked on a weekly basis
to evaluate  the  need for supplemental irrigation.  The relative
toxicity of the surface soil is evaluated through monthly micro-
tox analyses.
  Bi-annual soil core sampling includes collection of soil cores in
1 ft increments down to the bottom of the treatment zone (5 ft).
The soil cores, groundwater and lysimeter samples are analyzed
for indicator parameters and for a complete scan of base neutral
and acid extractable organics using GC/MS techniques. Inorgan-
ic metals also are included in the analyses using ICAP techniques.
  All sampling and analyses include triplicates from each plot to
determine statistical variability. The startup of the study was in
July 1985, and it is scheduled to be completed in 1986. Insuffic-
ient analytical data were available at the time of this writing to
evaluate degradation trends. However, the soil core, lysimeter
and groundwater analyses have not detected  any migration of
hazardous constituents from the treatment zone.

  This project involves a detailed laboratory  study using creo-
sote and pentachlorphenol  contaminated soils from an  aban-
doned wood preserving  plant in the northeastern United States.
The study was initiated in December 1985 and is scheduled to run
for 8 months.
  The laboratory  study program evaluating  land  treatment  is
divided into the following three areas:
• Evaluation of leachate movement through  soil  columns and
  contaminant volatilization
• Examination of contaminant degradation rates in soil pans
• Development of soil/water partition coefficient for PAH and
  pentachlorophenol associated with the site soil

Soil Column Evaluation
  The studies outlined below are designed to evaluate the rate
and extent of treatment, including biodegradation  and retarda-
tion. A mass balance of selected constituents will be determined
by taking into consideration the initial loading rate, the concen-
tration of constituents in the leachate for the  duration of the  8-
month experiment and the concentration of constituents in the
soil core at the termination of the experiment.
  The soil treatment process will be evaluated using 4-in. diam-
eter, 5-ft long glass column reactors packed with clean soils  to
represent actual site  conditions. The columns contain the same
                                                                                     CONTAMINATED SOIL TREATMENT    85

soil  profile materials that exist at  the site.  Native  soils  were
collected in 6-in. intervals and packed into the columns in the
same order as obtained in the field to obtain a representative
soil column profile.
   Before any contaminated material was applied to the soil col-
umns, chloride tracer studies, using sodium chloride, were  per-
formed. This work was completed to determine the proper hy-
draulic loading rate once contaminated material is applied.
   Actual  contaminated soil  from the site was mixed  into the
upper 6 in. of the clean soil in a porportion to give approximate-
ly 5% benzene extractables. Common commercial fertilizer and
manure also were added  and the  soil pH was adjusted to  and
maintained between 6-7 by lime additions. The columns are to be
operated for approximately 8 months with only one initial waste
application.  Watering of the columns is to be done with its pH
adjusted to approximately pH 4. The pH adjustment is done to
simulate the existing pH of rain water for the area.  In  addition
to routine watering, the columns are to be watered with the equiv-
alent of a 25-year, 24-hr storm event in January and May during
the study. The soil is to be tilled every other week. Volatile emis-
sions are to be addressed as part of the column work.
   For QA/QC purposes, two duplicate columns containing the
waste material are operated. A third column containing  only
clean soils serves as a control and is operated in the same man-
ner as the columns with contaminated material. All three columns
are to be sampled periodically for analyses.

Soil Pan Reactor Evaluation
   In conjunction with the operation of the soil columns, a soil
pan study phase is to run  for approximately 8  months  with
degradation rates of specific compounds addressed. Four soil pan
reactors are operated. The soil pans are constructed of aluminum,
have a surface area of approximately 1.5 ftj and contain approx-
imately 6 to 8 in. of soil. The soil is a mixture of contaminated
material and clean surface soil from the site in a proportion to
give a concentration of approximately 5% benzene extractables.
Fertilizer and manure have been added to give a C:N:P:ratio of
approximately 50:2:1, respectively. Manure was applied corres-
ponding to a loading rate approximately 10-20 dry tons per acre.
This is the same soil misture that is to be used in the column re-
actors. The soil pH is maintained  within  a range of 6-7 by lime
addition as needed, and the  soil moisture is  maintained within
50% of field capacity by laboratory tap water adjusted to a pH of
4. During the course of the experiment, hand tilling is to be done
every other week.
   For QA/QC purposes, three duplicate pans containing the soil
waste material lime, fertilizer, manure are to  be set  up. Two of
the pans are  to be maintained in a moist environment by water
applications as needed. The third pan is not to be watered and is
to serve as a contaminated  soil control.  A fourth pan, contain-
ing only clean soil, is to be operated in the same manner as the
contaminated soil pans.

Soil/Water Partition Coefficient Development
   Soil/water partition coefficients are to be developed for penta-
chlorophenol and PAHs as these compounds specifically relate to
the contaminated  site soils. For this work,  different solution
pHs are to be evaluated to examine the effect of pH on the leach-
abilities of PAHs and pentachlorophenol.
   This  information is needed to develop a better  theoretical
understanding  between pentachlorophenol and  PAH attenua-
tion by the soil matrix.
  Land treatment is a potentially technically effective and eco-
nomic means of treating and disposing of wood preserving wastes
from sites currently regulated under RCRA or CERCLA. Rigor-
ous 2-year studies are currently under way to determine appro-
priate engineering design features for these systems.  The data
from those studies will provide a sound scientific basis for eval-
uating the feasibility of this technology at other wood preserv-
ing sites.
1.  Sims,  R.C., "Land  Treatment  of Polynuclear Aromatic Com-
   pounds." Ph.D. dissertation. Depi. Biol. Agr.  Eng.. No. Carolina
   State Univ., Raleigh, NC, 1982.
2.  Sims, R.C. and Overcash,  M.R., "Fate of  Polynuclear Aromatic
   Compounds in Soil-Plant Systems," Residue Reviews 88. 1983, 1-68.
3.  Umfleet, el al..  "Reclamation of PAH  Contaminated Soils."  Pre-
   sentation made at  Environmental Engineering 1984. Specialty Con-
   ference, ASCE, Los Angeles, CA. June 1984.
4.  Umfleet, D.A.,  "Land Treatment of PAH Compounds,"  Unpub-
   lished M.S. thesis. Department  of  Civil and Environmental Engi-
   neering, Utah State University, Logan, UT, 1985.
5.  Edgehill and Finn, "Microbial Treatment of Soil to Remove Penta-
   chlorophenol. " Appl. Environ, .\ficrobiol. -IS.  1981, 1120-1125.
6.  Baker, M.D. and  Mayfield, C.I.. Water. Air and Soil Pollul., 13,
7.  McGinnis, G.D.,  "Biological and  Photochemical  degradation of
   pentachlorophenol and creosote," Mississippi Forest Products  Lab-
   oratory. Mississippi Slate, Ml, 1985.
8.  American  Petroleum  Institute. "Land  Treatment  Practices in the
   Petroleum Industry." API. Washington, DC, 1983.

                      Method  for  Determining  Acceptable  Levels
                                of Residual  Soil  Contamination

                                                 William A. Tucker
                                                   Carolyn Poppell
                                 Environmental Science and  Engineering, Inc.
                                                 Gainesville, Florida
  With the exception of PCBs, regulatory standards or estab-
lished numerical criteria that define acceptable levels of soil con-
tamination at  hazardous waste sites do  not  exist.  However,
numerical criteria are readily available for air and water. In con-
trast to air and water, soils are not fluid and are more readily sub-
ject to property rights. People experience little direct exposure to
contaminated soils.  For example,  adults inhale approximately
17,000 I/day and ingest  roughly 2  I/day  of  water, but they
generally are believed to accidentally ingest less than 0.00005 1 of
soil per day. For these and other reasons, Congress has not
enacted the "Clean  Soils Act," and  soil quality standards have
not been  implemented.
  The extent of human exposure to  residual contamination in
soils can be extremely  variable, depending on  such  factors as
climatic factors and  soil properties. The most important variable
is land use.  Contaminated soils in a  backyard playground and
subsistence garden (common in many residential areas) pose a
much greater health risk than in a paved shopping mall develop-
  Consequently, it is reasonable to determine acceptable levels of
residual soil contamination on a site-specific basis, considering
the probable (or  possible) future uses of the land  during  the
period that it will  remain contaminated. Many feasibility studies
for remediation of contaminated sites and remedial actions im-
plemented at sites reflect  this  concept. On the other  hand, im-
plications of this concept have been addressed only implicitly or
qualitatively in many Remedial Investigation/Feasibility Studies
(RI/FS)  conducted  to  date.  Systematic and  quantitative im-
plementation of this concept can be  accomplished with multi-
media exposure and risk assessment procedures that are readily
available  and well  documented.
  The purpose of this paper is to present these procedures in the
context of the endangerment assessment process as they currently
are implemented by the authors. The methods used by the authors
are similar  to the  PPLV method  developed  by Dr.  David
Rosenblatt and co-workers of the U.S. Army Medical Bioengi-
neering Research  and  Development Laboratory; the  interested
reader should consult Rosenblatt and Dacre7, Rosenblatt, et a!.'
and Small9 for further discussion of their procedures.  Following
the presentation of the method, a few examples will be presented
to illustrate some features of its implementation.

  Establishment of acceptable residual soils concentrations often
will be an essential element of the endangerment assessment pro-
cess. Endangerment assessment links the remedial investigation to
the feasibility study in two ways:
• The process results in determination that an endangerment or
  public health risk exists. If the  determination is positive,  it
  virtually rules out the no-action alternative and demonstrates
  the need for  the feasibility study.
• The endangerment assessment should clearly identify the en-
  vironmental  quality  criteria that would be required  for the
  CERCLA alternative; that is criteria which, if met, would sub-
  stantially eliminate the endangerment.
  The establishment of acceptable levels  of residual soils con-
tamination contributes to  both objectives of the endangerment
assessment. Endangerment may be suspected if the "endanger-
ment criteria" are exceeded. The role of "endangerment criteria"
in the RI/FS process is presented in Figure 1.
  The recommended approach for determining acceptable levels
of residual soils contamination is presented in Figure 2. Results of
the  remedial investigation  will identify the contaminants of con-
cern for the site. Inputs from  the endangerment assessment team
will be useful at an early  stage in the remedial investigation to
refine the list of contaminants  of concern to focus RI resources on
the  "riskiest"  contaminants.  Further refinement and  screening
undoubtedly will be required as the RI data are acquired and re-
viewed.  In some circumstances a formal hazard ranking system
may be useful to screen contaminants, and applications of a for-
mal screening  model may be a component of the uppermost
"box" of Figure 2.
   Once contaminants of  concern have been identified, investiga-
 tions proceed along two parallel and simultaneous paths which in-
 terface at intermediate points in the  process as well as in the  in-
 tegrative final steps. The two paths  are contaminant fate (down
 the right in Fig. 2) and contaminant  toxic effects (down the left).

   First, the available toxic effects data for the contaminants of
 concern are reviewed. It is important to identify any particularly
 sensitive subpopulation (e.g.,  children, women of childbearirig
 age, aquatic species or age classes). Considering this information
 on sensitive subpopulations  and  site-specific factors (e.g.,  in-
 dustrial site where children  would not be  exposed) permits the
 identification of populations at risk.  Populations at risk are those
 potentially exposed populations who also are sensitive to the con-
 tamination. The most sensitive population at risk normally will be
 considered as the basis of the endangerment criteria. However, if
 the number of individuals in  the most sensitive population at risk
 is likely to be very small, it may in some instances be reasonable to
 base the criteria on another  sensitive population while allowing
 for special or auxiliary protection of the special, rare subpopula-
                                                                                  CONTAMINATED SOIL TREATMENT    87


                                                                             NO FS NEEDED
                                                                          NO ACTION ALTERNATIVE
                                                             Figure I
                                           Role of "Endangcrmenl Criteria" in RI/FS Process
   The next step in the effects side of the process is to determine
 acceptable exposure levels for  the sensitive population at risk.
 Determination of an acceptable exposure level is often the most
 controversial step in the process depicted in Figure 2. Much of the
 effort, techniques and procedures frequently referred to as carcin-
 ogenic  or quantitative  risk assessment  are  almost exclusively
 devoted to this element of the process.  Later in this paper, it will
 be shown that other aspects of the process also introduce substan-
 tial uncertainty and variability in estimates of acceptable residuals
 levels. For carcinogenic contaminants, this step also introduces
 the controversial issue of acceptable risk: recent  policy guidance
 from the U.S.  EPA indicates  that a 10-*  lifetime  individual
 cancer  risk  level  should  be evaluated  as a cleanup goal  at
 CERCLA sites, consistent with  its guidance for groundwater
 Alternative Concentration Limits,  a site-specific standard setting
 mechanism under RCRA. Alternative cleanup standards resulting
 in lifetime individual cancer risks  in the range of  10 ~8 to 10-4
 also should be considered.

 Land Use Scenario
   As illustrated in  Figure  2, a critical requirement of the soil
 criteria setting procedure recommended here is the need to define
 an exposure scenario. As used in this discussion,  an  exposure
 scenario is a characterization of  the behavioral  patterns  of  the
 population at risk that could result in, or reduce, exposure to  the
 contaminants at the site.  The  critical  element of the  exposure
 scenario will be  the land use scenario used in the analysis.
   Depending upon  the degree  of conservatism  that may seem
 warranted to affected parties, and the persistence of site-related
 contaminants, the exposure scenario may be based on: (1) the  ex-
 isting land use, (2) probable future land  use over a  specified
 future period, based  on existing  or predicted patterns of land
 development and (3) the use (of all possible future uses) that leads
 to the greatest  potential for exposure.  The choice among these
 alternative scenarios  is not strictly a technical decision, but the
 technical implications of selecting among these alternatives can be
 addressed by a technical specialist. Ultimately, the  selection of a
 land use scenario for the analysis and the selection of an "accep-
 table risk" are political/regulatory decisions, not technical ones.
 It will be shown later that these  decisions can affect the ultimate
soils criteria by factors of a thousand or more.

 Environmental Fate
  Parallel to and  simultaneous  with the review of toxic effects
data, the environmental physical chemical properties of the con-
                                              RCVIEW ENVIRONMENTAL
                                             BEHAVIOR OF COMPOUNDS:
                                              ENVIRONMENTAL SETTING
                                               IDENTIFY DOMINANT
                                              PATHWAYS OF MIGRATION
                                                  FROM SITE
                            EXPOSURE ALONG EACH PATHWAY
                       CONSUMPTION OF FISH • _< ACCEPTABLE DOSE
                            ACCEPTABLE SOIL CONCENTRATION
                           Figure 2
         Determination of Acceptable Soil Concentrations
                by Exposure and Risk Assessment
taminant are reviewed in the context of the environmental setting
of  the contamination/soils  to evaluate  the site-specific en-
vironmental behavior of the contaminants.
  Dominant pathways of migration from the site are identified by
screening level analyses and calculation. Screening level analysis
procedures are diverse, and flexibility must be maintained in their
application in order to derive the most information possible from
the available data,  while conserving "level of effort" resources.
Identification of the dominant migration pathways usually can be
deduced on the basis  of the judgment of senior environmental
scientists familiar with the site. Occasionally, formal screening
level calculations as represented, for example, by methods of En-
field, el at.' or Tucker, el a/.10,  can be applied at  this step.
  Integrating  the information  from the exposure scenario with
the dominant  migration pathways yields the dominant pathways
of exposure. These are the pathways, from soil to an exposed
member of the population at risk, that are capable of carrying

and delivering a significant dose of contamination. An example of
an exposure pathway would be "surface run-off erodes surficial
soils and carries both dissolved and adsorbed soil contaminants to
a nearby estuary;  fish and shellfish  bioconcentrate  the  con-
taminants from the water,  and the organisms are harvested for
human consumption."
  Once the dominant exposure pathways  are identified, each is
expressed as a mathematical equation relating the dose (exposure)
via the pathway to the soil contaminant level. These equations are
representations of actual processes and are designed to express the
relationship conservatively  (inferring a higher-than-expected ex-
posure for a given soil concentraton) to favor protection of public
health. The most important pathways  should be evaluated with
the greatest  degree  of sophistication  to  make their  exposure
estimates more reliable.
  Finally, the exposures via all of the individual pathways are
added together, and the sum (total exposure) must be less than the
acceptable exposure. This step links the exposure assessment with
the toxic effects evaluation to produce the risk assessment-based
criterion. Each pathway is defined as a factor times  the soil con-
centration, and   the  acceptable  exposure  and  the  factors are
known. Consequently the equation may be inverted  to derive the
acceptable  soil concentration.

  In  most applications  of  this procedure  conducted by the
authors, the  following pathways of exposure  have required con-
• Incidental  direct soil ingestion
• Dermal absorption
• Uptake by plants in the human food chain
• Leaching by infiltrating rainwater leading to contamination of
  shallow water table aquifer
• Groundwater or  surface  run-off discharge to surface  water
  bodies supporting a fishery
• Resuspension of contaminated  soil particles to the atmosphere
  by wind erosion or mechanical entrainment  associated with
  earth moving activities
  Consequently, the assumptions related to exposure via the pro-
cesses and  the uncertainty  associated  with procedures used to
quantify these processes, become critical issues in determining ac-
ceptable levels of residual soil contamination.

Importance of Land Use Scenario
  The land use issue can be investigated with reference to these
processes by  contrasting the assumptions involved for a residen-
tial land use scenario vs.  an alternative  land  use.  Consider  a
hypothetical example: rural land now used for lumber production
has been determined to  be contaminated. Assume it  has  been
estimated that the contaminants may persist in soils for 50 to 200
years. If the existing land use is maintained for that  period, then
the exposure  potential is very low. Workers involved in lumbering
will be exposed to soil contamination via dust inhalation, inciden-
tal  direct soil ingestion and dermal absorption. Small9 recom-
mended the use of 10 mg/m3 as a reference exposure level for this
situation. Since levels higher than 15 mg/m3 are not permitted in
the workplace, this is a useful conservative estimate. Actual levels
may  be  much lower. Depending on the importance of this
pathway relative to others, a more realistic estimate considering
level of activity, soils and climatic factors may be warranted. Ex-
posure to airborne dust also depends on the pulmonary ventila-
tion rate. Data are presented in Anderson, et al.' and Small.*
  Adults probably ingest very little soil. Soil ingestion occurs as a
result of contamination of food and food wrappers, dirt on hands
(especially  prior  to  eating) and  ingestion of inhaled particles
trapped in the upper respiratory tracts. Definitive data on typical
or extreme rates of soil ingestion among adults are not available.
Small' recommends the use of 100 mg/day, an estimate consistent
with Ford and Gurba" recommendations. Although there is very
little basis for this estimate, it is commonly used. General support
for the estimate relates to studies of juvenile soil ingestion that in-
dicate ingestion rates from 100 to 1,000 mg/day.2 It is generally
accepted that most adults ingest less soil than most children.
  A basis  for estimating dermal absorption  exposure  to  con-
taminated soils was presented recently by Layton, et al.6 Their
analysis indicates that it is conservative to assume that adults are
exposed to the contamination contained in 43 mg of soil per day
via this exposure pathway. Alternative approaches to estimating
dermal absorption  are available, but few address the details of ab-
sorption from contaminated soils, so the Layton, et al.6 method is
a useful basis for screening level analysis.
  The information presented above may be combined to estimate
the total dose experienced by the individual as follows:
Inhalation   Dust
 Dose = [0.1 g/day  + 17 mVday (0.01 g/m3 + 0.043 g/day]Csoil  (1)
Expressed in terms of the soil concentration:

  DoseO»g/day) = 0.31 g/day x Csoilf4g/g
  To determine the acceptable soil concentration, ACS, we sim-
ply require that the actual dose be less than or equal to the accep-
table dose:
  Dose = 0.31 g/day x'ACS S Acceptable Dose
implying ACS  <  Acceptable Dose
                     0.31 g/day

  To  complete the hypothetical example, insert an  acceptable
dose for a 70 kg adult of 2,000 /ig/day. Then the acceptable soil
concentration is 6,450 /tg/g.
                           Figure 3
      Characteristic Elements of a Residential Land Use Scenario
                                                                                    CONTAMINATED SOIL TREATMENT    89

  The recommended residual contamination level would have to
be much lower  if it were  anticipated that  the  land  would  be
developed residentially. Figure 3 illustrates some of the exposure
pathways  operating  under a  residential land use.  It  must  be
assumed that children may spend most of their time playing on
contaminated ground. Subsurface contamination can be exposed
during normal residential landscaping which also may  resuspend
dust into the air. Families may derive a significant fraction  of
their diet from a backyard garden. The residential water supply
may be derived from shallow individual wells, depending on local
hydrogeologic conditions and the availability of community water
supplies. These factors can  lead to substantially higher exposure,
especially for children.
  Assume that the population at risk is children ages 1 to 3 having
a typical body weight of 12.5 kg. Their acceptable daily dose for
the (same) contaminant would be 350 /ig/day. Based on the data
presented by Binder, el at..1 it seems  likely that the average child
ingests 100 to 200 mg of soil per day and a value of 150  mg/day
will  be  used in this  sample calculation. Kolbye, et  al.'  provide
estimates that a 2-year-old ingests 0.4 I/day of water and  125 g of
potatoes and other vegetables  per day  (20 g/day, dry weight);
while Layton, el a/." estimate that a  child  is exposed to the con-
tamination found in 41 mg  of soil per day, by the dermal absorp-
tion route. These exposure  factors will account for the major ex-
posure  pathways to children in a residential setting.  To relate
these exposure  factors back to the  soil concentration, it  is
necessary to establish the relationship between soil concentrations
and: (1) plant concentrations (plant bioconcentration factor) and
(2) groundwater/drinking water concentrations.
  The best method for estimating plant bioconcentration  factors
will depend on the combination of contaminant and plant(s) and
the data available for that combination.  Assume that Small's"
guidance has been used to indicate that plant concentration on a
dry weight basis will be approximately five times soil concentra-
  Groundwater/drinking water concentrations may or may not
be related to soil concentrations.  However, if the shallow water
table aquifer yields water of acceptable quality and quantity for
residential potable use, then it usually would be assumed to be the
source of drinking water. Further, and for the sake of simplicity
in this example, if the contaminated land is situated at  a ground-
water divide, then the quality of water in the wells will be similar
to the quality of the soil solution in equilibrium with the con-
taminated soils. Then the concentration in drinking water will be:
where: Kj is the adsorption coefficient. Assume that Kj =  125
for the contaminant in site soils. Combining this information for
the various pathways, the total dose for the 1-3 year old children
comprising the population at risk can be calculated in proportion
to the soil concentration.

             Soil          Drinking         Garden         Dermal
           ingestion         water         vegetables       absorption
Dosc=   CSO,| (0.15 g/day  + 0.4 I/day li'** + 2(1 g/day • 5  • 0.041 g day)  (6)
Expressed in terms of the soil concentration:

  Dose O^g/day) =  103 Csoi, (/tg/g)

97% of the dose (100/103) is calculated to come from the garden
vegetable pathway.
  To calculate an acceptable soil concentration (ACS), the  dose
must be less than the acceptable dose.

  Dose  = 105 g/day Csoi, < 350 ^g/day
  Comparing this value with the ACS for the lumber production
land uses scenario (ACS S 6,450/ig/g), it is seen that the land use
scenario assumption can affect the recommended soil criterion by
more than three orders of magnitude. This degree of sensitivity to
assumptions is apparently at least as important as the uncertainty
regarding  toxic effects and quantitative carcinogenic risk assess-
ment procedures.

Refinement of Exposure  Factors/
Reduction of Uncertainties
  The results of any predictive method, such as the one presented
here, always  will be uncertain.  In multimedia exposure  and risk
assessment,  the number of variables  and  uncertainty  in each
results in large uncertainties. A  traditional response to uncertain-
ties in environmental  assessment  has been  a reliance on conser-
vative estimates. The true situation is unknown, but estimates and
decisions arc  based  on extremely adverse assumptions.
  The simple predictive methods presented in the examples above
can be refined by acquisition of more data or application of more
sophisticated predictive methods. As the estimates become more
reliable, the degree  of conservatism can  be reduced. This process
usually  will result  in   a  less stringent,  yet  more  firmly based,
  As an example, consider the basis for  estimating inhalation ex-
posure for the lumber production scenario.  It was assumed that
air  concentration of  dust that had  been stirred  up from con-
taminated areas could be as high as 10 mg/m'. The basis for that
estimate was conservative:  if  levels were higher,  the  workers
would wear respirators. This estimate could be  refined either by
sophisticated wind erosion/fugitive dust modeling or by measur-
ing air  concentrations under   the conditions of  the exposure
scenario.  Assuming that  fixed air  monitors were placed in a
similar,  though  uncontaminated  area,  and  that  personnel
monitors also were employed in a 1-month study of worker ex-
posure to nuisance dust, a more realistic dust concentration of 0.4
mg/m-1  was  determined. Applying  this  finding, the previous
analysis can be  refined. The  dose  is now estimated to be:
           ingestton        inhalation
Dose -  CSO,| (0 I g da>  *   17 m.Y das •
     = 0.15 mg da> C'ioil I eg g)
  Airborne        Dermat
    dim         absorption
0 IXKM g m» *  0.043 mg day) (7)
and the acceptable soil concentration would be:
                                                                                             =  13.400
  Such refinements almost invariably should result in a relaxation
in the recommended criterion, as long as uncertainties always are
resolved with a conservative assumption.

  A method for determining safe  levels  of residual soils con-
tamination as a component of the endangerment assessment pro-
cess for hazardous waste sites has been presented.  The method
presented  is based closely on the PPLV approach developed by
Dr.  David H.  Rosenblatt  and  co-workers at the U.S.  Army
Medical Bioengineering Research and Development Laboratory,
Ft.  Detrick, MD."* The method has been  restated here, using
slightly different terminology, to illustrate its relationship to the
endangerment assessment process.
  Development of a hypothetical example illustrates the sensitiv-
ity  of  the  recommended criteria to  the exposure  assumptions
adopted, particularly those relating  to the future land use of the
site. The land  use  scenario  adopted for the analysis affected the
recommended  acceptable soils concentration by more than three

orders of magnitude in the example problem. It is possible to con-
struct alternative scenarios that affect the criteria by even greater
amounts. Thus,  uncertainty  regarding  future land  use  and the
assumptions adopted in the face of that uncertainty can influence
the endangerment assessment, the evaluation of the no-action
alternative and the costs of  remedial action just  as severely as
uncertainties regarding the "toxic effects/quantitative risk assess-
ment" aspects of the problem. The exposure  assumptions are
readily understood and thus readily debated by all concerned par-
ties, and so they can and should receive a thorough airing during
the RI/FS. Decisions regarding appropriate future land  uses for
the site are quintessentially political, rather than technical deci-
sions.  Affected parties should be consulted, and these decisions
should be made by regulatory authorities and/or affected parties.
   The second major  conclusion  of the paper is based on  an
analysis in which one of the criteria was refined by acquisition of
additional data.  This allowed a more realistic  analysis  and the
discarding  of an excessively  conservative assumption.  More
sophisticated theoretical analysis frequently can achieve a similar
result. To develop criteria in the face of great uncertainties, it fre-
quently is prudent to adopt  conservative assumptions.  This af-
fords the decision-maker the certainty that the criterion level is
safe and that concentrations  below the criterion do not pose an
unacceptable risk. The converse, however, is untrue: levels above
the derived criteria are not necessarily unsafe. The refinement of
estimates presented bears this out: with a conservative  assump-
tion, applied in lieu of site-specific data, it was determined that
soils having 6,450 jtg/g of contaminant "X" were "safe"  and
presented no endangerment. Upon acquisition of additional data,
it became apparent that soils  as high as  13,400 /ig/g can be left in
place. This does  not contradict the previous analysis.  Before ac-
quisition of the new data, the analysis indicated that 6,450 jtg/g
was a safe level. The  new  analysis confirms  this conclusion.
Generally, the public and their representatives interpret results of
this type to mean that soil contamination  at 6,500 /ig/g, or any
level above 6,450 /ug/g, is harmful. Drawing this interpretation
from the conservative analysis would be inappropriate and should
be avoided.

 1.  Anderson, E., Browne,  N., Duletsky, S. and Warn, T., "Develop-
    ment of Statistical Distributions or Ranges of Standard Factors"
    used in Exposure Assessments, U.S. EPA Contract No. 68-02-3510,
 2.  Binder, S., Sokal, D. and Maughan, D., "Estimating the Amount of
    Soil Ingested by Young Children Through  Tracer Elements,"  Un-
    published Manuscript. Div. of Env. Hazards and Health Effects,
    Center for Env. Health, Centers for Disease Control Public Health
    Service, U.S. Dept. of Health and Human Services, Atlanta, GA.
 3.  Enfield,  C.G., Carsel, R.F.,  Cohen, S.Z.,  Phan, T. and Walters,
    P.M., "Approximating Pollutant Transport  to  Ground Water,"
    Ground Water, 20, 1982, 711-22.
 4.  Ford, K.L.  and Gurba,  P., "Health Risk Assessments for Contam-
    inated Soils," Proc. of the Fifth National Conference on Manage-
    ment of Uncontrolled Hazardous  Waste Sites,  Washington,  DC,
    Nov. 1984,  230-731.
 5.  Kolbye, A.C., Mahaffey,  K.R., Fiorino, J.A., Corneluissen, P.C.
    and Jelinek,  C.F., "Food Exposures to Lead," Environ.  Health
    Perspectives,  May 1974, 65-74.
 6.  Layton,  D.W., Hall,  C.H.,  McKone,  T.E., Nelson, M.A.  and
    Ricker, Y.E., "Demilitarization of Conventional  Ordnance: Priori-
    ties for Data Base  Assessments of Environmental Contaminants,"
    UCRL-53620 Draft prepared  for USAMBRDL,  Ft.  Detrick, MD,
 7.  Rosenblatt,  D.H. and Dacre, J.C., "Preliminary Pollutant Limit
    Values  for  Human Health Effects," Environ. Sci.  Techno/.,  14,
    1980, 778-84.
 8.  Rosenblatt,  D.H.,  Dacre, J.C. and Cogley, D.R., "An Environ-
    mental Fate Model Leading to Preliminary Pollutant Limit Values
    for Human Health Effects" in R.A. Conway  (ed.), Environmental
    Risk Analysis for Chemicals,  Van Nostrand Reinhold, New York,
    NY, 1982, 475-505.
 9.  Small, M.J., "The Preliminary Pollutant Limit  Value Approach:
    Procedures and Data Base," U.S. Army Medical Bioengineering Re-
    search and Development Laboratory, Ft. Detrick, MD, Tech.  Re-
    port 8210, 1984.
10.  Tucker, W.A., Dose, E.V., Gensheimer, G.J.,  Hall, R.E., Pollman,
    C.D. and Powell,  D.H.,  "Evaluation of Critical Parameters  Af-
    fecting Contaminant  Migration through  Soils,"  Report AMXTH-
    TE-TR-85030, prepared by Environmental Science and Engineering,
    Inc. for U.S. Army Toxic and Hazardous Materials Agency, 1985.
                                                                                         CONTAMINATED SOIL TREATMENT    91

                  Objective Quantification of  Sampling Adequacy
                           and  Soil  Contaminant Levels  Around
                               Point Sources Using  Geostatistics
                                                  Jeffrey C. Myers
                                          Geostat Systems  International
                                                 Golden,  Colorado
  This paper describes the events and conclusions resulting from
a study focused on two lead smelters within the Dallas city limits.
These smelters were implicated  as the sources of elevated con-
centrations of lead in the surrounding soils. The presence of high
lead is of considerable concern to nearby inhabitants.
  This report focuses on  the application of geostatistical tech-
niques to sampling around point sources. The application of geo-
statistical methods to smelter contamination had not been prev-
iously attempted. Many  of the questions concerning applicability
of geostatistics,  geometric and spatial  structure of the contam-
inant plume, modeling of the area, sampling strategies and gen-
eralization to future sites have been resolved. Geostatistical tech-
niques have proven to be very useful in the resolution of these

  Increased environmental awareness dictates that virtually any
type of industrial "smokestack," either  existing or planned, is
subject  to public scrutiny. Coal-burning  power plants and steel
mills are being blamed for acid rain problems. Other heavy metals
such as uranium that are emitted from coal-powered plants also
are causing concern. These emissions  pose a threat  to nearby
plant and animal life. Insights gained from modeling heavy metal
airborne contamination  hopefully can  improve models in these
other situations or provide a starting point  for understanding
these unique problems.
  The results and conclusions of  the Dallas Lead Project have
provided analytical tools and valuable insights into the nature of
contamination at other  locations.  This information then can be
applied to future modeling projects.
  This paper discusses the techniques applied to the modeling of
the Dallas project  in detail as well as the insights that have been
gained by  the authors. The U.S. EPA commissioned this study
as an experiment to determine applicability of geostatistical tech-
niques in an attempt to  develop more objective procedures. The
following conclusions are reached in the paper:

• Geostatistical  methods  are very objective  and effective  in
  modeling  contamination around  smelters.  Two separate
  smelters in the Dallas  area have been studied. Both sites have
  produced excellent  results and  nearly  identical geostatistical
  models indicating that other sites may follow a similar pattern.
• Geostatistical  modeling has many advantages over  traditional
  estimation techniques.  Its Best Linear  Unbiased  Estimator
  (BLUE) property provides more objective answers to questions
  that are difficult to provide otherwise. Results of studies such
  as this one influence  decisions  in cleanup  phases that  affect
  human health and safety and should be made based  on the
  best possible data. The results also will have significant legal
  ramifications and are subject to close scrutiny in courts of law.
  Thus the need for objectivity and reproducibility becomes in-
  creasingly important as issues enter the courtroom.
• Modeling with regionalized variable techniques permits the
  earth scientist to isolate different phenomena acting simul-
  taneously within the  same geographic area. Different patterns
  of dispersion from the smelter can be recognized and isolated
  from other  sources.  Trivial contributions from  secondary
  sources (e.g., auto emissions) can be identified. In this way
  liability can be correctly assessed.
• Objective methods to determine sampling sufficiency easily can
  be defined. Sufficient grid sizes to model and assess  the area
  with  an  acceptable level of  confidence  can be obtained with
  minimum sampling.  If too many samples already  have been
  taken, sampling can  be reduced to determine sufficient levels
  for obtaining both variograms and desired accuracy as a start-
  ing point for future projects. The technique known as the com-
  position of extension variances has been applied and discussed
  with respect to overall precision.
  The results can  be applied to the design of future smelters;
they also provide a method of limiting liability in environmental
impact  situations by isolating sources; finally, the results can be
used by the environmental and legal professions to further the
cause of public health and safety within a framework of cost con-
sciousness and objectiveness.

  During the last  quarter of  1980,  a  sampling  team from the
University  of Texas at  Arlington,  under contract with the U.S.
EPA's regional office in Dallas, Texas, collected samples of soil,
house dust, paint and tap water from a number of private resi-
dences  and schools.  These sample  locations were situated near
and in the general vicinity of two lead smelters and one battery
plant in Dallas,  Texas. In addition to the two smelter sites and
the battery plant, a fourth site identified  as the Reference area
(REF) also was selected as a control  area for sampling. The pur-
pose of this initial phase of the investigation was  to determine
whether additional effort was warranted, i.e.,  did  a  problem
  All four sampling sites were located within the Dallas city limits
and all except  the REF site were geographically located  in the
city's western section. Figure 1  shows the location of the two lead
smelters: RSR Corporation (RSR) and the  Dixie Metal Company
(DMC). General Battery Corporation (GBC) is located  near the
intersection of Bengal and Amelia Streets.  The GBC site was not
analyzed in the geostatistical portion of the study, however. The

                          Figure 1
     Location of the DMC & RSR Smelters and the Reference Area

REF site was selected for its similarity to the RSR,  DMC and
GBC sites in automotive emissions, population and ethnic back-

  Data collected by the University of Texas at Arlington have
shown some soils in the  immediate vicinity of the Dixie Metal
Company (DMC) and the RSR Corporation (RSR) to have lead
levels that exceed 1,000 ppm, believed to be the critical danger
level for exposure. Due to the presumed behavior of airborne
lead from smelting facilities such as these, the sampling area was
confined to  a 2-mile radius around each site. To better evaluate
the degree of contamination of these two industrially  exposed
areas, a site known as the Reference area (REF) was selected.
  The sampling scheme, except for the total number of samples
collected in  the reference area, was identical to that used at the
DMC and RSR sites.  In  addition, a 2-mile area surrounding a
third site, the General Battery Corporation (GBC), was sampled
using the same scheme.

  The information collected in sampling phase  1 (20 data points)
was used to construct a variogram for each of the three study
areas.  The  calculated and recommended  sampling scheme  to
identify the geographical extent and level of soil lead contamina-
tion for phase 2 of the project was as follows:
• The dimensions of the square grid should be 750 ft x 750 ft.
• The  total  area covered by the grid should be contingent upon
  the minimum lead concentrations of concern.
• The grid should be oriented along the axis of the plume.
• The grid used in the REF area should be identical to that used
  in the DMC and RSR areas.
• Soils within a radius of 100 ft of each grid intersection should
  be sampled using procedures identical to  those described in
  phase 1.
• Six samples should be collected for each site to be composited
  into a single sample.
• Duplicate samples should be collected at 5% of the grid loca-
• Samples should be separated into three  "splits" at 5% of the
  grid locations and sent to different laboratories.

  During the summer of 1980, the U.S. EPA collected more than
3,000 individual core  samples that were composited into more
than 500 grid samples throughout the three study areas. These
data are the foundation  of the subsequent geostatistical analysis
and mapping portion of the project.

  Variograms of the  Ln-transformed soil sample  values were
computed in each of the three areas. In each case, an average
variogram and four directional variograms were produced. The
four directional variograms have  been calculated along the east-
west  (0°),  north-south  (90°), northeast-southwest (45°)  and
northwest-southeast (-45°) directions. For each directional vario-
gram, an angular tolerance window of 22.5 ° (total width of win-
dow) around the given direction was allowed. A distance lag sep-
aration of 800 ft also was used. For the average variogram, which
considers all points, the angular tolerance was 180°.
  The results of the calculations were very good. Salient features
of the variograms include the following:
• All the variograms show  some kind of  continuity between
  holes. They all tend to increase with distance until they reach
  a sill value.
• Continuity in the DMC and RSR areas shows very similar pat-
  terns when compared to the reference  area. Modeled range
  and sill values bear this out.
• As expected, the REF area shows a much lower sill and a lesser
  degree of continuity.
                         Figure 2
    DMC Variograms of Ln-Transformed Lead Values for Phase 2
                                                                                   CONTAMINATED SOIL TREATMENT   93

                                                    t i-coo  Q I s i
                          Figure 3
     RSR Variograms of Ln-Transformed Lead Values for Phase 2

   Isotropic spherical models have been fitted to the experimental
average variogram curves. In all cases, they are the sum of a small
nugget effect (calculated from the duplicate samples) and  one or
two spherical structures of varying ranges. The curves of the fitted
models are shown as a solid line in Figures 2 to 4.
       ^ LCAO -   K' I
  Finally, one anticipated problem in the calculation of the vario-
grams did not arise. This "non-problem" was the possible pres-
ence of a drift or a  trend. A drift is defined in geostatistical
terms to be a systematic increase or decrease in  the data along
given directions.  This drift causes  problems in the variogram.
Modeling such a structure is difficult and also causes problems in
the kriging process. Special  forms of kriging (universal,  general-
ized IRF) have been created just to handle this problem. Although
the possibility of a drift seemed reasonable, one did not arise.

  The  soil  sample data were kriged in  each  of the  three areas
using their respective variograms. The kriged area  has been made
large enough to include three  reference  locations of geographic
interest. These locations have been used for map orientation.
  The contour maps resulting from  the kriged grid are shown in
Figures 5 to 7. The round symbol in the center represents  the lead
smelter. The contour lines show the estimated soil lead concentra-
tion in ppm.  In the DMC area (Fig. 5),  a highly visible  concen-
tration or plume has developed around and to the  north and east
of the source. The large number of concentric contours encircling
the  smelter shows a steep gradient of rapid change in a short dis-
tance between a low of 200 ppm outside and a high of 3,000 in-
  The effects  of  flooding by the  Trinity River are well  defined
(to  the north) as a distinct area of low lead concentration which
can be seen along the flood  plain of the river. On either side of
the  river, the concentrations are noticeably higher. Several pos-
sible anomalous sample values also occur to the north of the river.
In at least one case a sample whose concentration was 10,400 ppm
contamination was found later to  have been exposed to automo-
tive contamination.
  In the RSR area, the smelter is located within the major con-
taminant plume that disperses rapidly east-west and more slowly
north-south (Fig.  6). The  southern end of the plume also contains
                                                                    SCO	2700	SJOO	7900	»»OO
                          Figure 4
    REF Variograms of Ln-Transformed Lead Values for Phase 2
                                                                    »002700^00        TioO       9900
                          Figure 5
         Isomap of Estimated Lead Values in the DMC Area

                29OO       S3OO      7TOO
                                                                            1500   3OOO  45OO   60OO   75OO   90OO  IOSOO
                            Figure 6
          Isomap of Estimated Lead Values in the RSR Area

a possibly anomalous value that has the effect of extending the
plume in this direction. This finding suggests that the true dis-
persion nature of the plume is primarily north and slightly west
from the smelter.
  As expected, the REF area shows no systematic areas of lead
concentration (Fig. 7). Local highs and lows  fluctuate randomly
and at considerably reduced levels overall from those  seen in the
DMC and RSR sites. There is no evidence to suggest  that this is
anything other than a background lead level map.
  Accompanying each isopleth map is a second contour surface
that  contains isovalues for the kriging standard deviation (Figs.
8 to  10). The values shown are the logarithmic kriging deviation
multiplied by a factor of ten (to facilitate plotting). The "swiss
cheese" effect noted here is common for cases such as this where
the sample grid spacing is a relatively large percentage of the
range of the variogram. Because the quality of estimation is solely
a function of geometry and the variogram, estimates are better as
they approach a sample point.
37OO       6IOO       85OO       \O9OO
            1300        3700       6100       6500       I09OO
                           Figure 7
         Isomap of Estimated Lead Values in the REF Area
                                                                   9000 -
                                                                   3OOO -
                                                                            I50O    3OOO   450O   6OOO   75OO    9OOO  K55OO
                                                                        Figure 8
                                                      Isomap of the Estimation Variance in the DMC Area
                                                                                     IOIOO     "TOO
                                                                                                                        THE MKTRft TRACK
                                                                                                           IOIOO     II7OO
                                                                                              Figure 9
                                                                            Isomap of the Estimation Variance in the RSR Area
                                                                           /?      °
                                                                           y * iLLiNoia AVC a
                                                                                                          6500     IO9OO
                                                                                 •ucxtrr AVE
                                                                                                                      JNNTV»LE IT.
                                                                                                           85OO      IO9OO
                                                                       Figure 10
                                                      Isomap of the Estimation Variance in the REF Area
                                                                                       CONTAMINATED SOIL TREATMENT    95

  The results of this first mapping stage proved that kriging and
geostatistical  techniques apply very well to soil data  and point
source contamination. The second stage of this study addressed
the concerns of the U.S.  EPA and others as to the adequacy of
sampling: have enough samples been taken to assure reasonable
level of certainty?  Also, what changes might be made in future
studies of similar phenomena? These questions will  not be ad-
  The techniques used to answer these questions will be the analy-
sis of extension variance and the use of reduced data sets to de-
termine the effects on the variogram. The results of the prelim-
inary geostatistical analysis provided models of the spatial con-
tinuity structure in each area through their respective variogram
models. These  models will  now be  assumed  to be the "true"
underlying spatial structure. The data sets of 206 and 208 samples
in the DMC and RSR sites respectively will be taken as the "uni-
verse" of data.
   The procedure, then, is to reduce the level of sampling  in the
areas of study and to  recompute the variograms using a smaller
data set. Multiple data sets will be produced and analyzed at each
level of reduced sampling to minimize any  possible selection bias.
The new parameters of the variograms then will be compared to
the "true" values to see how accurately the subset can predict the
"real" structure.  This information from  small data sets can  be
used to design future two-phase sampling schemes.
   The two smelter areas have been divided into equal sized grid
cells from which one sample per cell has been taken at random.
This random selection has been  done by  computer to eliminate
human  bias.  Since the "true" variograms are isotropic (equal
zone of influence in all directions), the sampling grids were de-
signed to be  isotropic (square) so that  this relationship can  be
   From the variograms,  it is  possible to  quantify the  variances
of polygonal (block) extension errors and then combine  them,
since they are the variances of independent variables. This has
been used as a check on total precision.
   Small sample subsets were selected in  both the DMC and RSR
areas on grids of 11 x  11, 7 x 7, 5 x 5 and 4x4. These represent
sample  sizes of approximately 100, 50, 25 and 16 respectively.
The 11x11  grid was selected as the first  grid  size with the idea
that subsequent sizes would be enlarged or reduced as needed de-
pending upon the outcome of this first attempt. Ten sample sub-
sets were taken using the same ten random seeds in both smelter
areas. The ten new data sets in each area then were variogram-
med, their sample locations  were plotted  and their  histograms
were plotted.
  The 20 variograms then were modeled.  The  results showed re-
markable similarity to the original data  sets both visually and in
variogram parameter comparison. The results of the 11 x 11 grid
data were so successful that the data set was again halved by us-
ing a 7  x 7 grid.  Excellent variograms were produced although
more variation was noticed in the model  parameters. The RSR
models  tended  to  have a lower  sill  than  the  true one, but the
ranges were generally the same as for the  true  model. The  DMC
models showed on the average the same  sill but generally showed
a shorter range. Because the  object of this exercise  was to de-
termine sampling grid sizes from small initial data sets, the range
was the  most  important parameter. For  this grid size, both areas
provided accurate  ranges with the DMC  models  being a little
more conservative.
  Again the data set was  halved by going  to a  5 x 5 grid. At this
level of sampling, the directional variograms had deteriorated to
such a point  that  they were valueless.  A  lack of data was the
cause.  However,  the average variogram  still could be used to
advantage  as  good structures  still were found.  Sill  variance
showed a much greater fluctuation, with a number of low-valued
variances where sampling failed to pick up high values near the
smelters. Ranges generally were shorter than the  "true" ranges.
This corresponded to the shorter range found in phase 1  of the
sampling where only 20 samples were taken. A few very  long
range values did appear, but they resulted mainly from an attempt
to fit a good model at short distances where the variance  rose
quickly for a time and then slowly crept toward the sill.
  The  final level of sampling was produced on a 4 x 4 grid. Again
directional  variograms  were of no value. The  sill values  in the
RSR area showed a large fluctuation,  but the ranges were quite
stable and not too far from the  "true" parameters. In the DMC
area, sills were generally near or above the "true" parameters
and the ranges were at or below  those expected. Overall, this grid
produced about the same quality estimates as the 5x5 grid and
showed slightly better stability of ranges.
                                 DMC AREA

                   0    497    H   14 19
                      NUMBER  OF GRID CELLS

                          Figure 11
       Global Precision vs. Sampling Density in the DMC Area

                               RSR AREA
                         497     II    14 19
                       NUMBER OFGRIDCELLS

                          Figure 12
        Global Precision vs. Sampling Density in the RSR Area

  An estimate of global precision provides a measure of the cer-
tainty with which the global mean is known. This value is anal-
ogous to the "standard error of the mean" in classical statistics.
In the geostatistical sense, obtaining a high value for precision of
the global mean suggests that local estimates also will remain
quite accurate. This condition is not strictly true in all local cases
but offers a starting point for decision making.
  The precision for symmetric grids of sides 4 through 15 have
been calculated and have been plotted in Figures 11  and 12. The
14 x 14 and 15 x 15  intervals represent the precision when con-
sidering a full sample set of 206 or 208. The graphs show that in
terms of global estimates the sampling has reached  the point of
diminishing returns. Both figures indicate that further sampling is
unnecessary.  What also is critical here is  that variograms have
been produced for even the smallest of data sets.
  Geostatistical techniques have proven to be an extremely valu-
able tool in the analysis of pollutant contamination. A strongly
identifiable plume of lead is associated with smelter source. No
such structure is found in the REF area. The plume is marked by
a central core of concentrated high values, generally within 1 mile
of the source, with a continual decrease in lead values to the
borders of the areas. It appears certain that these two smelters
were indeed the source of the contamination  and that automo-
bile emissions are of negligible contribution.
  The smelter areas have been adequately sampled with the 800-ft
grid. The  range of the variograms was  five to seven times this
sampling interval and proved quite sufficient to model the spatial
continuity of the lead values.
  Global estimates of precision for the second stage sampling
campaign  are very conservative and  indicate  that additional
sampling would have been of little value. Variogram parameters
remain surprisingly constant down to extremely small levels of
  Two-stage sampling, though already in regular use in the en-
vironmental industry, can be greatly enhanced by the use of geo-
statistical technique. Optimal sampling geometries can be  devel-
oped for a variety of situations and constraints. As a result, cost
benefit  curves can be produced to select development drilling
grids. These analyses will provide a measure of confidence and
                                                                                     CONTAMINATED SOIL TREATMENT    97

                 Innovative Application of  Chemical  Engineering
                   Technologies  for Hazardous  Waste Treatment
                                                   Robert  D. Allan
                                                  Michael  L. Foster
                                                   IT Corporation
                                                Knoxville, Tennessee
  Regulatory and public pressure are creating an increasing de-
mand for more innovative methods of hazardous waste manage-
ment. In response to this need, some proven technologies are
being adapted to provide alternatives to the more classical remed-
iation techniques in certain applications. This paper reviews the
site-specific parameters that need to be considered before select-
ing one of these alternatives. The operational details of the can-
didate technologies are discussed, and specific examples of appli-
cations are given.

  Application of the RCRA and CERCLA statutes often creates
the need for corrective action, particularly with regard to ground-
water treatment and surface impoundment closure. Of particular
interest  to industries are the on-site treatment alternatives due to
the cost and liability factors of land disposal and the lack  of con-
venient incineration options.
  This paper concentrates on proven treatment  solutions avail-
able to meet the needs of corrective action situations.  Particular
emphasis is placed on those technologies offering the opportun-
ity  for cost-effective,  on-site treatment or waste volume reduc-
tion. Included is a review of  technologies  utilized in  mobile
and/or in situ applications.

  Unfortunately, no single treatment technology is going to be
universally  applicable  for  all corrective action  situations. A
thorough evaluation of the site-specific  parameters involved is
necessary in order to choose the most technically suitable  and
cost-effective treatment technology for the job.  Because major
corrective  action  programs, especially  those involving  aquifer
restorations, are  likely to take several years to accomplish, the
investment  of time and effort to properly evaluate the situation
initially  will more than pay for itself in the long run.
  Among the major site-specific parameters that should be de-
fined are those discussed below.

Contaminant Composition and
  An accurate analysis of the site contamination is  important.
The contamination could be a single component from an under-
ground solvent storage tank leak or a wide variety of compon-
ents from  a disposal site leachate. The  level of contamination
could range from  low >jg/l to several hundred mg/1. The treat-
ment costs of technologies such as carbon adsorption and oxida-
tion are dependent on concentration  and type  of compound,
while costs for technologies such as air stripping are more closely
related to the treatment rate and desired removal efficiency.
Desired Cleanup Criteria and
Effluent Disposition
  Stringent cleanup criteria (e.g., drinking water standards for
aqueous streams) will eliminate some technologies from consid-
eration immediately,  but  it must be recognized that different
levels of cleanup are likely to be required for many remedial pro-
grams. The ultimate disposition of the effluent for aqueous treat-
ment will have significant  bearing on the required effluent qual-
ity. If the treated  stream can be discharged to the sewer for
further treatment by a municipal waste facility (POTW), effluent
guidelines are likely to be less strict than if the stream is to be re-
charged  to an aquifer or discharged directly to  a surface water
source under an NPDES permit.

Volume to be Treated
  Treatment volumes for remedial programs can vary widely. A
small volume may be treated best by a technology with low capital
costs and high operating costs, while economic remediation of a
larger magnitude is more likely  to be accomplished utilizing  a
technology with low operating  costs. An evaluation of treatment
options  should  include consideration of leasing equipment to
treat small volume sources.

Fate of Treatment Byproducts
  Local  requirements for disposition of treatment byproducts
will have a major impact on the economics and viability  of vari-
ous treatment technologies. Air stripping  can be an extremely
cost-effective technology if the organic-laden air can be  emitted
directly to the environment; but strict air emissions standards may
require that air stripping be followed by a vapor-phase  adsorp-
tion system, significantly increasing treatment costs. Similar con-
siderations  exist  with  the generation of  potentially toxic by-
products from chemical oxidation processes and the possible need
for nutrient removal (phosphorus and nitrogen) from the  effluent
of biological  oxidation processes. The economics of activated
carbon adsorption  (especially on a throwaway carbon basis)
can be affected by  the accessibility of an approved disposal site
for the contaminated carbon.

Utilities Availability and Cost
  The selection of certain technologies will be influenced by the
availability of utilities such as  steam at the treatment site. Like-
wise, some processes such as UV-catalyzed  oxidation can require
a substantial source  of  electrical power.  The  energy-intensive
processes are likely to be more attractive in the southern part of
the country where electrical costs are approximately $0.05/kWh
than on the West Coast where costs can approach $0.15/kWh.
  A clear definition of the goals of a remedial action program
along with an understanding of the above site-specific parameters

and the technical limitations of the technologies under considera-
tion  will allow  selection  of the most  cost-effective treatment

  A number of technologies are potentially applicable to remed-
iation  programs. Some are best suited  for specific applications
and others may be eliminated from widespread consideration be-
cause of the inability to meet stringent discharge standards or
other technical limitations. In this section, the technologies with
the greatest potential in remedial actions are identified and spe-
cific applications are discussed.

  Incineration, or thermal destruction, is an effective way to de-
stroy organic liquids and sludges and organics in a solid matrix.
It can  be used for a relatively broad waste profile depending on
the type of incineration system utilized. Incineration is defined
generally as a controlled, high-temperature, oxidation reaction.
The  end products of the incineration reaction are generally car-
bon  dioxide and water from hydrocarbon wastes; but industrial
and hazardous wastes also can generate hydrogen chloride, sulfur
dioxide and less desirable combustion byproducts depending on
the nature of the wastes, the operating features of the specific in-
cineration system and the compatibility of the waste profile  and
incineration system.
  The technology of high-temperature incineration is in  wide-
spread use, both at private and commercial installations. It often
is considered in  remedial  action alternative evaluations, but the
waste volume must be substantial to justify the expense of permit-
ting and building an on-site unit. For more limited volumes, com-
mercial incineration  facilities offer the advantage of destruction
of the wastes involved.

Thermal Desorption
  Thermal desorption removes organic  chemicals from soil  and
solids by heating the soils to temperatures sufficient to convert the
organics to vapors and then holding the soil at that elevated temp-
erature long enough for the removal process to occur. Desorbed
organic and water vapors  can be recovered by solvent scrubbing
or condensed and collected in an accumulator where the  water
and  organics are separated into their respective phases. Water
can be treated with activated carbon before discharge,  while the
organics phase can be sent to a commercial incinerator or treated
chemically or photolytically.
  Thermal desorption is effective for any organic  contaminant,
with more volatile compounds such as solvents, being more read-
ily removed. The types of compounds for which this technology
is expected to be applicable range from chlorinated  solvents  and
gasoline/fuel components  to  phenols,  aromatic hydrocarbons,
PCBs  and dioxin. The  concentration of the contaminant does
not have a  significant influence on  the applicability. Although
the soil characteristics will impact the operating rate and process
design requirements, thermal  desorption is amenable to a wide
range of soil/site conditions.
  IT  has  developed and demonstrated thermal desorption in
bench  scale and pilot scale equipment. Laboratory testing  has
been performed  on a variety of soil types containing such chem-
icals as Agent Orange (containing chlorinated dioxins), penta-
chlorophenol and creosote materials (PNAs).  Pilot scale work,
conducted  as part of the  Air Force's  Environmental Restora-
tion Program, will be reported at an upcoming American Chem-
ical Society meeting.1 Briefly, a pilot plant was assembled to treat
up to 100 Ib/hr  of soil contaminated with 2,3,7,8 TCDD from
Agent  Orange. The desorber was operated at 460 to 560 °C. The
vaporized contaminants were captured by a solvent scrubber and
destroyed in a UV photolysis reactor.  Soil analysis  showed  a
greater than 99% reduction in the TCDD concentration and resid-
ual soil concentrations of less than 1 ppb.
  The thermal desorption process offers several advantages,  as
compared to thermal  oxidation,  for decontamination of organ-
ically contaminated soil. These advantages include lower temper-
ature operation resulting in less energy consumption and simpler
design criteria, relative ease of regulatory approval due to classif-
ication as a physical/chemical treatment process and the ability
to easily skid or truck mount the equipment  for mobile treat-

UV Photolysis
  Ultraviolet light can catalyze normal chemical oxidations by
creating a supply of "free radical" species. IT has applied this
technology to the destruction of dioxin at the Syntex Agribusi-
ness plant in Springfield, Missouri.2 The dioxin destroyed was in
a distillation residue which was a byproduct  of previous pro-
duction of  hexachlorophene by the Northeast Pharmaceutical
Company. The process involved extraction of dioxin from the
residue  using a common  solvent and  then destruction  of the
dioxin in reactors equipped with  10-kW,  high  intensity  ultra-
violet lamps. Approximately 13 Ib of dioxin were destroyed using
this process.

Chemical/Biological Oxidation
  Oxidation technologies  use  chemical  or biological  means to
completely oxidize organic contaminants to carbon dioxide and
water, or partially oxidize them to non-toxic intermediates. Most
of the oxidation options are best suited to specific applications.
  Chemical  oxidation  accomplishes  detoxification  of  waste
streams through the addition of chemical oxidation agents. Near-
ly all chemical oxidation processes use either chlorine or oxygen
as the oxidizing agent.
  Chlorine oxidation  is of questionable value  in most remedial
applications because of the difficulty in assessing the impact of
potentially toxic chlorinated byproducts. Oxygen itself has very
limited effectiveness but may be  of use in the form of ozone or
hydrogen peroxide.
  As an example of an application of chemical  oxidation, IT has
utilized hydrogen peroxide to successfully treat  a sizeable spill of
formaldehyde. The spill resulted  from a leak in an underground
pipeline and had contaminated both the subsurface soil  and the
groundwater. Because the  spill contamination was below a ship-
ping and receiving area, common cleanup techniques involving
excavation would have caused an expensive disruption of plant
operation. Instead the formaldehyde was chemically oxidized  in
situ by injecting a hydrogen peroxide solution  into  the contam-
inated zones. Formaldehyde contamination levels in the soil, orig-
inally as high as 8% (80,000 ppm), were reduced to 15 ppm or less
by this method.
  Biological oxidation uses active microorganisms to biodegrade
organics to acceptable forms. The two major forms of biological
treatment are aerobic (which produces carbon dioxide and water)
and  anaerobic  (which produces  carbon dioxide  and methane).
Biological treatment is getting increased attention as a remedial
alternative because  of its  potential for in situ treatment. Bio-
reclamation basically is the use of indigenous soil bacteria to de-
grade organic contaminants. Nutrients,  such as oxygen, and spe-
cific biological cultures can be added to enhance the degradation.
  Vandalism of a rail car  parked on a railroad siding in Ukiah,
California resulted in  an opportunity for the application of bio-
logical oxidation to an emergency situation. Twenty thousand
gallons of a strong formaldehyde solution drained from the rail
                                                                                                  ON-SITE TREATMENT   99

car into and through the rail-track ballast and into the associated
drainage system. Initial efforts focused on containing and isolat-
ing the spill.
   Once the spill had been controlled, the more difficult task of
decontamination  of the affected area  was addressed. The rail-
way company was vary concerned that normal excavation pro-
cedures would disrupt rail service on the adjacent mainline track.
Thus,  IT developed an alternative approach which initially in-
volved treatment of the  area with alkaline  hydrogen peroxide.
This chemical oxidation  process was successful in reducing the
formaldehyde concentrations to a  level where biological oxida-
tion could be used.
   The biological process made use of a portable aeration tank, a
spray system and the railroad ballast itself. Liquid was pumped
from a sump dug next to the ballast, passed through the aeration
tank and sprayed back onto the ballast. The system was inocu-
lated with  a  specially cultured microorganism that  degraded
formaldehyde but was free of pathogens. The ballast, being com-
posed of coarse rock, supported the growth of the biological med-
ium in a manner similar to a trickling filter. Nutrients were added
to the system as needed. Over a 25-day period, the system suc-
cessfully reduced the formaldehyde concentration from several
hundred ppm to less than 1 ppm.'

Carbon Adsorption
   The use of activated carbon adsorption for aquifer  restora-
tion programs has drawn widespread attention. Several literature
sources point to  its ability to achieve exceptionally good efflu-
ent quality, and the U.S. EPA has endorsed it as  the preferred
treatment  method for meeting drinking water standards.  This
section discusses three basic ways in which carbon can be used:
•  Throwaway carbon
•  Thermal regeneration
•  Nondestructive regeneration

Throwaway Carbon
   One way to consistently ensure good effluent quality is to use
activated carbon absorption on a once-through carbon  basis.
Virgin carbon is capable of removing a broad range of organic
contaminants to low ug/1 levels. A once-through carbon  adsorp-
tion system is easy  to operate,  requires a minimum of operator
attention and its capital cost requirements are relatively low.
   Unfortunately, carbon replacement costs associated with once-
through carbon adsorption systems are very high. The large treat-
ment volumes and/or high concentrations usually associated with
remedial programs result in a high carbon consumption  rate. In
addition,  hazardous substances, when loaded onto activated car-
bon, make the carbon a hazardous waste, requiring disposal in
an approved hazardous waste facility.

Thermal Regeneration
   The most common regeneration technique for activated car-
bon is thermal oxidation, usually accomplished  in a multiple
hearth, fluidized-bed or rotary-kiln furnace. A thermal regenera-
tion unit can be built at the treatment site, but the level of carbon
consumption associated  with most remedial programs  usually
makes it more economical to utilize an off-site thermal regenera-
tion service.
  The advantages of activated carbon adsorption with a  thermal
regeneration service are low capital requirement and ease of oper-
ation,  as with throwaway carbon,  but  the disadvantages of
thermal regeneration are numerous.  Using a thermal regenera-
tion service basically  amounts to "renting"  the carbon. How-
ever, in many cases, the  cost per pound of carbon for  thermal
regeneration is only slightly less than that for purchasing virgin
carbon. Additionally, carbon losses associated with thermal re--
generation are typically 5 to 10% per cycle due to handling losses
and  carbon that  is destroyed during the regeneration  process.
However, thermally regenerated carbon can be expected to have
different performance characteristics and reduced removal effic-
iency as compared to virgin carbon.
  A major problem associated  with thermal regeneration is that
many hazardous substances cannot be thermally regenerated for
one reason or another. Because the hazardous substances are de-
sorbed from  the carbon and destroyed in an afterburner during
thermal regeneration, air  permits for thermal regeneration sys-
tems often contain restrictions on the types of materials that may
be accepted.  Consequently, thermal regenerators often refuse to
deal with carbon that  has been used to adsorb toxic materials
such as PCBs.

Nondestructive Regeneration
  There are three basic ways  that granular activated carbon can
be nondestructively regenerated:
• Using pH shift for weak acids or bases
• Using steam for volatile organics
• Using a solvent for a wider variety of organics
  Steam regeneration, which would be applicable in many remed-
ial situations, is accomplished by passing steam through a spent
adsorber to a condenser and  then to a decanter where the con-
densate  and  immiscible organics  are separated. Carbon losses
associated with nondestructive regeneration are significantly less
than those associated  with thermal regeneration. The carbon is
not physically altered by the nondestructive regeneration process,
and if the carbon  adsorbers are made of the proper material, the
regeneration  can  be accomplished  in  situ, eliminating carbon
handling losses.
  The process is relatively easy to operate and the nondestruc-
tive  nature of the process allows  organic recovery, if  desired.
Even if recovery is not desired,  the disposal requirements are re-
duced from several thousand pounds of contaminated carbon to a
few gallons of organic material.
  IT has evaluated the applicability of carbon adsorption systems
with nondestructive regeneration in a variety of remedial  situa-
tions. A recent feasibility study compared  air stripping, air strip-
ping with vapor-phase adsorption, adsorption using virgin car-
bon, adsorption  using steam-regenerated carbon,  steam  strip-
ping and  chemical oxidation  using UV peroxide for an aquifer
restoration. The critical design parameters were: a contaminated
groundwater  concentration of  30  mg/1 or halocarbons includ-
ing EDB (1,2-dibromoethane);  a maximum flow of 200 gal/min
and  an  average flow of 100 gal/min; and a mandated  effluent
concentration of ^1 ug/1 of each halocarbon present.'
  Details on the  result  of this evaluation can be  found in the
referenced paper. Based on these design parameters, the costs of
air  stripping and steam-regenerated  carbon adsorption were
basically equal and fell in  the range of $300-400/day of system
operation. Subsequent laboratory work,  using samples of the
actual groundwater, verified that  carbon adsorption with  steam
regeneration could achieve acceptable effluent quality and  main-
tain a stable working adsorption capacity.
  Laboratory evaluation of the air  stripping option was post-
poned  until the  issue of  exhausting the  vaporized organics to
the atmosphere could be resolved. This issue is frequently a re-
striction in the application of air stripping.

Air Stripping
  Air stripping is a contaminant removal technique based on con-
centration differentials between a liquid phase and a contacting
gas phase. As air is contacted  with a wastewater stream in a strip-
ping tower, the concentration differential drives the organic con-

taminant from the liquid to the gas phase. The major advantage
of air stripping is its low overall treatment cost. Both capital and
operating costs  requirements are low compared to most other
  The key to air shipping's low overall treatment cost is the
assumption that it can stand alone as a treatment technology. In
many cases, however, air emission standards will require that air
stripping be used in conjunction with a vapor-phase adsorption
unit, significantly affecting the cost-effectiveness of the technol-
  Air stripping  is most applicable to compounds  of  a volatile
nature with relatively low solubility in water. Chlorinated organ-
ics such as trichloroethylene, 1,1,1 trichloroethane and aromatic
compounds such as benzene and toluene are  good examples of
chemicals that can be successfully removed by air stripping.  The
overall applicability of the technology can be  expanded to com-
pounds with lower Henry's law constants by preheating the water
prior to its entering the stripping column.
  IT has evaluated air stripping a number of times in treatability
evaluations for a wide variety of clients. In a study conducted for
the American Petroleum Institute, IT engineers conducted side-
by-side, bench-scale evaluations of air stripping and carbon ad-
sorption with steam regeneration to remove  dissolved gasoline
components  from  groundwater.5 Air stripping effectively  re-
moved the aromatic gasoline components from groundwater even
at the normal groundwater  temperature  of SOT.  However, as
would be expected, air stripping was ineffective in removing more
soluble components of gasoline such as tert-butyl alcohol.

  As the expense and regulatory limitations of waste disposal be-
come increasingly important factors, the innovative option of
mobile, on-site  treatment has become an  important  remedial
alternative. Many of the technologies described above can be, and
have been, adapted to mobile applications.
  The U.S. EPA and a limited number of commercial firms  cur-
rently operate mobile incineration systems.  A number of aque-
ous treatment technologies have been mobilized  for on-site re-
mediation including activated carbon, air stripping and ion ex-
change. In addition, mobile  equipment is available to dewater
sludges and reduce the volume of material that ultimately must be
sent to disposal or destruction facilities. These options include
mobile centrifuge and mobile belt filter press equipment.

  A number of technologies  exist and have been proven  in re-
medial action situations. Experienced, knowledgeable personnel
in the hazardous  waste field  are  able  to efficiently evaluate
options for  corrective action programs, but, in order  to  do so
properly, they must evaluate the site-specific parameters involved
and fully  understand  the constraints, options and goals of the

1. Helsel, R., Alperin, E.,  Geisler, T., Groen,  A., Fox, R., Stoddart,
   T. and Williams,  H., "Technology Demonstration of a Desorption/
   UV Photolysis Process for Decontaminating Soils Containing  Herbi-
   cide Orange," To be presented before the Division of Environmental
   Chemistry, American Chemical Society, April 1986.
2. "Destroying Dioxin: A Unique Approach,"  Waste Age, Oct. 1980,
3. Sikes, D.J., McCulloch,  M.N. and Blackburn, J.W., "The Contain-
   ment and Mitigation of a Formaldehyde Rail Car Spill Using Novel
   Chemical and Biological In Situ Treatment  Techniques," Proc. of
   the Hazardous Materials Spill Conference, Nashville, TN, Apr. 1984,
4. Parmeli, C.S.,  Allan, R.D. and Mehran,  M.,  "Steam-Regenerated
   Activated  Carbon: An  Emission-Free, Cost-Effective  Groundwater
   Treatment Process," Paper presented at American Institute of  Chem-
   ical Engineers Annual Meeting, Chicago, IL, Nov. 1985.
5. Parmeli, C.S. and Allan, R.D., "Treatment Technology for Removal
   of Dissolved Gasoline Components from Groundwater," Third Na-
   tional Symp. and Exposition  on Aquifer Restoration  and Ground-
   water Monitoring, Columbus,  OH, May 1983.
                                                                                                  ON-SITE TREATMENT    101

                            Field  Studies  of In Situ  Extraction
                          and Soil-Based  Microbial Treatment
                               of  an Industrial  Sludge  Lagoon

                                                 David S.  Kosson
                                               Erik A. Dienemann
                                        Robert C. Ahlert, Ph.D., P.E.
                                         Rutgers,  The State University
                          Department of Chemical and  Biochemical Engineering
                                            Piscataway, New Jersey
  Over a period in excess of 10 years, several industrial sludges
were disposed of by landfilling in a surface impoundment. The
resulting  lagoon contains  more  than  30,000  yd3  of sludge.
Leachate from the  sludges has impacted the local groundwater,
thus requiring remediation. Laboratory  studies indicated that in
situ extraction with aqueous sodium hydroxide is a viable method
for removal of organic materials from the sludges. Treatment of
recovered extract was accomplished through use of a soil-based
microbial treatment system. Sequential  aerobic and anaerobic
biodegradation in the laboratory reduced TOC by greater  than
  A field, pilot-scale treatment system has been designed,  con-
structed and operated to demonstrate feasibility of the proposed
renovation program. Field results indicate that exhaustive extrac-
tion of the sludges is possible. Up to a 50-fold increase in the rate of
removal of contaminants  over natural processes can be achieved.
Biodegradation treatment efficiencies in excess of 95°'o have been

  Over a decade or more,  several  sludges  were  disposed of  by
landfilling in a surface  impoundment.  During  this period  of
operation, the composition and  rate of deposition of sludges
varied. The resulting lagoon contains more than 30,000 yd1  of
sludge. The principal sludges in the lagoon are primary (lime neu-
tralized) and secondary (biological) sludges from treatment of ef-
fluent from diverse chemical  manufacturing  operations.  The
sludges range from solid  to gelatinous in physical state and are
layered in the lagoon. Leachate from the sludges has impacted the
local groundwater.
  Cleanup of the lagoon  is viewed a,s two interrelated problems.
The first problem is the removal of contaminants from the lagoon
without major excavation. The second problem  is treatment  of
the stream containing the stripped contaminants, including  both
organic and inorganic species. To achieve these goals, a process
corsisting of in situ extraction of the sludges followed by on-sile
treatment of recovered extract has been developed.
  Laboratory results have indicated that exhaustive  leaching of
the sludges with an alkaline aqueous extractant (sodium hydrox-
ide) is possible.1 Natural leachate and the sludge extracts can  be
treated effectively  for total  organic  carbon  (TOC) removal
through  use of a soil-based aerobic/anaerobic  microbial  treat-
ment process employing an acclimated mixed microbial popula-
tion in a soil matrix.2 Aerobic biodegradation occurs near the soil
surface where oxygen is available. Anaerobic biodegradation oc-
curs at greater depths. Conversion  of organic solutes to carbon
dioxide and methane is the ultimate result. Treatment efficiencies
in excess of 99% were attained.'
  Preliminary design estimates for implementation of these pro-
cesses indicated that renovation  of the lagoon could be ac-
celerated  to a period of less than 5 years with substantial cost
reductions compared to traditional treatment technology. Thus, a
field pilot plant  was designed, constructed and  operated for
demonstration of the process.

  The pilot system consisted of several sequential process steps
(Fig. 1). The first process step is the extraction of sludges present
in a representative section of the lagoon. Sodium hydroxide solu-
tion is mixed batch-wise in a 200 gal process tank (Tank 1). This
solution is applied to sludges present in the Extraction Bed. The
solution either can be applied to the surface of the Extraction Bed
through perforated pipe or injected  into the sludges through six
well-points. Extract is recovered from the Extraction Bed through
Wells 1 and 2. Extract removed from each well is pumped through
a basket strainer and cartridge filter and into a  1,000 gal storage
tank (Tank 2).
  The  second  process  step is adjustment  of pH, dilution  if
necessary and addition of  nutrients  to extract stored in Tank 2.
This process occurs as a continuous  process in Tank 3 which is a
200 gal process tank, divided into  four equal sections by two
overflow  baffles and one  underflow baffle.  Extract is pumped
from Tank 2 to the first chamber of Tank 3. Carbon dioxide  is
bubbled through the extract to adjust the pH to between 7.0 and
7.5.  Extract passes under an underflow baffle  to the second
chamber  of Tank  3. In  the second chamber, floe created  by the
recarbonation process is allowed to settle. Clarified extract passes
over an overflow baffle into the third chamber of Tank 3. In the
third chamber, dilution with  recycle or potable water and addi-
tion of nutrients  occurs.  Finally, extract underflows into the
fourth chamber from which it overflows and is applied to the sur-
face of the Treatment Bed.
  The third process step is treatment of the modified extract (ef-
fluent  from Tank 3) through an  aerobic/anaerobic  soil-based
microbial treatment process. The treatment process consists of a
soil bed in which an aerobic microbial population is maintained in
the upper region and an anaerobic microbial population is main-
tained in the lower region.1 Extract  applied to the surface of the
Treatment Bed  percolates through the soil  bed where it  is
biodegraded. Effluent  from  the Treatment  Bed  is  recovered
through Well 3. Recovered effluent is pumped to a 500 gal storage
tank (Tank 4) from which it can be recycled onto the Treatment
Bed, recycled to Tank 1 or discharged.
  The  entire process was designed  to be sufficiently  flexible  to
compensate  for fluctuations in either the extraction or treatment
processes as well as weather conditions. Ranges of expected condi-
tions and flow rates for each process step are presented in Table 1.

                                                       * BLEED 
- n




. ,



or^fr VCNTdWD
rxwii *

ii l



i J




^-— GRA
^1C ^•r°00

JIIIUHVilllllllllll [&&
KOTO.** LAC«» ..J[J?T^~
                                                                                              i:  O
                                                           Figure 2
                                            Extraction and Treatment Bed Construction
placed in the center. An additional 2 ft of sand were backfilled
around the well. Two 1-in. methane vents were installed on top of
the sand, and the remainder of the bed was backfilled with 5 ft of
soil mixture.
  The soil mixture used to backfill the Treatment Bed consisted
of local top soil,  sand  and  granular activated carbon IS: 1:1  by
volume, respectively. The soil mixture was mixed  on-site  by a
bucket loader prior to backfilling.
  Modified extract is applied to the surface of the Treatment Bed
through two 0.5-in. perforated PVC pipes. Two water level sen-
sors are located on the surface of the bed to prevent flooding. Ef-
fluent from the process is collected through Well 3 which contains
a stainless steep deep well pump, a low water level sensor and a
high water level sensor.  Water collected from Well  3 is pumped
directly to Tank 4.

Process Area Design
  The Process Area consists of  a 30  x  35-ft compound  located
orr top of shale fill approximately 100 ft east of the Extraction and
Treatment Beds at an elevation of approximately 10 ft above the
surface of the lagoon (Figs. 3 and 4). Tanks 1,  2, 3 and 4, a
storage shed for CO2 cylinders and nutrient solution, electrical
power supply and controls, a wash basin and safety shower are
located within the compound.
  Interconnections between tanks and the Extraction and Treat-
ment  Beds consist of PVC  solvent-cemented pipe. All piping,
where possible, is buried in shallow trenches approximately 1  ft
deep and backfilled with top soil. The surface of the compound is
covered with roadstone.

Process Controls
  The process controls consist of four operationally independent
controllers. Three of the controls are identical units,  i.e., one
pump controller each for Wells I,  2 and  3. The well pump con-
trols are a high water level sensor and a low water level sensor
(float switches) in each well.  Liquid rising to the level of the high
water level sensor activates the pump in that well. When the liquid
level falls below the level of the low water level sensor, the pump
is deactivated. The actual operating time of each pump is ac-
cumulated on a timer located on the control panel.

                               TOP OF BERM
                                        •TO UTILITIES
                                         (POWER AND WATER)
                                                                         TO DISCHARGE
                                            EXISTING ROAD
                                       LAGOON SURFACE
                                            	GRADE AND LEVEL WITHIN PERIMETER

                                            	CHAIN LINK FENCE
                                                                                      I' DEEP TRENCH FOR PIPING
                                                                                      (TO BE BACKFILLED)
                                                          Figure 3
                                                      Pilot Plant Layout
  The fourth controller controls operation of pumps, valves and
sensors associated with Tanks 2, 3 and 4 and sensors on the sur-
face of the Treatment Bed. Tank  1 is operated manually, only.
The primary function of the fourth controller  is to regulate ap-
plication of influent to the Treatment Bed. A master 24 hr/7 day
timer regulates  operation cycles  for application. When set on
automatic, the controller will activate whichever of the Tank 2, 3
and 4 valve and pump combinations are set on  automatic during
the preset cycle. These combinations include:
                                  • The actuation valve and pump on Tank 2 (extract feed to Tank
                                    3 or Tank 1)
                                  • The actuation valve and pump on Tank 4 (recycle to the third
                                    chamber of Tank 3, Tank 1 or discharge)
                                  • The valve controlling potable  water for dilution in the third
                                    chamber of Tank 3
                                  • The pump for  addition of nutrients to the third chamber of
                                  • The valve controlling recarbonation in the first chamber of
                                  • The valve controlling floe removal from Tank 3
                           LINK /=jr/vcr
•v i
v » i . .e






ti-L£ D

\. 	 Sfi^S 	 ^
r«e ivcr* fo «
                                                         Figure 4
                                                  South View of Pilot Plant
                                                              AMD EXTAAcr/0/J
                                                                                               ON-SITE TREATMENT    105

   The system will be shut off when set on automatic when:

 • A  sensor  on  the  surface of  the Treatment  Bed  indicates
 • A sensor in Tank 2 or 4 indicates the tank is empty, if the valve/
   pump combination for that tank is set on automatic
 • An application cycle is completed

 In addition,  each pump and  valve combination can be activated
 or deactivated  manually. The  system  includes  accumulating
 timers to monitor operating time  for each pump in the system. A
 main power  disconnect is provided for this controller, also.

   Construction of the pilot-plant began in May 1985.  Brush was
 cleared from the surface of the lagoon in the area proposed for
 placement of the Extraction and Treatment Beds. Approximately
 1-2 ft  of  cover were scraped from the lagoon  surface. The
 resulting surface was leveled.  In addition, an area on top of shale
 fill on the eastern portion of the lagoon was leveled for placement
 of the Process Area.
   Wells 1 and 2 were installed on May 24 and 25, 1985. Three bor-
 ings were conducted before an appropriate location for placement
 of the Extraction Bed was found. It was considered important to
 locate the Extraction Bed in a section of the lagoon that was filled
 continuously with sludges rather than in an area containing semi-
 continuous lenses of shale  fill.  Sheet piling was installed around
 Wells 1  and  2 between June  10 and  13, 1985.  Construction and
 water batching of the pilot-plant was completed on July 15, 1985.

   The pilot-plant has been inspected either daily or every second
 day during operation.  Field crews consist of at least two trained
 workers. Routine maintenance and inspection of the pilot-plant

 • Recording of elapsed time on the control timers for each  pump
 • Recording of the liquid levels in Tanks 2 and 4
 • Visual inspection of the Extraction and Treatment Beds
 • Measurement of the water  level elevation in Wells 1, 2 and  3
 • Mixing NaOH solution in  Tank 1 and feeding  the Extraction
   Bed, if required
 • Adjusting  feed stream  controls and timers for Tank 3
 • Sampling Wells 1, 2 and 3, and Tanks 2, 3 and 4 at times re-
^ 3
 9 -
 8 -
 .7 -
 6 -
 5 -
 4 -
 .3 -
 .2 -
 . I
o.a -
0.7 -
0.6 -
o.s -
0.4 -
0.3 -
0.2 -
0 I -
                          10        GO

                              TIME (DAYS)
                                                             Routine maintenance and inspection usually requires 1 to 2 hours
                                                             of site work per day.

                                                             Extraction Bed Operation
                                                               Recovery of leachate from Wells 1 and 2 began on July 17,
                                                             1985. Application of potable water to the surface of the Extrac-
                                                             tion Bed began on July 23. The purpose of the potable water addi-
                                                             tion was to test the application system from Tank I and the
                                                             hydraulic response of the Extraction Bed. It was indicated that
                                                             the apparent  hydraulic response  of the  bed was  greater  than
                                                             several days.
                                                               On  July  31, influent to  the Extraction Bed was changed to
                                                             0.005  N NaOH.  Throughout August,   low  infiltration  and
                                                             recovery rates  of extractant  were observed (Figs. 5 and 6). Pond-
                                                             ing frequently  occurred at the sludge surface. In response to this
                                                             problem, six well-points were installed in the Extraction Bed to ef-
                                                             fect subsurface injection of extractant. Installation of the well-
                                                             points was  completed  on  August  30 (Day 40). Subsequently,
                                                             sodium hydroxide solution  was injected through  the well-points
                                                             and applied to the surface of the Extraction Bed. Substantial in-
                                                             creases in extract  recovery  were observed (Fig. 6).  On Sept. 19
                                                             (Day 60), influent  to the bed  was increased in strength  to 0.1
                                                             N NaOH. This change  was  made in an attempt to increase TOC
                                                             concentration and quantity  of extract recovered. Recent data in-
                                                             dicate the permeability  of the Extraction Bed is increasing.
                                                                90 -

                                                                BO -

                                                                70 -

                                                                CO -

                                                                50 -

                                                                tO -

                                                                30 -

                                                                20 -

                                                                10 -
                                                                                     40        60
                                                                                         r/.ur (DATS)
                           Figure 5
                 Extraction Bed Influent Volume
                                                                                           Figure 6
                                                                                 Extraction Bed Effluent Volume
Treatment Bed Operation
  Addition of potable water to the Treatment Bed surface began
on July 15. This application continued until field status of the soil
system was established. On July 19, influent to the Treatment Bed
was  changed to a nutrient solution  containing dextrose as the
primary carbon source (Table 3). Inoculation with an activated
sludge culture occurred on July 24.

                           Table 3
       Makeup of Nutrient Solution Applied Beginning July 19

    50   g/1   Dextrose
     8   g/1   NH4C1
    45   g/1   Ca(NO3)2 • 4H2O
     0.3  g/1   K2HP04
This solution is diluted approximately 100:1 in Tank 3

  10. S -

    to -

   9.5 -

    9 -

   8.5 -

    8 -

   7.5 -
                             TIME (DAYS)
                   D   WELL I       +  ITJiY.i 2

                          Figure 7
               Extraction Bed  Well 1 and 2 (pH)
  Operation with nutrient solution as influent continued until
adequate  quantities of  extract  for  treatment  had  been  ac-
cumulated in Tank  2. Treatment of extract began on Sept.  13.
The target TOC concentration for influent to the Treatment Bed
(effluent from  Tank 3) was  1,000 mg/1. The nutrient solution
composition during this  phase of operation included 45  g/1
Ca(NO3)2 and 0.3 g/1 K2HPO4. This solution was diluted by a
factor of approximately 100:1 in Tank 3. The target influent rate
to the Treatment Bed  during this phase  of operation was 1
in./day. The typical application cycle was 4 hr/day.
  On Oct. 17,  the  flows to Tank 3 were adjusted to the target
Treatment Bed influent TOC of 2,000 mg/1. The target influent
rate  of application  was reduced  to 0.5  in./day. These changes
were made so that operation at a similar mass loading with a dif-
ferent hydraulic flux and influent concentration could be exam-
ined. Target flow and  concentration were  difficult to  control
because of variability in water pressure.

Extraction Bed
  Influent and effluent volumes for  the  Extraction  Bed  are
presented in Figures 5 and 6. Initially, influent volumes exceeded
the  volume  recovered  through Wells 1 and 2 by considerable


14 -
13 -
12 -
10 -
a -
8 -
7 -
6 -
5 -
4 -
3 -
9 _


% + + + * ++ +
D +
+ •*• +
D + ++ ++
DmD ++ + 0°° Q+ &
DaDfiQD P a Da++
— . -i 	 1 	 r- 	 1 	 1 	 1 	 1 	 1 	 1 	 1 	
amounts. These differences are attributed to several factors. In-
itially, the Extraction Bed was unsaturated and a large volume of
liquid was required to saturate the sludges. Leading to the dif-
ference between the two flows were losses through the joints of
the sheet piling.  Although the area for leakage was considered
small, a significant hydraulic gradient may have existed across the
joints. Losses through the Extraction Bed bottom also accounted
for more liquid loss.
  Recently,  increases  in the volume of extract recovered from
Wells 1 and 2 have been observed. These increases correlate close-
ly to the installation of the well-points for subsurface injection of
extractant. Additional increases have been observed, most likely
the result of increased sludge permeability occurring from the ad-
dition of more concentrated sodium hydroxide solution. Current
influent application rates are approximately 0.5 in./day;  current
effluent recovery  rates are approximately 0.3 in./day. Both in-
fluent application rates and effluent recovery rates are increasing.
During the interval between Days 60 and 100, extract recovery
was approximately 60%.
  Effluent pH,  TOC and  TDS for  the Extraction Bed  are
presented in Figures 7, 8 and 9. Initially, all three quantities were
high. The TOC and TDS of extract recovered from Well 2 were
approximately 11,000 mg/1 and 28,000 mg/1, respectively; the pH
was 10. However,  all three parameters decreased between Days 40
and 60. On Day 60, the TOC and TDS of extract recovered from
Well 2 were 5,200  and  11,000, respectively; the pH was 9.2. Initial
high pH and TDS values are attributed to lime previously applied
to  the  sludges  in the Extraction  Bed.  Decreases  in these
parameters between Days 40 and 60 are considered to be the result
of the lime dissolution and removal from the sludges.








p E


15 -
14 -

13 -
12 -
II -
9 -
B -
7 -

e -


B.S -
g .

7.5 -
7 -

S.S -

s -
5.5 -

z -
.9 -
.8 -
.7 -

6 -
S -
.4 -
.3 -
2 -

0.9 -
0,8 -
0.7 -
0.5 -
0.4 -
0.3 -
0.2 -
O.I -
0 -


*** '*
«m +
D* 3 +
0 * *
_ +

D ^ + ft * *
D ri
cP* o of
	 1 	 1 	 1 	 1 	 1 	 1 	 1 	 1 	 T— T —I 1 1 '
8 8.4 8.8 3.2 9.6 10 10.4 108
D ITELL i » "ELL 2
Figure 10
Extraction Bed - Well 1 and 1

0 D D^ D

^^ i — " — i 1 1 1 1 1 1 ' 20 40 SO tO 100 TlUf fDArsi Figure 12 Nutrient Tank Tank 3 (to Treatment Bed) Figure 15 Treatment Bed - Well 3 (Effluent TOC) attributed to development of the microbial population and modification of the target application rate to 0.5 in./day. Signifi- cant rainfall occurred during this period, also. Effluent TOC for the Treatment Bed is presented in Figure 15. The peak observed between Days 70 and 90 is the response ex- pected during development and acclimation of the microbial 108 ON-SITE TREATMENT



7.8 -
7.B -
7.4 -
7.2 -
7 -
C.8 -
6.B -
B.4 -
6.2 -
6 -

s.a -
S.B -
5.4 -
S.2 -
S -


00° D ° °
one ° D D
D QDnD n

                                   so        ao        100
                             TlilE (DAYS)
                          Figure 16
              Treatment Bed   Well 3 (Effluent pH)

population. This peak was accompanied by a decrease and subse-
quent increase in effluent pH (Fig. 16). This variation in pH also
was an expected response of the system.  Both phenomena were
observed in previous laboratory column experiments.
  Throughout operation, process effluent pH was between 6 and
8 and effluent TOC and TDS were  less  than 50 mg/1 and 900

  A pilot-plant  to demonstrate  a state-of-the-art process  for
cleanup of an industrial sludge lagoon has been designed, con-
structed and operated successfully for a period of over 100 days.
The cleanup process consists of two coupled treatment steps. The
first step is removal of contaminants from the sludges through in
situ extraction with aqueous sodium hydroxide. Extractant is in-
jected into the sludges through well-points or applied to the sur-
face through a perforated pipe distribution  network.  Extract is
recovered by means of two wells screened near the bottom of the
sludge deposits.  Operation of this system indicates that up to a
15-fold increase in removal efficiency, over natural processes, can
be obtained through increased  extract concentration. An addi-
tional 4-fold increase in the rate of renovation is obtained through
increased hydraulic flux through the sludges.
  Extract recovered from the sludges is  treated for removal  of
TOC through a soil-based sequential aerobic/anaerobic microbial
treatment system. Treatment occurs on-site,  immediately adja-
cent to the extraction process. A diverse, mixed microbial popula-
tion is developed in the soil system. Neutralized extract is applied
to the surface of the Treatment Bed and allowed to percolate
through the soil system. Aerobic and anaerobic microbial pro-
cesses metabolize organic contaminants to carbon dioxide and
methane.  Treatment efficiencies in excess of 95%  have been

  The authors would like  to  acknowledge  several people  at
Rutgers who have performed principal roles in the design, con-
struction and daily operation of the pilot-plant. These people are
Nick Bosko, Betsey Brush, Jim  Claffey, Irene Legiec,  John
Magee,  David Robertson, Judy Svarczkopf and  Valerie Ver-
miglio. In addition, the authors would like to acknowledge Joe
Evankow and Terry Burke for design and fabrication of the con-
trol panel.

1. Kosson, D.S., Ahlert, R.C., Boyer, J.D., Dienemann E.A. and Ma-
  gee II, J.F., "Development and Application of On-Site Treatment
  Technologies for Sludge Filled Lagoons,"  Proc. International Con-
  ference on New Frontiers for Hazardous Waste Management, EPA/
  600/9-85/025, Sept. 1985, 118-127.
2. Kosson, D.S. and Ahlert, R.C.,  "In-situ and On-site Biodegradation
  of Industrial Landfill Leachate," Environ. Prog. 3, 1984, 176-183.
3. Dienemann, E.A., Magee II, J.F.,  Kosson, D.S. and Ahlert, R.C.,
  "Rapid Renovation of a Sludge Lagoon," paper  presented at the
  AIChE Annual Meeting, Chicago, IL, Nov. 1985.
                                ON-SITE TREATMENT    109

                Cleanup  of  Contaminated Soils and Groundwater
                                   Using Biological Techniques
                                                 Paul E. Flathman
                                             Jason A. Caplan, Ph.D.
                                                O.H.  Materials Co.
                                              Biotechnology Division
                                                    Findlay, Ohio
  On-site  biological cleanup following spills  of biodegradable
hazardous organic compounds in lagoon, soil and groundwater
environments is a cost-effective technique when proper engineer-
ing controls are applied. Biodegrdation of hazardous organic con-
taminants by microorganisms minimizes liability by converting
toxic reactants into harmless end products.
  The two case histories presented in this paper describe:
• Bench-scale evaluation of the potential for biological cleanup
  in the spill  site matrix
• Field implementation of biological cleanup systems
  O.H. Materials Co. has performed biological cleanups of spilled
substances since 1978 when a railroad incident resulted in spillage
of acrylonitrile. Subsequent biological environmental restoration
projects have included additional acrylonitrile spills and cleanup
of other materials such as methylene chloride, methylethylketone,
crude oil, petroleum hydrocarbons, butylcellosolve, ethylacry-
late, n-butylacrylate,  toluene,  styrene, acetone,  isopropanol,
tetrahydrofuran and various phenolics.
  Cost-effectiveness, minimal disturbance to existing operations,
on-site destruction of spilled contaminants and permanence of the
solution are several of the advantages identified for implementing
biodegradation as a  technique for spill cleanup and environmen-
tal restoration.

  Many current and emerging technologies are available for the
on-site removal of groundwater contaminants. For cost-effective
cleanup following spills of hazardous organic materials, combina-
tions of treatment alternatives often are employed.
  The case history presented here describes the physical and bio-
logical  treatment of methylene chloride-contaminated ground-
water following  rupture of an underground  pipeline. After 2
months of field operation, air stripping techniques provided an
estimated  97% reduction  in  the  concentration of methylene
chloride in the groundwater. In a  downgradient monitoring well
located within 20 ft of the pipeline break, the methylene chloride
concentration was reduced from 9,300 mg/l to 300 mg/  by the
end of the second month.
  Since it became increasingly difficult  to remove the residual
methylene chloride from the groundwater by physical treatment
methods, biological  techniques were initiated.  Biological treat-
ment, using adapted microbial strains, was implemented by the
end of the third month and achieved an estimated 97% reduction of
the residual methylene chloride in groundwater. After 4 months of
field operation, greater than a 99.9% reduction  in the concentra-
tion of methylene chloride was attained.
  On Aug. 16, 1983, it was discovered that a buried  methylene
chloride line, located in proximity to a water main, had ruptured
and an undetermined amount of methylene chloride had leaked
into the soil and groundwater. Following discovery of the methy-
lene chloride leak, OHM was summoned to conduct an emergen-
cy response site investigation. Free product was collected first and
staged in vessels for eventual on-site treatment. Monitoring wells
that were installed established that the contamination was within
a confining clay  layer at the 20-ft level and did not reach the
aquifer located at 100 ft. Monitoring wells were utilized to iden-
tify the contaminant plume, and pumping techniques  were em-
ployed to contain the spilled methylene chloride.
  Emergency response to the spill involved only containment. An
investigative phase was initiated immediately to examine various
alternatives for environmentally restoring the site once the spill
was contained. Several alternatives were considered,  including:
excavation and  disposal; physical containment; and  biological
techniques. After reviewing ihe alternatives, it was determined
that  a  combination  of the alternatives  would  be  the  most
beneficial for site remediation.
  Following removal of 160 yd-' of highly contaminated soil to a
Class A secure landfill, positive placement and suction lift tech-
niques were  implemented.  Air stripping was  used to remove
methylene chloride from the recovered groundwater. When  it
became increasingly difficult to remove methylene chloride from
the groundwater,  biological techniques utilizing adapted micro-
organisms were employed.

Physical Treatment
  Air stripping was selected as the preferred treatment technology
for  removing methylene chloride from the recovered ground-
water. This decision was based on several important factors. The
first  factor was OHM's documentation of the strippability of this
compound in past emergency and remedial cleanup operations.
Second, the regulatory agencies did not require the implementa-
tion  of  vapor phase  scrubbing units to cleanse the air  emissions
from aeration devices. Finally, unlike granular activated carbon,
the   use of  air  stripping  is generally  maintenance-free, and
regeneration and/or disposal of  contaminated media is not

Bench-Scale Evaluation
  Characterization and bench-scale testing of recovered ground-
water was performed prior to final selection of the packed column
air  stripping  treatment alternative.  Among  the wastestream
parameters characterized were:

• Influent concentration of volatile organic(s)
• Presence of phased product(s)
• Hardness and alkalinity



                                                                      t\  HEAT
                                                                                            AIR STRIPPER
                                 TO PLANT
                               -+• WATER
                                                           Figure 1
                         Physical Treatment System Schematic for Methylene Chloride-Contaminated Groundwater
• Suspended solids concentration
• Scale-forming tendency at influent pH
Characterizing these parameters determined if pretreatment was
necessary prior to implementation of the primary treatment pro-
cess.  A  particular primary  treatment  process  may  be  cost-
effective, but the pretreatment costs may render the process cost-
  Methylene chloride was obtained when the groundwater was in-
itially pumped from several monitoring wells.  The recovered
groundwater often  had a murky brown color with  visible par-
ticulate  matter that slowly settled  with the heavier-than-water
methylene chloride.
  Following  the  separation  of methylene  chloride (by phase
separation) from recovered groundwater in quart jar samples, the
supernatant was analyzed for methylene chloride concentration.
The methylene chloride concentration approached the accepted
limit of solubility at ambient temperature.
  Dilling et al.1  calculated the volatilization rate of methylene
chloride  for concentrations ranging  from  several hundred  to
several  thousand mg/1 from  the  liquid-air interface in  open
laboratory vessels. OHM assumed that similar volatilization rates
of methylene chloride would  occur in process settling chambers
following gravity separation. OHM  also assumed that the action
of system pumps and surface winds blowing over water surfaces
into open vessels would result in even greater rates of volatiliza-
tion. Evaluations of the remedial alternatives were based upon
those assumptions.
Field Implementation
  Figure 1 is a schematic of the treatment system designed to treat
groundwater at a rate of 10 to 15 gal/min. The recovered ground-
water was pumped through a mixed-media filter bed to remove
sand and other particulate matter. Three types of porous media
were used in this filter: (1) anthracite, (2) silica sand and (3)  pea
gravel, with the  sand placed above the gravel and the anthracite
above the sand.  This media configuration provided a large pore
size near the filter surface to capture particulates with a  relatively
low head loss. Smaller sized particles were trapped in the lower
sand layer.
  Groundwater leaving the mixed-media  filter bed was piped to
an equalization vessel having a residence time of 14 hr. This vessel
was basically a rectangular holding tank which allowed the denser
methylene chloride to separate from the water by gravity. The
vessel was equipped  with sludge collection equipment to remove
accumulated solids and float controls in the last chamber to main-
tain a constant flow  to the next treatment step.
  Liquid from the equalization vessel was then pumped through a
skid-mounted  shell  and  tube heat  exchanger  to  raise  the
temperature of the  water from 10°C to more  than 40 °C. The
heated water was then pumped to  the top of the air stripping  col-
umn for removal of  soluble methylene chloride.
  Data collected during the project demonstrated that methylene
chloride,  as  pure product,  settled  in the equalization  vessel
together with  fines  and other  particulate matter that  were  not
trapped by the mixed-media filter. While the total volume of the
                                                           REMOVAL EFFIENCY

                                                           —O CALCULATED
                                                             •  OBSERVED
                                                         40        45        50

                                                        COLUMN OPERATING TEMPERATURE

                                                           Figure 2
        Removal Efficiencies of Methylene Chloride from Groundwater by Air Stripping at a Function of Column Operating Temperature
                                                                                                   ON-SITE TREATMENT    111

pure product collected in this manner did not exceed more than a
few gallons,  the vessel operated efficiently. Methylene chloride
concentrations in the water entering the air stripping column rare-
ly exceeded 150 mg/1. The mixed-media filter removed virtually
all paniculate matter  in the water, and backwashing of the filter
was infrequently required. The tower operating temperature was
maintained between 27 and 60 °C.
   Figure 2 is a summary of actual removal efficiencies obtained
during the 3  month operation of the air stripper. The same plot
contains predicted  data. The predicted  removal efficiency was
calculated  from the mass transfer model of Onda et a/.2-M At op-
erating temperatures less than 38 °C, removal efficiencies agreed
with those predicted.
   However,  significant deviations from expected values  were
observed at the higher operating temperatures. One explanation
for those deviations is the non-adiabatic nature of the air strip-
ping system. The column was constructed of mild steel and was
not insulated since  the system was meeting the discharge criteria
of 20 mg/1 methylene chloride. As ambient temperatures became
progressively  colder,  column heat losses  were greater. Those
observations  suggested  that  substantial deviations  from the
predicted removal efficiency would occur as the average operating
temperature of the air stripper was increased.
   The method used for predicting mass transfer rates of methy-
lene chloride during the design phase of the project was adequate
for estimating removal efficiencies  at elevated temperatures. This
same procedure has been implemented  by OHM in numerous air
stripping operations where mobilization of available equipment
was used for emergency situations.  Unlike other volatile organics,
equilibrium and mass transfer data for methylene chloride have
not  been  documented for  aeration  applications at elevated
   Results for the removal of methylene chloride from monitoring
well B-5 are  presented in Figure 3. This well was located  20 ft
downgradient from the ruptured methylene chloride line. An ex-
ponential decay curve was  used to quantify the removal rate
(decline in concentration) of methylene chloride from the ground-
water surrounding the well. Holding other variables constant, the
rate of decrease was assumed to be  a  function  of methylene
chloride concentration, i.e.
    —   =  -kC                                       (!)


   C  = the concentration of methylene chloride remaining
   t = time (days)
   k = the rate constant (day-')

   The curve generated  was  fitted to  the  following first-order

    C = C0e-i«                                         (2)

   C0 = methylene  chloride concentration at time zero (mg/1)
   The  first-order rate constant, k,  was determined  by  linear
regression using least  squares, and the first order equation was
converted to:

   InC = InC0 - kt                                      (3)

   A  50%  reduction  in methylene chloride  concentration was
observed within the first 11  days of physical  treatment. After  2
months of  field operation, air stripping techniques provided an
estimated  97%  reduction  in the  concentration of methylene
chloride in the groundwater being extracted. Since it became in-
creasingly  more  difficult  to remove  the  residual  methylene
chloride from the groundwater,  biological techniques were con-
sidered for implementation.

Biological Treatment
Bench-Scale Evaluation
  Prior to initiation  of a  methylene chloride biodegradation
feasibility study, the following screening analyses were performed
on 17 selected water samples collected from the site: methylene
chloride,  ammonia-nitrogen, nitrate-nitrogen,  orthophosphate-
phosphorus, pH and aerobic heterotrophic bacterial plate counts.
  Results  of the screening analyses indicated that the site matrix
samples contained viable microbial populations; i.e.,  the  site
matrix did not appear toxic to microbial growth (geometric mean
= 2.0 x  104 CFU/ml, n =  17).  The analyses also indicated that
inorganic  nitrogen and phosphorus nutrient additions would be
necessary  for effective biological treatment of the contaminated
groundwater. Ammonia-nitrogen, nitrate-nitrogen  and ortho-
phosphate were not detected  in any of the groundwater samples
analyzed.  The average pH for all 17 wells was 7.7, an acceptable
value for microbial growth.'
  The biodegradation study was performed using the electrolytic
respirometer  (Exidyne Instrumentation Technologies, Inc., Col-
orado Springs,  CO). For this study, both the  fate of methylene
chloride and  chloride release, resulting from the biodegradation
of methylene chloride, were used as the test parameters. Chloride
concentration was determined using an Orion model 96-17B com-
bination chloride electrode.
                       PHYSICAL TREATMENT
                          DAYS 0-Si
                         UETHYLEMC CHLORIDE
                         r* m 0.7t
                         M C • » 11000 - 00«It
                                      INJECTION OP
                                     BIOLOGICAL MEDIUM
                                     INTO GROUND WATCH
                20      40
                                              too      ito
                           Figure 3
             Physical Removal of Methylene Chloride
                   from Monitoring Well B-5

  An injection pool sample collected in November 1983, was used
for the electrolytic respirometer study. Sludge collected from the
plant's industrial wastewater treatment basin was used to inocu-
late the respirometer vessels. According to reports, the sludge in
the wastewater treatment basin had been continuously exposed to
low levels of  methylene chloride  for years  and was, therefore,
thought to contain populations of microorganisms acclimated to
methylene chloride. The electrolytic respirometer vessels  were
prepared as  follows:

Poisons added were KCN and NaN3
 S  =  wastewater treatment basin sludge
IP  =  injection pool

  Methylene chloride was added to all respirometer vessels at a
concentration of 100  mg/1.  This  concentration approximated
average methylene chloride concentrations found in the ground-
water at the site. Vessels 1 and 2 were replicates that demonstrated
the feasibility of using the natural microbial flora. Any nonbio-
logical loss of methylene chloride from the test environment
would be quantified with the abiotic control (i.e., vessels 3 and 4).
Respirometer vessels were prepared using procedures described by
Young and Baumann34 with diammonium phosphate and sodium
dihydrogen phosphate  added to each culture vessel. Nitrification
was not suppressed.
                           Figure 4
            Biodegradation of Methylene Chloride with
             Chloride Release in a Microbial Reactor
  On a  periodic  basis, 40  ml aliquots were  removed from
respirometer  vessels  and analyzed for methylene chloride and
chloride  concentrations. Throughout the study, aliquots were
periodically removed from culture vessels and analyzed for pH
and for ammonia-nitrogen, nitrate-nitrogen and orthophosphate
concentrations. When necessary, pH and inorganic nitrogen and
orthophosphate concentrations were adjusted to ensure that the
chemical environment remained optimal for microbial growth.
  Methylene  chloride biodegradation and chloride release data
for respirometer vessel  1 are presented in Figure 4.  Results for
vessel 2  were similar.  Neither loss  of methylene chloride nor
chloride release was observed in the abiotic control vessels.
  The theoretical chloride release for mineralization of methylene
chloride was calculated from the following relationship:
                           + 2HC1
The theoretical chloride release is 0.835 mg chloride per mg of
methylene chloride following complete oxidation.
  Methylene chloride was  biodegraded  stoichiometrically  to
chloride in respirometer vessels 1 and 2 (Fig. 4). Using the follow-
ing relationship, it  was determined that 130 and  120%  of the
theoretical chloride release for methylene chloride biodegradation
was met in respirometer vessels 1 and 2.
  The results indicated that methylene chloride  was completely
oxidized to carbon dioxide, water and the chloride  ion. The lack
of  a  significant lag  period  indicated  the  presence of adapted,
naturally occurring microbes that could biodegrade methylene

Field Implementation
  Use of  the underground  recovery  and  treatment system for
aquifer restoration has been described previously.n'15' 25>27> 28 A
schematic of the biological treatment system is shown in Figure 5.
  A modified activated sludge system was the preferred method
for aboveground treatment. In addition to providing efficient
biological treatment, the activated sludge system provided a sup-
ply of adapted microorganisms for inoculation  of the ground-
water through the injection system. The injection system was used
to inoculate the groundwater with naturally occurring  microbes
capable of biodegrading methylene chloride and  to  provide  a
suitable chemical environment necessary  to support microbial
                          Figure 5
           Biological Treatment System Schematic for
         Oxidation of Methylene Chloride in Groundwater
                                                                                                  ON-SITE TREATMENT    113


                         POOL WHILE BEING REFILLED
                     INJECTION OF    CONTINUOUS
                    MTO GROUND WATER TREATMENT
                       HT1ATED      INITIATED
                           Figure 6
        Chloride Release in the Bioreactor Resulting from the
        Biodegradation of Methylene Chloride in Groundwater
  The recovery  system  was  used to withdraw contaminated
groundwater for  aboveground treatment. The supernatant from
the treatment system was reinjected into the subsurface environ-
ment,  creating  a  closed-loop   system.   Biodegradation   of
methylene chloride took place in the ground as well as above the
ground in the biological treatment system.
  On day 82 of field operation, the bioreactor was prepared for
batch  treatment  of methylene chloride-contaminated ground-
water to  provide a large quantity of adapted microorganisms for
injection. To maintain a methylene chloride concentration in the
bioreactor capable of supporting microbial growth, air containing
methylene chloride vapor was bubbled into the bioreactor.  Be-
tween days 82 and 87, an 80.4% increase in the chloride concen-
tration was observed as a result  of  chloride release from  the
biodegradation of methylene chloride. Between days 87 and 90,
the bioreactor  was drained  into the injection pool while being
refilled. Injection of methylene chloride adapted microorganisms
into the groundwater environment began on day 89 (Fig.  6).
  Batch biological treatment of methylene chloride-contaminated
groundwater again was performed in the bioreactor between days
90 and 92. A 110% increase in chloride concentration was  ob-
served as a result of  methylene  chloride  biodegradation. The
bioreactor again was drained into the injection pool while being
refilled. On day 95, continuous biological treatment was initiated.
Over the 10-day operational phase, bioreactor effluent chloride
concentration averaged 131 ±65% greater than influent concen-
tration. Therefore, the bioreactor contained  a population  of
microorganisms adapted to methylene chloride.
  Removal of  methylene chloride from  the groundwater  was
rapid. A 50% reduction in methylene chloride concentration was
observed in monitoring well B-5 (Fig. 7) within 8 days after the
commencement of biological treatment.  The rate of decrease was
assumed to be first-order;" i.e., holding other variables constant,
the rate of decrease was assumed to be  a  function of methylene
chloride concentration.
  With  the onset of winter, biological treatment was temporarily
suspended on day  123 (Jan. 2, 1984). At that time, a 91% reduc-
tion in methylene chloride concentration had been achieved since
the injection of adapted microorganisms into the groundwater.
  The theoretical chloride release for biodegradation of 186 mg/1
methylene chloride (i.e., from 192  to 6 mg/1 methylene chloride)
is  155 mg/1.  The  observed chloride release was 156 mg/1; i.e.,
chloride concentration in monitoring well B-5 increased from 175
mg/1 on day 80 to  331 mg/1 on day 123. Therefore, we concluded
that biological  treatment was  responsible  for the removal of
methylene chloride from the groundwater.

  At the Naval Air Engineering Center in Lakehurst, New Jersey,
biodegradation techniques were used successfully to treat ethylene
glycol-contaminated   groundwater  following the  loss  of  an
estimated 4,000 gal of cooling water from  a lined surface storage
lagoon.   The cooling  water  was  estimated  to contain 25%
(vol/vol) ethylene  glycol.
  The problem developed on Jan. 5, 1982, following a storage
lagoon  liner  break.  A  subsequent investigative program con-
firmed soil contamination around the  lagoon and identified a
180-ft long by  45-ft wide contaminant  plume extending to the
east. At the start of the project,  the average ethylene glycol con-
centration in the groundwater was  1,440 mg/1. Approximately 85
to 93%  of the ethylene glycol was removed from the groundwater
within the first 26 days of biological treatment. By the completion
of the project, ethylene glycol was  reduced to below the limits of
detection (<,50 mg/1) in all production wells at the site.
       I ,0
                 INJECTION OF >IO(.OaiCAl
                 UEOAJUINTO OROUHD
                 WATER INITIATED
                                 METHYLCNE CHLORIDE

                                 •* • 0.71
                                 In C B In 100000 - 0.0851
           •o    »o    100    no   no   iJo    MO   i»o

                           Figure 7
     Biodegradation of Methylene Chloride with Chloride Release
         Based on Samples Taken from Monitoring Well B-5

Biofeasibility Evaluation
  The biofeasibility study assessed the biodegradation potential
of ethylene glycol in the ground as well as the presence of a toxic
or inhibitory environment to microbial growth. In addition, tech-
niques to increase the biodegradation rate in the ground also were
evaluated. Such techniques commonly include the use of surfac-
tants, cosubstrates, primary substrates, vitamins, trace elements
and selected microbial strains.  The results of the  biofeasibility
evaluation form the basis for subsequent field implementation, as
discussed in this case history.
  Ethylene glycol can be used as a carbon and energy source for
aerobic migroorganisms.6'10.17^1 The aerobic metabolism  of ethy-
lene  glycol is  relatively  common,  and  the  pathways of its
metabolism are known.4,5,18,19,26,31-33
  Anaerobic metabolism of ethylene glycol has been reported by
Dwyer  and Tiedje.8  Using  a  sewage  sludge inoculum under
methanogenic  conditions,  ethylene glycol was  converted to
ethanol, acetate and methane. The ethanol produced was further
oxidized to acetate with methane as the final end product. It has
been shown that clostridium glycolicum fermentation of ethylene
glycol yields equimolar amounts of acetate and ethanol.16
  Prior to initiation of the feasibility study, screening analyses, as
previously described for the first case history, were performed on
representative soil and groundwater samples collected from the
site. Ethylene  glycol  concentration was determined by direct
aqueous injection into a gas chromatograph.
  Screening analyses indicated that the groundwater contained a
viable microbial population; i.e., the environment did not appear
toxic  to microbial growth. Aerobic  heterotrophic bacterial
population densities ranged from 102 to 106 colony forming units
(CFU)/ml.  Bacterial population densities of soil samples collected
within the spill area varied from 102 to 106 CFU/g oven-dry soil.
Screening also indicated that both pH adjustment and inorganic
nitrogen and phosphorus nutrient additions would be necessary
for biological treatment of ethylene glycol. The average pH of the
13 samples  was  4.5.
                                   •  TOTAL BOD
                                       n = 40
                                       r2 = 0.85B
                                       y = 1740 - 4480 •"'
                                     ETHYLENE OLVCOL
                                       n = 6
                                       r2= 0.813
                                       In y = In 44100 - 0.7»»t
  The biodegradation study was performed using the electrolytic
respirometer with the composite groundwater sample. All samples
were prepared in a liquid nutrient medium containing basal salts
and yeast extract.
(1000 mg/1)
*Poisons added were HgCl2, KCN, NaN3
**GW = Composite groundwater sample
tLGW = Laboratory grade water

  Results from vessel 1 would measure the rate of biodegradation
of ethylene glycol by the natural microbial flora. Comparison of
oxygen uptake rates in vessels 2 and 3  would demonstrate the
presence of significant quantities of substances toxic or inhibitory
to microbial growth in the composite groundwater sample. Any
nonbiological loss of ethylene glycol would be quantified with the
abiotic control (i.e., vessel 4).
  Ethylene glycol biodegradation data for respirometer vessel 1 is
presented in Figure 8. A loss of ethylene  glycol was not observed
in the abiotic control (vessel 4). The  theoretical oxygen demand
(ThOD)  for  aerobic mineralization of ethylene  glycol was
calculated from the  following relationship:

  2 (CH2OH)2 + 5  O2 	> 4 CO2  + 6 H2O
The ThOD is 1.29 mg O2 per mg of ethylene glycol. With an in-
itial concentration of 1440 mg/1 ethylene glycol,  1860 mg O2
would be required for complete oxidation. The ultimate BOD
(BODu)  was determined by fitting  oxygen uptake data to  a
modified crescent curve:30
  BOD = a + be-kt                                     (4)

BOD =  the amount of BOD expressed or exerted at time t
    a =  BODu, the ultimate amount of oxygen uptake to be
         expressed (mg/1)
    b =  a lag period parameter (mg/1).  (Note: b = a if the fitted
         crescent curve intersects the origin)
    k =  the rate constant (hr-i)
    t =  time (hr)
The ultimate BOD is defined as  the total amount of oxygen re-
quired to biodegrade the  immediately available organic matter
present in a sample.24
  Using the following relationship, it was determined that 94% of
the theoretical oxygen demand for ethylene glycol biodegradation
was met:

          x 100
                          Figure 8
         Biodegradation of Ethylene Glycol in Composite
                Groundwater Sample (Vessel 1)
Thus, the results indicate that ethylene glycol was completely ox-
idized to carbon dioxide and water without the accumulation of
incomplete oxidation products. With a BODu of 1700 mg/1, an
estimated minimum of 170 mg/1 NH3-N and 17 mg/1 PO,}-?
would be required in the groundwater to prevent nitrogen and
phosphorus  from  limiting  microbial  growth during biological
                                                                                                 ON-SITE TREATMENT    115




                                      -»-• VESSEL,
                               COMPOSITE OROUNO WATER SAMPLE
                                4- BASAL SALT! 4- OLUCOSE
                                  Ts In SB.4 + O.OSB7I
                             • BASAL SALTS » OLUCOSE
                                t MICROaiAL INOCULUM
                                 n   7
                                 rf = O.BBS
                                 In y = hi BS.1  4- 0.04781
                           Figure 9
    Oxygen Uptake Data for the Composite Groundwater Sample
           Containing Glucose and for a Glucose Control
   The lack of a significant lag period in both oxygen uptake and
 ethylene glycol biodegradation by the natural microbial flora in
 vessel 1  indicated the presence of adapted microbes that could
 biodegrade ethylene glycol. That Finding was significant because
 it indicated that in situ biodegradation of ethylene glycol already
 was occurring and that the management approach should be to in-
 crease the natural biodegradation rate.
   The  groundwater appeared  to  be slightly stimulatory  to
 microbial growth, as evidenced by the slightly greater rate of oxy-
 gen uptake in the composite groundwater sample compared to the
 glucose-basal salts control  (Fig.  9). At  the termination of the
 study, initial oxygen  uptake data in the toxicity/inhibition control
 (respirometer vessels 2 and 3) were compared on a semi-log plot.
 A comparison was  made  of the  slopes of the lines-of-best-fit
 through the linear portion of the natural log transformed oxygen
 uptake data. In the linear portion of the semi-log plot, the rate of
 oxygen uptake in the glucose-containing composite groundwater
 sample was 1.2 times greater than in the  glucose-basal salts con-
 trol. Thus, evidence for an environment toxic or inhibitory to
 microbial growth was not found.
   The feasibility study results indicated that biodegradation tech-
 niques were a viable  option for reducing  ethylene glycol concen-
 trations in the groundwater.

 Field Implementation
   The biological treatment  program at the site was divided into a
 14-day operational phase,  a 3-month monitoring phase and a
 9-month  maintenance program.  The operational  phase was de-
 signed primarily to provide maximum recovery, treatment and en-
 hanced bacterial growth in  the groundwater both within the spill
 area and within the contaminant plume (Fig. 10). The monitoring
 program  was designed to assess  the ethylene glycol degradation
 rate  in  the groundwater following  nitrogen and phosphorous
 nutrient addition, pH adjustment and enhanced  bacterial growth.
The maintenance program  was designed  to  provide an environ-
ment suitable for the continued  biodegradation of any residual
ethylene glycol remaining in the groundwater environment.

V" ,

• 0
                                                                                                                     OnOUNO WATCH
                                                                       O RECOVERY WELL

                                                                       O (OIL SORINO

                                                                       • BOIL IORINO / MONITOR WELL
                           Figure 10
            Schematic of Ethylene Glycol Storage Lagoon
               and Contaminant Groundwater Plume
   Ethylene glycol contamination at the site was divided into two
 zones. The first zone was the unsaturated zone between the sur-
 face and the water table where ethylene glycol had been retained
 through  capillary  action.  The  highest  contamination  level
 detected in that zone was 4,900 mg/1 ethylene glycol. Surface con-
 tamination also was indicated  following analysis  of shallow
 samples (0 to 2 ft) collected adjacent to the lagoon.
   The second  zone of  contamination  was  the  groundwater.
 Groundwater samples within the spill area had ethylene glycol
 concentrations as high as  2,100 mg/1. Monitoring wells were in-
 stalled, and  subsequent  water analyses  indicated significant
 groundwater contamination. A downgradient contaminant plume
 was estimated to be 180 ft long by 45 ft wide.
   A two-phased approach was implemented to deal with both soil
 and groundwater contamination. Using  injection  and recovery
 wells,  initial  efforts addressed  the highly contaminated  soils
 underlying the storage lagoon. The second phase concentrated on
 groundwater cleanup. Its goals were (1) to prevent further migra-
 tion  of contaminated groundwater  and (2)  to  treat any  con-
 tamination released during treatment of the unsaturated zone.
   The injection system was used to adjust groundwater pH as
 well as provide the inorganic nitrogen and phosphorus necessary
 to support microbial growth. The recovery system withdrew con-
 taminated groundwater  for  above ground treatment in an ac-
 tivated sludge treatment system (Fig.  11). Supernatant  from the
 treatment system then was reinjected into  the subsurface environ-
 ment, creating a closed-loop system.
                            pH ADJUSTMENT
                     PREMIX I NUTRIENT ADDITION
                    1 TANK '
                                                   WASTE SLUDGE
                                                    FOR SURFACE
                           Figure 11
      Flow Diagram for the Activated Sludge Treatment System
           of Ethylene Glycol-Contaminatcd Groundwater

                                                                   ETHYLENE GLYCOL
                                                                    O BELOW DETECTION LIMIT

                                                                   MICROBIAL POPULATION

                                                           Figure 12
     Ethylene Glycol Concentration, Bacterial Population Density, and pH as a Function of Time for a Contaminant Plume Production Well
  Based  on the information obtained during the site investiga-
tion, five recovery wells were installed to recover contaminated
groundwater (Fig. 10). Three wells located near the lagoon pro-
vided zones of attraction for treated water injected in that area.
The  remaining two recovery wells  were positioned  east of the
lagoon to  recover contaminated water  from  the contaminant
plume and aid in the distribution of treated water from the plume
injection system.
  Following pH adjustment and nutrient addition, supernatant
from the biological treatment system was recharged through a
three-phase  injection  system.  The lagoon  injection  system  was
employed to flush contaminated soil and thereby transmit con-
taminated water to the three recovery wells located in the vicinity
of the lagoon. The plume  injection system  was similar to the
lagoon system in  both construction and operation. The primary
functions of the plume injection system were to enhance bacterial
growth through pH adjustment and nutrient addition  and to
create a  gradient from the fringe of the plume toward the  two
recovery  wells located at the center  of the contaminant plume.
The third injection phase was implemented through surface ap-
plication. Surface application was used primarily in the lagoon
area to flush the unsaturated  zone and  enhance bacterial growth
in the contaminated soil.
  The activated sludge biological treatment system was designed
to reduce ethylene glycol concentration through microbial bioxi-
dation using the indigenous microbial flora. A second role of the
bioxidation reactor was to  provide adapted microorganisms for
the three-phase injection system (Fig. 11). The biofeasibility study
had demonstrated that the  natural microbial flora were adapted
to biodegrading  ethylene glycol and that the management ap-
proach  for  the project  should  be  to  increase the natural
biodegradation rate using  microbes indigenous to the ground-
  Throughout the  operational  phase, treatment system and
recovery well samples  were analyzed for  aerobic heterotrophic
bacteria, pH, dissolved oxygen, inorganic nitrogen, phosphorus
and ethylene glycol. Based on the results of those analyses, en-
vironmental parameters were modified to maintain an effective
                                                                        ETHYLENE QLYCOL
                                                                          O BELOW DETECTION LIMIT
                                                                        MICROBIAL POPULATION
                                                           Figure 13
    Ethylene Glycol Concentration, Bacterial Population Density, and pH as a Function of Time for a Downgradient Spill Area Production Well
                                                                                                   ON-SITE TREATMENT    117

 rate of biological treatment. During the monitoring phase, those
 same  parameters were quantified  on  a monthly  basis from
 recovery well samples.
   In the initial phases of treatment, ethylene glycol concentra-
 tions in two plume recovery wells were reduced from 690 and 420
 mg/1 to below limits of detection (LOD  = 50 mg/1) within 26 days
 of treatment. With  the exception of one monitoring phase data
 point (day 40),  ethylene  glycol concentration remained below
 detection limits. As shown  in Figure 12, groundwater pH  was
 maintained within an optimal range for bacterial growth (i.e., pH
 6 to 7). Bacterial population density was returning to background
 levels (i.e., 105 CFU/ml) following a high of 10« CFU/ml.
   In the two downgradient recovery wells adjacent  to the storage
 lagoon, the groundwater concentration  of ethylene glycol  was
 reduced by more than 85% and 92%, respectively, within the first
 26 days of treatment (Fig. 13). Initial concentrations were 3,400
 and 1,100 mg/1, respectively. In the upgradient recovery well ad-
 jacent to  the lagoon, the ethylene  glycol concentration in the
 groundwater  increased from 100 mg/1  on day 0 to 880 mg/1 by
 day 13. However, monitoring data showed a continued decrease
 in  ethylene glycol concentration after day 13. The increase in
 ethylene  glycol  concentration  prior  to day  13  reflected  the
 project's  aggressive operational phase, whereby,  the  injection
 system flushed pockets of ethylene glycol from the unsaturated
 zone into the groundwater.
   The maintenance  program focused on the removal of those re-
 maining pockets of contamination,  particularly in  the  lagoon
 area. As part  of that program, both lime and diammonium phos-
 phate were applied to the soil surface in the lagoon  and contami-
 nant plume areas. Lime increased the pH of the soil and ground-
 water to a  favorable range for microbial growth. Diammonium
 phosphate,  a  readily  available  source  of  nitrogen  and
 phosphorus,  supported  the growth  of an increased microbial
   By the completion of the project, ethylene glycol was reduced
 to below the limits of detection in all production wells at the site.

   The two case histories presented demonstrate the application of
 biological techniques for environmental restoration of areas con-
 taminated  with spilled organic materials.
  The combination  of physical  and biological techniques effec-
 tively removed methylene chloride from the groundwater  at a
 relatively rapid rate. If it  had been necessary to remove the con-
 taminated  soil from  the site, transportation and disposal (T&D)
 costs for 14,000 yd3 of contaminated soil would have been at least
 $1,050,000. The estimate for traditional treatment was based on a
 transportation cost of $3/mile in 20-yd3 sized trucks to a hazar-
 dous waste disposal facility 100 miles from the treatment site. The
 disposal cost for the contaminated soil was $60/yd3.
  By successfully decontaminating the soil and allowing ii to  re-
 main on-site, a two-fold cost savings was achieved and all future
 liability was reduced substantially. This cost  estimate  for T&D
 was conservative since removal of soil from an area increases void
 volume for transportation and disposal often by a factor of 1.2 to
 1.3. Moreover, hazardous materials cannot be shipped off-site in
 bulk,  but  most be  packaged into  drums. Additional  chemical
 analyses generally are required to characterize the waste prior to
 off-site transportation and disposal. Inclusion of these additional
 factors that must be considered for off-site disposal  resulted in an
 even greater cost benefit for in situ cleanup.
  Biological techniques, as described in  the second case history,
were  also effectively used to remove ethylene glycol  from  the
groundwater at a relatively rapid rate. The flexibility of the injec-
tion/recovery system in maintaining an environment conducive to
 biodegradation while flushing ethylene glycol from the ground-
water was a key factor in the removal of scattered pockets of con-
tamination. If it had been necessary to remove the soil from both
the spill area and the contaminant plume to a hazardous waste
disposal facility 500 miles from the treatment site, the T&D cost
would have been more than 16 times the cost for on-site treat-
  Biodegradation as a method  for spill cleanup and environmen-
tal restoration is considered to be a  promising technology. Land
treatment techniques have been engineered and are accepted as an
economical and environmentally sound means of destruction for
many types of industrial wastes.  With regard to the cleanup of
contaminated soil and groundwater, excavation has been a com-
mon method for remediation. However, combinations of physical
and biological techniques now are gaining increasing acceptance as
a practical, cost-effective alternative for environmental restoration.
 I.  Adas, R.M. and Bariha, R.. Microbial Ecology: Fundamentals and
    Applications, Addison-Weslcy Publishing Company, Reading, MA,
 2.  Bolles, W.L. and Fair. J.R., "Improved Mass-Transfer Model En-
    hances Packed-Column Design," Chem. Eng.. 84(14), 1982, 109-116.
 3.  Bordner, R., Winter, J. and Scarpino, P.,  Eds. "Microbiological
    Methods for  Monitoring  the  Environment-Water  and Wastes,"
    EPA 600/8-78-017. Environmental Monitoring and Support Labora-
    tory. U.S. EPA, Cincinnati, OH, 1978.
 4.  Caskey, W.H. and Taber, W.A., "Oxidation of Ethylene Glycol by
    a Salt-Requiring Bacterium," Appl.  Environ. Microbiol.. 42, 1981,
 5.  Child,  J. and  Willetts, A.,  "Microbial Metabolism of Aliphatic
    Glycols.  Bacterial  Metabolism  of  Ethylene  Glycol."  Biochem.
    Biophs.  Ada. 538, 1978, 316-327.
 6.  Claus, D. and  Hempel, W.. "Specific Substrates for Isolation and
    Differentiation  of Azotobacler vinelandii," Arch. Mikrobiol, 73,
    1970, 90-96.
 7.  Dilling,  W.L.,  Tefertiller, N.B. and  Kallos.  G.J., "Evaporation
    Rates of Methylene Chloride. Chloroform,  1,1,1-Trichloroethane,
    Trichlorocthylene,  Tetrachloroethylene,  and  Other  Chlorinated
    Compounds in  Dilute Aqueous Solutions," Environ. Sci. Technol.,
    9, 1975, 833-837.
 8.  Dwyer,  D.F. and Tiedje. J.M., "Degradation of Ethylene Glycol
    and Polyethylene Glycols by  Methanogenic Consortia," Appl. En-
    viron. Microbiol., 46, 1983, 185-190.
 9.  Anon.,  "Current Developments,  Hazardous Waste," Environ.
    Reporter. May  6. 1983. 11-12.
10.  Fincher, E.L. and Payne, W.J., "Bacterial Utilization of Ether Gly-
    cols," Appl. Microbiol.. 10, 1962, 542-547.

11.  Flathman. P.E.,  McCloskey.  M.J., Vondrick. J.J. and Pimlett,
    D.W., "In Situ Physical Biological Treatment  of Methylene Chlor-
    ide (Dichloromethane) Contaminated Ground Water,"  Proc.  of the
    Fifth  National Symposium on Aquifer Restoration and Ground
    Water Monitoring, Columbus. OH,  May 1985, 571-597.
12.  Flathman, P.E. and Caplan, J.A., "Biological Cleanup of Chemical
    Spills,"  Proc.  of HAZMACOM 85.  Oakland, CA, Apr. 1985, 323-
13.  Flalhman, P.E. and Githens, G.D., "In Situ Biological Treatment
    of Isopropanol, Acetone, and Tetrahydrofuran in the Soil 'Ground-
    water Environment," E.K. Nyer,  Ed., Ground  Water Treatment
    Technology. Van Nostrand Rcinhold  Company,  New York. NY,
    1985, 173-185.
14.  Flalhman, P.E..  Quince,  J.R. and Bottomley,  L.S.. "Biological
    Treatment of Ethylene Glycol-Contaminated Groundwater at Naval
    Air Engineering Center, Lakehursl, New Jersey," Proc. of the Fourth
    National Symposium on Aquifer Restoration and Ground  Water
    Monitoring. Columbus, OH, May 1984. 111-119.

15-  Flathman, P.E.,  Studabaker,  W.C., Oithens,  G.D. and  Muller,
    B.W., "Biological Spill Cleanup," Proc. of the Technical Seminar
    on Chemical Spills, Toronto, Ontario, Oct. 1983, 117-130.
16.  Gaston, L.W. and Stadtman, E.R., "Fermentation of Ethylene Gly-
    col by Clostridium glycolicum sp. /. Bacterial., 85,  1963, 356-362.
17.  Gonzalez,  C.F., Taber, W.A. and Zeitoun, M.A., "Biodegrada-
    tion of Ethylene  Glycol by a Salt-Requiring Bacterium," Appl.
    Microbiol., 24, 1972, 911-919.
18.  Harada, T. and Nagashima, Y., "Utilization of Aklylether Com-
    pounds by Soil Bacteria," J. Ferment. Techno!., 53, 1975, 218-222.
19.  Jones, N.  and Watson, G.K.,  "Ethylene  Glycol and Polyethylene
    Glycol Catabolism by a Sewage Bacterium," Biochem. Soc. Trans.,
    4, 1976, 891-892.
20.  Kemmer,  F.N., Ed., Nalco Water Handbook, McGraw-Hill  Book
    Company, New York, NY,  1979.
21.  Klecka, G.M., "Fate and Effects of Methylene Chloride in  Acti-
    vated Sludge," Appl. Environ.  Microbiol., 44, 1982, 701-707.
22.  Kobayashi, H  and Rittman,  B.E., "Microbial Removal of Haz-
    ardous Organic Compounds," Environ.  Sci.  Tech.,  16, 1982, 170A-
23.  Larson, R.J., "Role of Biodegradation Kinetics in Predicting En-
    vironmental Fate," A.W. Maki, K.L. Dickson, J. Cairns, Jr.,  Eds.,
    Biotransformation and Fate of Chemicals in the Aquatic Environ-
    ment, American Society for Microbiology, Washington, DC,  1980
24.  Mitchell,  R.,  Introduction  to Environmental Microbiology,  Pren-
    tice-Hall,  Inc., Englewood Cliffs, NJ, 1974.
25.  Ohneck, R.J. and Gardner, G.L., "Restoration of an Aquifer Con-
    taminated  by  an Accidental Spill of Organic Chemicals," Ground
    Water Monitoring Review, 2(4), 1982, 50-53.
26.  Pearce, B.A.  and  Heydeman,  M.T., "Metabolism of Di(Ethylene
    Glycol) [2-(2'-Hydroxyethoxy) Ethanol] and Other Short Poly (Ethy-
    lene Glycol)s by Gram Negative Bacteria," J. Gen.  Microbiol, 118,
    1980, 21-27.
27.  Quince, J.R. and Gardner, G.L., "Recovery and Treatment of Con-
    taminated Ground Water: Part I," Ground  Water Monitoring Re-
    view, 2(3), 1982, 18-22.
28.  Quince, J.R. and Gardner, G.L., "Recovery and Treatment of Con-
    taminated Ground Water: Part II," Ground  Water Monitoring Re-
    view, 2(4), 1982, 18-25.
29.  Roberts, P.V., Hopkins, G.D., Munz, C. and Riojas, A.H., "Evalu-
    ating Two-Resistance Models for Air Stripping of Volatile Organic
    Contaminants  in  a Countercurrent, Packed Column," Environ.
    Sci. Tech., 19, 1985, 164-173.
30.  Shammas, N.C., "Modified Crescent Curve Fitting, Program Num-
    ber 341C," Users' Library, Hewlett-Packard Company, Corvallis,
    OR, 1983.
31.  Thelu, J.,  Medina, L. and Pelmont, J., "Oxidation of Polyoxy-
    ethylene Oligomers by an Inducible Enzyme from  Pseudomonas
    P400," FEMSLett., 8, 1980, 187-190.
32.  Wiegant, W.M. and DeBont, J.A.M., "A New Route for Ethylene
    Glycol Metabolism in Mycobacterium  E44," /. Gen. Microbiol.,
    120, 1980, 325-331.
33.  Willetts, A., "Bacterial Metabolism of Ethylene Glycol.  Biochem,"
    Biophys. Acta, 677, 1981, 194-199.
34.  Young, J.C. and Baumann, E.R., "The Electrolytic Respirometer—
    I Factors Affecting Oxygen Uptake Measurements," Water Res.,
    10, 1976, 1031-1040.
   The authors wish to acknowledge the contribution of their col-
league,  Lucy  S.  Bottomley,  P.E.,  Environmental  Engineer,
Public  Works Department,  at  Naval  Air Engineering  Center,
Lakehurst, New Jersey. At O.H. Materials Co., we specifically
thank Robert H. Panning and Robert J. Ohneck for their support
throughout these projects.
                                                                                                       ON-SITE TREATMENT    119

        Physical/Chemical  Removal  of Organic  Micropollutants
            from  RO Concentrated Contaminated  Groundwater
                                                  L. Simovic
                                            Environment Canada
                                     Environmental Protection Service
                                       Wastewater Technology Centre
                                             Burlington, Ontario
                                                  J.P. Jones
                                          University of Sherbrooke
                                             Sherbrooke, Quebec
                                               I.C. McClymont
                                              ICAN Consultants
                                             Burlington, Ontario
  The reverse osmosis (RO) concentrate of contaminated ground-
water from the Gloucester Landfill site near Ottawa, Ontario, was
treated by air stripping (AS), ozonation (O3) and ultraviolet (UV)
irradiation for the removal of organic micropollutants.
  Fourteen compounds, predominantly volatiles, were monitored
during this 6-week study. The concentration ranges of some of the
compounds in the reverse osmosis concentrate were:  trichloro-
ethylene  2-280 /ig/1;  benzene  28-1,350  pg/1; chloroform
800-40,100  pg/l;  diethyl ether  136-2,300  /ig/1  and acetone
750-36,000 jig/1.
  Based on the results of this study,  the air stripping process
alone is considered to be adequate for treating the RO-concen-
trated contaminated groundwater at the Gloucester Landfill site.
Residual concentrations of four of the six compounds with pro-
posed groundwater quality  objectives met  or exceeded those
  The removal efficiency achieved by all three  unit processes
combined, however, was better than that obtained by AS alone.
For the degree of additional improvement that might be achieved,
it is doubtful that the addition of Os/UV to the AS process would
be economically justified.

  From 1969 to 1980, federal government laboratories in Ottawa,
Ontario, disposed of their  hazardous  wastes (mainly organic
solvents) in a Special Waste Compound (SWC) at the nearby
Gloucester Municipal Landfill.1 In  1981 contaminants including
diethyl ether,  acetone  and some  chlorinated  solvents  were
detected in the groundwater migrating from the landfill site. This
contamination  was  a potential danger to the drinking water
source of nearby residents. Subsequently, Environment Canada
and Transport Canada undertook a number of studies to define
the extent  of the problem  and identify possible  remedial
  In 1984 Environment Canada's Wastewater Technology Centre
(WTC),  in cooperation with the  Environmental Emergencies
Technology Division, carried out a pilot scale study on the
removal of toxic organics from the contaminated groundwater
within the SWC. To define one remedial action, the objectives of
this study were: (1) to evaluate the effectiveness of selected reverse
osmosis (RO) membranes for removing toxic organics from the
contaminated groundwater and (2) to evaluate the effectiveness of
air  stripping,  ozonation  and ultraviolet  (UV) irradiation for
treating the concentrate from the RO system.
  The results related to the RO  tests  have been  presented
elsewhere.' This paper presents the results for the removal of
organics from the RO concentrate by air stripping, ozonation and
ultraviolet irradiation. For each treatment process, the results are
discussed with  reference to specific toxic organic  compounds
which were present in the groundwater at concentrations higher
than the proposed groundwater quality objectives.4

  A trailer facility outfitted for this project was set up at the
Gloucester site beside test well 36 W. This well, located within the
SWC, was found to have the  highest concentrations of organic
micropollutants.' A schematic  of the pilot plant used during this
study is shown in Figure 1.
  The air stripping unit in the pilot plant consisted of two 1.22 m
by 10 cm (I.D.) plexiglass columns connected in series. The col-
umns were packed with 1.3  cm  Intalox*  saddles.  A  Whispair
1707Jt blower was used to force air through the stripping columns.
  The ozonation experiments were conducted using a bubble col-
umn 91 cm by 10 cm (I.D.). Ozone was generated from oxygen us-
ing a Welsbach T-816J ozonator. The Oj concentrations in the inlet
gas  and  the off-gas were measured using a standard iodometric
method.' Ozone in the liquid was measured using the indigo
  The UV irradiation chamber  used was a Trojan System 2000.**
It consisted of four 9 W low pressure mercury lamps. The total UV
output of all the lamps in the reactor was determined by ferriox-
alate actinometry'and was found to be 4.1  x 10 ~4 einsteins/min.
The flow of liquid through the reactor was maintained during the
tests to give a theoretical retention time (RT) of 22 min.

Air Stripping
  During air  stripping experiments, the following three variables
were studied: air flow rate, the packing height and the water flow
rate. A full 23 factorial experimental design was used to allow for
the determination of the relative significance of the effect of these
variables (Table 1). The air and water flow rate levels were chosen
to bracket a  wide range below the conditions at which the col-
umns would become flooded.  From the calculated stripping fac-
tor  for the volatile compounds present in the concentrate, the
values for the  packing height were chosen to give a  sufficient
number of transfer units for ?  99% removal of these compounds.

• Norton Chemical Company, Akron, OH
t Roots Blower Operation, Dresser Industries, Inc., Cornerville, IN
t Polymetrics Inc., Sunnyvale, CA
"Trojan Technologies Inc., London, Ont.






' t


                                                           Figure 1
                                        Process Schematic for Groundwater Treatment System
                           Table 1
                 Air Stripping Process Variable
Air flow rate (1/min)
Packing height (m)
Water flow rate (1/min)
Ozonation/Ultraviolet Irradiation
  The effects of ozonation, UV irradiation,  air stripping and
water flow rates on the removal of organic micropollutants were
observed using a full 24 factorial experimental design (Table 2).

                           Table 2
                AS/Os/UV Experiment Variables
Air stripping
Os dose (mgO3/l of gas)
UV irradiation
Water flow rate (1/min)


4.1 x 10-4
•Air now rate = 300 1/min; packing height 1.22 m
Sampling and Methods of Analysis
  The locations where samples of pilot plant influent and treated
effluent were collected during the tests are shown in Figure 1. For
each experiment, two sets of samples were collected: one set was
collected at the beginning of the test and another set after 2 hr of
continuous operation of the unit process(es) being tested.  After
collection,  the samples  were stored at  4°C in glass vials with
teflon-lined septums. Sample analyses were completed within 10
days of collection of a given set of samples.
  The organic compounds, predominantly volatiles, were analyzed
by the purge and trap  method using a  Hewlett Packard model
5830A Gas Chromatograph coupled to a Flame lonization Detec-
tor (FID). The following pairs of compounds coeluted during the
analysis  and could not  be  quantified  individually:  (1) 1,1-di-
chloroethane (RT  = 13.6) and tetrahydrofuran (THF) (RT =
13.7;  (2) 1,1,1-trichloroethane (RT = 17.4) and 1,4-dioxane (RT
= 17.4).

Water Characteristics
  During this 6-week study the concentrations of various com-
pounds  in  the  RO concentrate varied  widely. These  data  are
presented in  Table 3, together with the proposed groundwater
quality objectives for these compounds.

  Forty-seven tests were conducted during this study.  Of these
tests,  27 used AS, 7 used O3 and 3 used UV alone to remove the
organic micropollutants  from the concentrate. In 11 tests, O3  was
                                                                                                  ON-S1TE TREATMENT     121

                           Table 3
 The Influent Concentration for Some of the Organic Mlcropollulanls
Organic Contaminant

1,1 -Dichloroethylene
Diethyl Ether

NA — Nol Available
RO Concen- Quality Objectives
Irate 0*g/l)  (/'g/l)
  9-  1,670
  2-  1,531
136-  2,300
 32-   653
  2-   280
 28-  1,350
 followed by UV irradiation. Air stripping was followed by UV in
 2 tests, by Oj without UV in 5 tests and by O3 in the presence of
 UV in 8 tests. Not all of the samples collected during these tests
 could  be analyzed for all the parameters, so some of the data
 points required to complete the factorial analysis of the data are
 missing.  A  stepwise  regression  using backwards  elimination
 routine was, therefore, applied for the analysis of the data.
   Although the concentration of organic compounds in the RO
 concentrate varied greatly (see Table 3), the removal efficiency of
 the various processes tested was not affected by this variability in
 the concentration. Variations in the removal efficiencies were due
 to the experimental conditions applied.

 Air Stripping
   Table 4 summarizes the results  for air stripping experiments.

                           Table 4
       Percentage Removal and Residual Concentration Ranges
          Achieved During AS Treatment of  Groundwaler

1,1 -Dichloroethylene
1 , 1 -Dichloroethane/THF
Diethyl Ether
1,1,1 -Trichloroethane/
1 ,4-Dioxane
Range of
57- > 99
32- 99
25- 99
31- >99
25- >99
24- 99


2.3- 38.2
0.5- 130.2
6.4- 667.9
1.6- 228.6

12.5- 813.7
0.5- 63.4
1.2- 78.2
  In addition to the chemicals listed in Table 4, the removal effi-
ciencies of  four other  compounds,  listed in  Table 5, were
estimated. The estimates were based on the percent reduction in
the peak  areas for  the compounds  observed during the  gas
chromatographic analysis of the RO concentrate and the air strip-
ping samples.
  Analysis of  the data from the  AS experiments showed that
water flow rate had  the greatest effect on the removal of most
compounds. The best removals  were obtained in  the experiments
where the water flow rate was low (0.5 1/min). Packing height and
air flow rate did not have a substantial effect on the efficiency of
removal within the range of values used in these experiments.
                        Under the best conditions (water flow rate of 0.5 1/min), four
                      of the compounds  were removed to below  their groundwater
                      quality objective concentrations (see Table 6).

                                                Table 5
                            Percent Reduction Range Estimated from the Peak Areas

Removal 0
 7-  98
 5- >99
                                                               Table 6
                                            Effluent Concentrations Achieved with Air Stripping

1 ,2-Dichloroethane
Ground water

99%,  having residual concentrations of 67.4  /tg/1
                                     and 0.7 /»g/l, respectively. Since there are no proposed ground-
                                     water quality  objectives  for  these compounds, it is not clear
                                     whether the residuals obtained through AS  alone would be con-
                                     sidered acceptable. If they  were not acceptable,  then AS would
                                     have to be augmented by some other process(es).

                                       Results from direct ozonation of the RO concentrate are shown
                                     in Table 7.
                                       The  removal efficiencies for  ozonation alone were generally
                                     either similar to or lower than those achieved by AS alone.  The
                                     reason for lower  removal during ozonation  was probably due to
                                     the lower gas flow rate through the column. In AS, the gas flow
                                     rate was in the ranges of 15 to  300 1/min; for ozonation, the max-
                                     imum gas flow rate was 2 1/min.
                                       In one  of the tests using  Oj alone, the ozonation column was
                                     operated  at a gas  flow rate of 2 1/min but the ozone generator was
                                     turned off. From the results of this test, it appears that approx-
                                     imately 30%  of the removal observed was due to the stripping ef-
                                     fect of the gas and 70% was due to the chemical oxidation of the
                                       With ozonation of air-stripped  RO concentrate (Table 7), the
                                     removal of the organic micropollutants generally improved,  par-
                                     ticularly for acetone and diethyl ether. It is not clear whether this
                                     additional removal could be sufficient to justify the use of ozona-
                                     tion in conjunction with air stripping.

                                     Ultraviolet Irradiation
                                       Ultraviolet irradiation was  used alone, as well as in combina-
                                     tion with  O3  or AS. For most compounds, the removal was not as

good as that achieved by AS alone. For this reason, the data from
these tests are not reported. The  reader is referred to an earlier
data report.8 However, since the best removal efficiency for most
of the organic compounds was achieved when all three unit pro-
cesses were applied, the removals and residuals from these tests
are presented in Table 8.

                           Table 7
   Average Removals of Organic Chemicals from Groundwater after
              Ozonation and Air Stripping/Ozonation
1 , 1-Dichloroethylene
1 , 1-Dichloroethane/THF
Diethyl Ether
1 ,2-Dichloroethane
1 , 1 , 1-Trichloroethane/
1 ,4-Dioxane
Dichloropropane (1,2- or 1,3-)
After AS

After Oa

After AS/
03 (%)

                           Table 8
      Percentage Removal and Residual Concentration Ranges
      Achieved During AS/Os/UV Treatment of Groundwater
1 , 1-Dichloroethylene
Diethyl Ether
Range of
Efficiency (%)
89- 99
38- 95
87- > 99
84- > 99
63- > 99
99- > 99
85- > 99
Range of Residual
  Generally, the range of removal efficiencies obtained by the
combination, AS/O3/UV, was higher than AS alone (Table 4).
The residuals were either similar or lower than the ones obtained
by AS alone.
  Based on these results, AS alone appears to be adequate to
remove volatile  organic compounds from the RO concentrate.
However, all three unit processes combined provided marginally
better removal efficiencies. It is doubtful, though, that the degree
of improvement would justify the increase in cost that would oc-
cur by including O3 and AS units in the treatment train.

  The conclusions drawn from this study follow below.
  Two pairs of compounds (1,1-dichloroethane/tetrahydrofuran
and  l,l,l-trichloroethane/l,4-dioxane)  coeluted during  the gas
chromatography analysis and could not be quantified individually.
  The results of data analysis by stepwise regression showed that
during air stripping experiments  the water flow rate had the
greatest effect on the removal of most organic compounds. The
best  removals were obtained at a water flow rate of 0.5 1/min.
  During  air  stripping  experiments,  four  compounds
(dichloromethane,  1,2-dichloroethane,  trichloroethylene and
benzene) were removed to below the proposed groundwater quali-
ty objectives.
  In the absence of groundwater quality objectives for acetone
and diethyl ether, it is  not clear whether the effluent concentra-
tions obtained by air stripping (67.4 and 0.7 /xg/1) would be con-
sidered as acceptable residual values. If they were not acceptable,
O3/UV in conjunction  with air stripping could be considered.
  Removal efficiencies observed after O3 or UV irradiation alone
were generally either similar to or lower than the results obtained
by AS alone. Combinations of AS/O3, AS/UV and O3/UV unit
processes compared to AS alone did not  offer significant im-
provements in the removal efficiencies in most of the tests.
  The removal  efficiency achieved by  all three  unit  processes
combined was better than that obtained by AS alone. For the
degree of improvement that can be achieved, it is not  clear that
the addition of O3/UV to the AS  process would be economically
  The data obtained in this study can be used to predict process
performance for those situations where the contaminated ground-
water has large  variations in the micropollutants concentration

1. Jackson, R.E., Patterson, R.J., Graham, B.W., Bahr, J., Belanger,
   D., Lockwood, J. and Priddle, M.,  "Contaminant Hydrogeology of
   Toxic Organic  Chemicals at a Disposal Site, Gloucester,  Ontario,"
   NHRI Paper No. 23, IWD Scientific Series No. 141, Ottawa,  Ont.,
2. GEC, WESA, Canviro, "Gloucester Landfill Waste Site Engineering
   Study—Problem Definition and Remedial Alternatives,"  submitted
   to Transport Canada as project #  L10137-1174 by A.J. Graham En-
   gineering Consultants Ltd., Water and Earth Science Associated Ltd.,
   and Canviro Consultants Ltd., Sept. 1984.
3. Whittaker, H., Adams, C.I., Salo, S.A. and Morgan, A., "Reverse
   Osmosis at the Gloucester Landfill," Proc. of the Technical Seminar
   on Chemical Spills, Environment Canada, Toronto, Ont.,  Feb. 1985.
4. Canviro Consultants Ltd., "Treatment of Organic Contaminants in
   Landfill Leachate:  Final Report,"  submitted  to Wastewater Technol-
   ogy Centre, EPS, Burlington, Ont.,  May  1984.
5. APHA, AWWA, WPCF, Standard Methods for the Examination of
   Water and Waste-waters, New York, NY, 15th ed., 1980.
6. Bader, H.  and Hoigne, J., "Determination of Ozone in Water by the
   Indigo Method; A Submitted Standard  Method," Ozone: Science
   andEng.,  4,  1982, 169-176.
7. Hatchard, C.G. and Parker, C.A., "A New Sensitive Chemical Acti-
   nometer II.  Potassium Ferrioxalate as a Standard Chemical Acti-
   nometer," Proc. of the Royal Soc. of London, Series A.,  235,  1956,
8. Simovic, L. and Jones, J.P., "Removal of Volatile Organic from RO
   Concentrated Contaminated  Groundwater—Gloucester Site," Un-
   published manuscript, Environment Canada,  Wastewater Technology
   Centre, Burlington, Ont., 1985.
                                                                                                   ON-SITE TREATMENT    123

             State-of-the-Art Technologies of Removal,  Isolation
          and  Alteration  of  Organic  Contaminants Underground
                                                 Walter W. Loo
                                                George N. Butter
                                      McKesson Environmental Services
                                             Pleasanton, California
  The rising cost and  increasing long-term liability of off-site
treatment and disposal  have caused a demand for on-site in situ
treatment of organic contaminants located underground. There
are generally three types of in situ treatment: removal, isolation
and alteration.
  The purpose of this paper is to provide the reader with a de-
scription of available state-of-the-art, in situ treatment alterna-
tives for the treatment  of various organic contaminants in  both
the vadose and saturated zones in the subsurface. Guided format
tables are developed for the reader to search for the application
that may work in an individual situation while considering regu-
latory, economic, risk and liability consequences.

  The treatment and disposal of hazardous waste and material
off-site has become increasingly difficult due to closure of an al-
ready limited number of qualified waste handling and disposal
facilities, tightening of regulatory standards, escalating transpor-
tation and disposal costs and risks of long-term liability. While
the technology of in situ treatment is advancing at a rapid pace,
there is a need to evaluate periodically the applicability of various
technologies to specific  on-site  contamination problems and
underground conditions.
  The purpose  of this  paper is to provide the reader with a de-
scription of available in situ treatment alternatives for applica-
tion on  various organic contaminants in both the vadose and
saturated zones  in the subsurface. However, the best technology
will work only if it is in agreement with regulatory remediation
standards and permits, economic feasibility and  long-term  risks
and liability as shown in Figure 1.
  The following sections discuss the three possible types of in situ
treatment alternatives:  removal, isolation and alteration. These
alternatives will encompass a broad combined spectrum of science
and engineering areas:
  Environmental Engineering
  An  initial requirement for on-site remediation always  is the
removal of the source and control of the contaminated plume,
whether it is in the soil  (vadose zone)  or in the groundwater
(saturated  zone) or both. The in situ removal process may in-
volve primary, secondary and tertiary methods.
     Regula tory
     Remedia tion
     Standards &
                    Risks & Liability
                        Figure I
  The primary methods include the following:
• Vadose zone—air vacuum venting
• Saturated zone—pumping of recovery well

  The air vacuum venting works well with volatile organic com-
pounds located underground,  in particular, permeable strata in
the vadose zone where the soil  can serve as a migration pathway.
In general, the venting process  will require an air emission permit
which may be prohibitive in some areas. If organic emissions to
the atmosphere are not allowed, an  oil stripper (OS) or an acti-
vated carbon adsorption unit  is required. If the vent stream is
steady and  loaded with high concentration volatile organics, it
may be advisable to recover the product by either an oil strip/de-
sorption or a regenerative activated carbon process (Fig. 2).
  The secondary method of removal involves both injection and
recovery wells. Usually, injection wells for the vadose zone should
be discouraged  because  the  injected  air/water  stream  may
"push" the volatile organics  toward  undesirable off-site  loca-
  The tertiary method of removal involves temperature enhance-
ment and addition of surfactants to  increase the release rate and
mobility of the organic compound movement through the porous
media toward ultimate recovery. The addition of surfactants may
create new  problems in recovery and surface treatment if not
properly applied. Temperature enhancement is generally a desir-

                          Figure 2
         Generic Flow Diagram of In Situ Removal Process
able process (if duration of treatment can be prolonged) by a
passive solar heating system. The system works as underground
space heating under steady circulation. The system cost and main-
tenance are usually very attractive.
  For the saturated zone, the application is similar to the vadose
zone with the exception of depth of the contaminated zone and
liquid handling.         \
  The following design criteria are important for  recovery well
field applications:

• The recovery rate must exceed the recharge/safe yield of the
  geohydrologic system
• Drilling and well installation design must now allow vertical
  migration of contamination from shallow depth
• Pumping of recovery wells must begin with shallow wells and
  move to deeper wells to avoid pressure induced vertical migra-
  tion of contamination
• It is not advisable to establish pumping wells in low perme-
  ability material (generally less than 1 X 10~4cm/sec)
• Well screen design must be designed specifically for "sinkers"
  or "floaters" organic compounds
  Unfortunately there  will never be 100% removal of the con-
taminants. Some portion of the contamination will remain affixed
to the fine grain matrix in the subsurface material.

  There  are  specific geohydrologic conditions which  will not
allow in situ removal treatment. Also, less volatile  organic com-
pounds may not be easily removed by even secondary or tertiary
treatment processes. The utilization of hydraulic isolation may be
the solution in these situations. Hydraulic isolation can be used as
a follow-up to in situ treatment after removal of a major portion
of the contamination.
  There are three different  in situ hydraulic isolation methods:
surface cover, subsurface barriers and sorption processes.
  The surface cover method generally involves placement of a
clay or concrete cover  over a defined contaminated area to pre-
vent percolation of surface water into the contaminated zone to
contact the contaminants below the surface.
  Subsurface barriers include a number of isolation techniques:

  Slurry trench walls
  French drain
  Sheet pile
  Interceptor ditch
  Curtain grouting
  The first four types of hydraulic barriers are best for hydraulic
isolation of low permeability media and shallow depths (20 to 40
ft maximum). Curtain grouting involves injection of impervious
material into permeable strata to block off potential  migration
pathways in the subsurface environment.
  Sorption in situ treatment is a relatively  new technology in-
volving  replacement of the permeable strata in a known contam-
inated area with a specially formulated fluid with both absorption
and adsorption properties. In the absorption process, the material
will expand up to 30 times its original volume while absorbing the
contaminant and blocking off or filling in the pore spaces. Due to
the microscopic size of the material, the large surface area of con-
tact with contaminants will make an effective sealant to the per-
meable  pathways when  contaminants are  released.  The formu-
lated fluid  will  work  for both polar and non-polar organic com-
pounds. Since this treatment  process is still  in the experimental
stage, it may not be very cost-effective at the present time.
  A shallow interceptor  ditch may work in relatively low perme-
able saturated material. The ditch may act and perform like a line
of hundreds  of wells. Sometimes a combination of the in situ
hydraulic isolation techniques may work for sites with multiple
organic  compounds of different physical and chemical properties
when other treatment processes do not provide a total solution.

  In situ alteration treatment involves the breakdown of haz-
ardous  organic compounds into  harmless compounds. It has
been  demonstrated  that  biodegradation  of  non-halogenated
organic  compounds to harmless compounds can be effected in
surface  aerobic treatment of sludge and waste water. The same
result can be  attained in  subsurface conditions which control  the
growth  of aerobic bacteria. The parameters controlling bacterial
growth and metabolism are:
  Dissolved oxygen

  The transport mechanism in the vadose and saturated zones is
the major problem that hinders subsurface treatment. If the con-
taminated fluid can be pumped to the surface, it is better to per-
form surface  treatment.  The application of the in situ alteration
process  should not be a primary means of treatment.
  The in situ alteration treatment of halogenated hydrocarbons is
still in the experimental stage.  The  recent advances are well docu-
mented  in the references  found at the end of this paper. The pro-
cess of breakdown of halogenated hydrocarbons depends on the
successful control of methanogenic conditions (addition of meth-
ane mixed  with air to the contaminated soil and  water).  When
aerobic  bacteria are exposed to methanogenic conditions, selec-
tive species  will degrade the halogenated organic compounds into
carbon  dioxide and  chlorides which are harmless compounds.
While the researchers are developing a better understanding of the
mechanism of degradation, forseeable problems of field applica-
tion are  as follows:
• Aerobic bacteria only occur in shallow depths
• A specially formulated carrier fluid must be coupled with meth-
  ane and other nutrient when introduced underground
• Introduction of methane gas  underground may create haz-
  ardous conditions (i.e., explosion) if not properly contained
                                                                                                ON-SITE TREATMENT    125

                            Table 1
     Practical Application of In Situ Treatment of Volatile Organic*
                            Table 2
    Practical Application of In Situ Treatment of Acid/Base Neutral
               Extraclables, Pesticides and Herbicides
In- si tu Trea traenl
Remova 1

• Flushing
tempe ra ture


• Subsur face
• SorplLon
Alt a tto
. Aerobic Process

NA Not applicable
(a) Air stripper
(b) Oil stripper
Vadosc Zone

ef f ec tlve

(b) or (c)
Yes, need (a)
(b) or (c)
(b) or (c)
Vol. n««d Id)
Saturated Zone

Yes, but


Tachn Ique a
Renova 1
Yti , need * Pu» ptng/e vacua 1 1 on
{a) (b) or (c) to create sink
Yet. need • Flushing
(a) (b) or (c)
(a) (b) or (c) temperature
enh«nc*aen t
I tola I i on
Yt i • Surface cover
bo r r I e r t
Yet, need (d) • Sorptlon

A 1 tera t Ion

( • ) Oil s tripper
Vadote Zone


N««d (c)
Long dura t Ion
Yet, r.«ed (c)
Saturated Zone

Need (a) or
Need («) or
(b) and (c)
Lont duration
Yea, can be

£«(,er .o»jntal for halo-enaled organic co« pound*. Need

(c) Detail geohydrologlc definition
   The understanding of the anaerobic degradation of halogen-
 ated organic compounds is in its infancy, but it has great poten-
 tial. Because of the low  cost of bacteriological processes, it is
 worthwhile to have engineers perform opportunistic  trial and
 error experiments on both laboratory bench and field pilot tests
 using this technique. Once demonstrated,  the cost saving over
 other in situ treatment alternatives may be substantial.
   At the present state of knowledge of biodegradation, there is
 immense room for improvement  and development of original re-
 search and application.  Hopefully, the economic  advantage  of
 innovative in situ technology will push the  alteration techniques
 to a major breakthrough in the near future.

   The above mentioned in situ treatment techniques may work
 individually or in combination for  specific  contamination cases.
 However, there  is no general formula to  produce an optimal
 remediation solution for a case-specific problem.
   An attempt is made to provide  a general guidance  for prac-
 tical application of the specific in  situ treatment techniques on
 organic compounds in the vadose and saturated  zones  in Tables
 1 and 2. These tables will not provide a total solution for a spe-
 cific site. However, they will provide an individual with an oppor-
tunity to evaluate the "cans," "cannots" and effectiveness of the
full spectrum of the available alternatives.
  These in situ treatment techniques are only one aspect of the
total remediation plan. The engineer will  still have to consider
regulatory standards and permits, economic feasibility and long-
term risks and liability for a total remediation program.

1.  Bouwer, E.J. and McCarty, P.L.. "Transformation of 1- and 2- Car-
   bon Halogenated Aliphatic Organic Compounds Under  Methano-
   gcnic Conditions," Appl. and Environ. Microbiol., 45, 1983, 1286-
2.  Haber, C.L. el al.,  "Methylotrophic Bacteria: Biochemical Diver-
   sity and Genetics," Science 221, 1983, 1147-1153.
3.  Parsons, F., el al., "Transformations of Tetrachloroethene and Tri-
   chloroelhene  in  Microcosms  and  Groundwater," JAWWA,  76,
4.  Wilson, J.T. and McNabb, J.F.. "Biological Transformation of Or-
   ganic Pollutants in Groundwater," American Geophysical Union EOS
   64, 1983,505-506.
5.  Wilson, J.T. and Wilson, B.H., "Biotransformation of Trichloro-
   ethylene in Soil," Appl. and Environ. Microbiol.. 49, 1983, 242-243.

               Assessment  of  Volatile Organic Emissions from a
                       Petroleum Refinery Land Treatment  Site

                                         Robert G.  Wetherold, Ph.D.
                                                 Bart M. Eklund
                                              Radian Corporation
                                                  Austin, Texas
                                          Benjamin  L. Blaney, Ph.D.
                                   U.S. Environmental Protection Agency
                          Hazardous Waste Environmental Research Laboratory
                                                Cincinnati,  Ohio
                                              Susan  A. Thorneloe
                                   U.S. Environmental Protection Agency
                               Office of Air Quality Planning and Standards
                                           Durham, North Carolina
  A field assessment was performed to measure the emissions of
volatile organics from a petroleum refinery land  treatment site.
As part of this study, the emissions of total volatile organics from
surface-applied and subsurface-injected oily sludge were meas-
ured over a 5-week period. The effect of soil tilling on the emis-
sions also was monitored.
  Volatile organics emission rates were measured using the emis-
sion isolation flex chamber method. Soil samples were collected
during the test periods to determine soil properties, oil levels and
microbe count. Soil surface and  ambient temperatures, both in-
side and outside the flux  chambers, were measured throughout
the test periods.

  The U.S. EPA currently is developing background information
on air pollutants at hazardous waste treatment, storage and dis-
posal facilities (TSDFs) to provide supporting data for  setting
standards as necessary under Section 3004 of RCRA as amended
in 1984. This regulatory development  is the responsibility of the
Office of Air  Quality Planning  and  Standards (OAQPS). The
U.S.  EPA Hazardous Waste Environmental Research Labora-
tory-Cincinnati (HWERL) has the responsibility of providing
technical support to OAQPS in the area of atmospheric emis-
sions determination from hazardous waste management.
  In order to assess emissions from hazardous waste TSDFs, a
field study was performed  at a land treatment facility located at a
large West Coast petroleum refinery. This study was intended to:
(1) measure the total volatile organics emission rates from sur-
face- and subsurface-applied oily refinery sludge and (2) to de-
termine how emissions of volatile organics  varied  with time.
Based on these measurements the total emissions over a 5-week
period were estimated.
  Little work has been performed to characterize volatile organic
emissions from large full-scale land treatment sites. The study
approach and selected  results  are summarized in this paper. A
more detailed  description of the work and its limitations in in-
cluded  in a previous U.S. EPA  report.1 The emissions of indi-
vidual  measured  compounds  are provided  in that  U.S. EPA
  The test site was located at a large West Coast refinery. The
land treatment area covers approximately 42,000 m2 (10 acres).
Three test plots, each approximately 420 m2  (0.1  acre)  in area,
were located side-by-side at one corner of the facility. The area
containing the test plot has been in land treatment service for
several years.
  During normal operation,  sludge from  refinery operations is
collected and  stored in a fixed-roof tank  at the land treatment
area. The sludge is injected subsurface using an 8,0001 (50-barrel)
capacity injector  vehicle equipped with  four separate  injector
tines. The injector vehicle  consists of a tractor and an 8,000 1
(50-barrel) capacity tank trailer. The tines are attached to the tank
trailer. The waste is injected  15 to 28 cm below the surface and
immediately disked by a disk unit pulled behind the injector ve-
hicle. The area is cultivated with a disk and/or rototiller several
times per week until the next sludge application.
  According to the operators of this facility, the annual  applica-
tion rate at this site is about 5.4 - 9.1 x 106 kg/year, and the in-
dividual application rates are typically about 16 1/m2 (400 bar-
rels/acre). The sludge typically consists of 50  to 75% DAF/API
separator float, 20 to 30% API separator sludge and about 5%
miscellaneous oily waste. The soil is  irrigated periodically to
maintain proper soil moisture content.

  The  experimental  design,  sampling  techniques,  analytical
methods and test procedures are summarized below.

Experimental Design
  The experimental design of this program was developed to ob-
tain data of sufficient accuracy to fulfill the major objectives of
the study. The primary emphasis of the program was on obtaining
volatile organics emission rate measurements. The test design was
a synthesis of two sampling strategies  for obtaining emission
measurements: totally randomized sampling over  the test plots
and semi-continuous sampling at a single location in each plot.
  To provide an overall estimate of the volatile organics emission
rate for a total test plot area,  a number of measurements at ran-
domly selected points  must be performed. One method of ac-
                                                                                          ON-SITE TREATMENT    127

complishing this is to divide the area into discrete segments and
then randomly select segments to be tested. On the other hand,
the best strategy for examining the change in emission rate with
time is the repeated sampling of a limited number of points over a
period of time.
Three test plots were used in this study. Sludge was surface-
applied to one plot (Plot A) and subsurface-injected at another
plot (Plot C). In between these two plots was a third plot (Plot B)
with no sludge applied. Plot B served as a baseline or control plot.
Each plot was approximately 420 m2 (0. 1 acre) in size, and each
was divided into 21 segments, as shown in Figure 1 , to provide for
randomized sampling.
Subturlac* Background Surlact
Application Plot Plol Application Plot




Table 1
Typical Dally Sampling Schedule
Flux Charter
ApprmlMt* Grid Cm F/C £MU?IM Cnll«rfpd
"TIB* Point* 10 ID" Canlitw Soil Syrlng* Couwnt*
OBOO 	 002 003 - - X OC
0843 A07e 002 003 - X X
0930 603 002 003 X XX
1013 A03 002 003 - - X
1100 AI9C 002 003 X X X
IMJ AOI 002 003 X
1230 	 002 003 - - - TIM
r 1400 A07e 002 003 X - X
1449 Mr 002 003 X - X
1330 B14 002 003 X X
0600 	 001 002 - - X OC
0845 C1IC 001 002 X X X
0930 CI3 001 002 X
1015 CI4 001 002 - - X
5 1100 CIS* 001 002 - - XX
E 1145 BIS 001 002 - - . X
: 1230 	 001 002 - - - Till
1315 C13e 001 002 X X
1400 Cll' 001 001 X
1449 C04 001 002 - X X
1530 C06 001 002 X X X
1615 BOS 001 002 X
i Prefix A. B, C refers lo plot A (surface applied). pk>l B (background or control plot), and
Eacn Plol . O3 lq m.l.fl (4.450 K). II.)
Eacn Grid > 20.1 M m.l«r« (217 M IIJ
b. F/C « flux chamber
C. Control points
                           Figure 1
       Test Plot Grid Assignment for Assess; ;ent of Emissions
             from Land Treatment of Petroleum Wastes

   Sampling schedules for each  plot  were developed for each
 separate day of sampling. Each sampling day was divided into
 two parts. During  each half day,  one  sampling crew  made
 measurements at the surface-applied plot, and the other crew per-
 formed similar measurements at  the subsurface injected plot.
 Each crew performed emission measurements  at  two  separate
 control points, at two randomly selected points  in their assigned
 plot and at one randomly selected point in the control plot. Ran-
 dom number tables were used to select the order of sampling, the
 crew used  on each plot and the equipment set used by each crew.
 The sampling order for composite soil samples, soil cores, gas
 samples and their associated duplicates within a given plot were
 also randomly  selected.  A  typical  daily  sampling schedule  is
 shown in Table 1.

 Sampling Procedures
   The main sampling technique employed at this site involved the
 direct measurement of emissions using the emission isolation flux
 chamber.  The enclosure emission measurement  approach has
 been used by others to measure emission fluxes of sulfur and
 volatile organic species.2-iA
   A diagram of the flux chamber is shown in Figure 2.  The flux
 chamber was placed on the emitting surface. Clean, dry air was
 passed through the chamber at a controlled  and measured rate.
 The concentration of the specie(s) of interest was measured at the
 outlet of the chamber. The emission rate of the measured specie(s)
 from the enclosed surface was then calculated from the flow rate
 and concentration in the gas.
                          Figure 2
      Cutaway Side View of Emission Isolation Flux Chamber
                  and Sampling Apparatus
   The flux chamber encloses a surface area of 0.13 m1 and has a
 volume of 0.031 m'. The sweep air  flow rate was 0.003 - 0.005
 mVmin. Air samples were collected from the outlet gas line using
 both  gas-tight syringes (for on-site  analyses) and evacuated
 stainless steel canisters (for off-site analyses).
   Liquid grab samples of the sludge and of the service water used
 to irrigate the soil were collected. Sludge from two tank trucks
 was transferred to the application vehicle (tractor) and applied to
 the soil. Each tank truck held two application tractor loads. After
 the first load was transferred from the tank  truck to the applica-
 tion tractor, 4-6 sample containers were filled with  sludge from
 the tractor. The samples were collected consecutively, while the
 tractor tank was full, from a gravity-fed line. While two tractor

loads were required to empty each tank truck, samples were col-
lected after only one of the two transfers from each tank truck
load. One service water sample was collected from the  plant ser-
vice water line to determine the concentration of organics in this
water which the facility applies  as part of its  site  management
  Two types of solid samples were collected: soil composites and
undisturbed soil cores. Both types of samples were  collected im-
mediately adjacent to points where flux chamber measurements
were made. Soil composites were used in the determination of oil
and moisture contents. These samples were collected using a hand
auger and boring to a depth of 30 to 40 cm. The undisturbed soil
samples were used for determining the physical properties of the
soils. They were obtained by driving a brass sleeve approximately
20 cm into the ground, removing the sleeve and capping both
  Type K thermocouples were used to monitor the air and soil
temperatures both inside and outside each flux chamber. Two
separate thermocouples  were  used  to  measure  soil  surface
temperatures inside and outside the flux chamber. Each was plac-
ed in the ground to a depth of 1.3 to 1.9 cm.

Analytical Procedures
  The on-site analyses were limited to gas-phase analyses of the
air  samples collected in gas-tight syringes from the  outlet of the
flux chambers. A Byron Instruments Model 401 Total Hydrocar-
bon (THC) Analyzer was used to determine the concentrations of
total hydrocarbons (THC), CO2  and methane in the effluent air
samples from the flux chamber.
  The off-site analyses included the  chemical speciation of the
flux chamber air samples collected in stainless steel  canisters and
of the liquid sludge and water samples collected at the land treat-
ment site. The oil, moisture and microbe levels in the soil  also
were determined,  as  were the physical properties of the  soil
  Chemical speciation of the air  and liquid samples for C2 and
C10 hydrocarbons  was performed with  a Varian  3700  Gas
Chromatograph. The chemical species were cryogenically concen-
trated to increase the sensitivity of the analyses. Liquid samples
were analyzed  using a purge-and-trap technique  modified to
utilize the cryogenic trap. For both air and liquid samples, total
volatile organics concentrations were obtained as the sum of the
species concentrations.
  The oil,  water and solids contents of the sludge samples were
determined by solubilizing the oil and water in the waste with
tetrahydrofuran (THF). This liquid  phase was then  separated
from the insoluble solids. Residual solids were weighed, and the
amount of water in the THF was determined by Karl Fischer titra-
tion. The oil content was  calculated by difference. The oil and
grease content of the soil samples was determined using U.S. EPA
Method 413.1. Standard methods were used to determine the bulk
density, particle density, total porosity, moisture content and par-
ticle size distribution.

Sludge Application
  The sludge was applied as evenly as possible to the surface-
applied Plot A  (SA plot) and to the subsurface-injected Plot C
(SSA plot). The application tractor started injecting sludge at grid
point (GP) C19 (Fig. 1) and moved across Plot C from GP  C19
through GP C21. The injector then was lifted above the soil, and
the tractor proceeded across Plot B (control plot) from GP  B19
across GP B21. No sludge was applied to Plot B. Sludge then was
allowed to flow from the raised injector tines onto the surface of
Plot A as the tractor passed across it from GP  A19 through GP
A21. The tractor then made a U-turn and returned across Plots A,
B and C while applying sludge.  Several application passes were
made to complete the application to the entire area of Plots A and
C. Three  complete tractor loads  and part  of another were re-
quired. The average waste loading was 1.4  x  104 kg/plot. All
three plots were tilled with a disk  tiller pulled behind the tractor
during the sludge application.
  Emission  sampling with the flux  chambers was started im-
mediately  after application and continued for 4 days.  Emission
sampling was performed during two other 4-day periods in the 3rd
and  5th weeks following  application. Each plot was  tilled 2-3
times per week during the 5-week test period. The plots also were
irrigated with water once between the first and second sampling
periods and three times between the  second and third sampling
periods. The complete tilling and irrigation  schedule for the test
period is summarized in Table 2.
                          Table 2
   Tilling Schedule for Land Treatment Facility During Test Study
Till Ing
Hours from Start a
Apply waste and till
Till and Irrigate
TIM and Irrigate
Till and Irrigate
Till and Irrigate
a. Rounded off to nearest 1/2 hour.
  The results pf the sample analyses for the gas, liquid and soil
samples are summarized below.

Gas Samples
  The results of the on-site Byron THC analyses agreed quite well
with the off-site Varian 3700 analyses. The results were correlated
with a correlation coefficient of 0.977. The Varian GC generally
gave slightly higher concentrations, with an average difference of
13.8% between the Byron and Varian instruments.
  No methane was detected in any of the on-site samples analyzed
with the Bryon 401, which is capable of detecting methane at
levels of 1 ppm.
  The CC>2 concentrations in the on-site gas samples were found
to range from  about 50 ppm to almost 500 ppm. Most values fell
between 100 and 300 ppm. Background CO2 levels,  as determined
by sealing  the bottom  of the flux chambers and  passing  air
through them, were in the range of 10 to 200 ppm.
                                                                                                 ON-SITE TREATMENT     129

Liquid Samples
  The sludge samples  were analyzed  several times to define the
sludge nonhomogeneity and obtain an average value for the level
of volatile organics. The analyses were performed over a period of
several months (during which the samples  were  kept  under re-
frigeration).  The  measured volatile organics content  of  the
sludge, as determined with the purge-and-trap technique, ranged
from 4,250 to 16,600 /ig/g. Most volatile organics concentrations
were between 4,000 and 6,000 /ig/g,  with the mean  value  being
7,800 jig/g. The variability was believed  to be at least partially at-
tributable to the difficulty of obtaining a representative sample in
small quantities for analysis.  The sludge volatile organic  com-
ponent was made up  principally of  paraffins (average concen-
tration = 3,600  pg/g),  olefins  (1,200 ug/g)  and  aromalics
(2,900 jag/g). The  remainder of the  volatiles were halogenated
and oxygenated hydrocarbons and unidentified organics.  From
the  characterization tests, the  sludge was  found to consist  of
3.9 to 4.5% solids, 57.5 to 59.1 % water and 37.0 to 38.0% oil.
   The sample of service water used to  irrigate the soil was also
analyzed  for volatile organics to  determine whether this  water
could be a significant source of volatile  organics (this sample was
kept under  refrigeration and analyzed  approximately  3  months
after collection).  The  volatile organics  content was found  to be
0.089 mg/1. The total amount of volatile organics applied with the
water over the duration of the test was estimated to be about 0.03
kg/plot. This amount is  equivalent  to less than 0.5% of the
volatile organics applied in the sludge.

Soil Samples
   The mean values for the physical properties determined for the
soil samples  are summarized in Table  3. Also included are the
mean moisture content  and oil/grease levels  for the collected
samples. The results are summarized for two of the three  sam-
pling periods.
                           Table 3
                     Mean Soil Properties

^g/cm )
lot J>
Density Porosity,
(g/ciT5) (J)
1.21 52
1.42 44
Oil and
  A statistical analysis was performed primarily to investigate the
effects of time, temperature inside  the chamber  (soil and air),
plot samples, location within each plot and chamber shading on
emission rates. The frequency distribution of the emission rate
data was highly skewed and a log normal distribution was used
to model the frequency distribution of emission rates. To de-
termine whether measurable differences existed among the emis-
sion rates from the test plots, an analysis  of covariance was per-
formed. The analysis of covariance technique eliminates the vari-
ation in emission  rates due  to other  variables  such  as  time,
temperature  (inside the chamber) and chamber surface shading.
In this way,  the average emission rates for the test plots can be
directly compared. A discussion of the  analysis  of covariance
can be found in Brownlee.'
  The statistical analysis indicated that the emission rates  were
significantly affected by test plot and elapsed time. The results of
the statistical  analysis are summarized in Table  4. Soil or air
temperatures within the chamber were marginally significant.
None of the other variables were found to have a significant af-
fect on the emission rates.

                           Table 4
                   Components of Variability
Source of  Variability
                                 Percent of
                      Component   Total  Variance
Percent of Total
 Minus Temporal
Temporal (day-to-day)      58.3          63.0

Air Temperature
 In Chamber               3.3          3.6

Plot                   28.1          30.4

Sanpl Ing Location
(•Ithln the plot)
Surface Chamber
Sampl Ing/ Analytical
a Variance components arc equal to standard dotations squared and ihus have units which are
  the squares of that used for the log-emivsion rate% f>ig m--sccp
b Relative contribution of parameters to the total tanance

  The oil and moisture contents of the soil samples were exam-
ined to determine whether there was any significant change with
time in these parameters. There appeared to be little change in the
oil content of the soil over the duration of the test. The moisture
content of the soil appeared to decrease by about 14 to 40% over
the test period. Both changes were so small that no  correlation
with emission rates could be determined.

  The volatile organics emission rates were examined with respect
to elapsed time since application, ambient conditions  at the time
of sampling and tilling activities.
                                            A  =  Surface
                                             3  =  Background
                                            C  =  Subsurface
          01234    5678     9  10 11  12
                             Sampling Days
            f = tilling episode

                           Figure 3
       Measured Emission Rates as a Function of Time Elapsed
                    Since Sludge Application

                           Table 5
       Average Measured Emission Rates by Plot by Half-Day
                                               Table 6
                                     Cumulative Measured Emissions
Fmlsslon Rate
Sampl Ing












A Plot
B Plot C Plot
(Background) (Subsurface) Comments


13.8 early start
295 till
91.3 till
12.1 early start
59.3 till
16.2 till
16.2 till
 Each Plot Area = 423 m2
 Effect of Elapsed Time from Application
 on Emission Rates
   The measured emission rates for the three plots are shown as
 functions of elapsed time in Figure 3. The average measured emis-
 sion rates are summarized for  each half-day of sampling. These
 rates are tabulated in Table  5.  The  first point (Day 0)  was
 measured before sludge application. The tabulated values and the
 figure show the approximately exponential decline  in  emission
 rates with time elapsed from application (and, to some extent,
 elapsed time from tilling  episodes). The emission rates from the
 control Plot B remained substantially below those of the test plots
 throughout the duration of the testing.
   The cumulative measured  volatile organics emissions  are sum-
 marized in Table 6 and shown graphically in Figure 4. An approx-
 imately exponential decline in the emission rates is evident.
             A = Surface Applied
             B = Background Plot
             C = Subsurface Injected
                         Sampling Days
10  11  12
Fml<;>:lon Rate (kn)
Sampl Ing
A Plot
B Plot
C Plot
(Subsurface) Comments
1.293 early start
3.646 till
5.072 till
5.136 early start
5.639 till
5.793 till
6.000 till
                    Each Plot Area = 423 m2
                   Effect of Tilling on Emission Rates
                     Tilling of the soil had a significant effect on the emission rate.
                   The effect can be seen clearly in Figure 3 as spikes in the emission
                   rates immediately after tilling. The percentage increase in average
                   emission rates by half-day immediately following tilling episodes
                   is shown for each plot in Table 7.  The five tilling episodes for
                   which emission measurements were made were distributed among
                   the three sampling periods.
                                                                                            Table 7
                                                                               Increase in Emission Rate Due to Tilling
A Plot
Increase In Emission Rate
B Plot

C Plot
                          Figure 4
   Cumulative Measured Emissions as a Function of Sampling Day
(includes only measured rates for days on which sampling was performed)
a. Incomplete data set

  The  emission  rates and  cumulative emissions  for  the  entire
5-week test period were estimated from the measured  rates. The
following assumptions were made to estimate the total  emissions:
                                                                                                  ON-SITE TREATMENT    131

• Tilling had no effect on emission rates beyond 6 hours after
  tilling occurred
• Each day had 12 hours of cooler temperatures (night)
The emission rates were estimated for three distinct times during
each sampling period. These were: (1) nighttime, (2) non-tilling
daytime and (3) tilling daytime. The nighttime emissions for each
sampling period were estimated from the emission rate data ob-
tained during the early morning hours of that period. The tilling
and non-tilling day emissions for each sampling period were de-
rived  from  the appropriate emission rate measurements. Emis-
sions  for the interval between sampling periods were obtained by
interpolation among the three  emission rates for the three sam-
pling periods. Summing all the night-time, non-tilling daytime and
tilling daytime emissions, the total emissions for each of the plots
for the 5-week period were estimated to  be:
   Plot A (surface applied):       33.3 kg
   Plot B (control):               5.2 kg
   Plot C (subsurface inj.):        39.0 kg
   The estimated total emissions per plot per week are shown  in
Figure 5.
» 18
= o IK
eo* ID
1 Jf 14
sl 12
Q) 10 _
Estimated We
Organic Emi
to *. a> CD a
A = Surface Applied
„ B = Background Plot



C = Subsurface Injected



I/A/' \~ fVyfTPxJ f77t7~T^
                          Figure 5
        Total Estimated Weekly Emissions from Each Plot
   The results  of  the  detailed sampling and  analysis of total
volatile organic emissions from this land treatment  facility in-
dicate that emissions from both surface-applied and subsurface-
injected  waste varied in a similar, approximately exponentially
decaying manner over the 5-week period of this test. This result is
in agreement  with the predictions  of the Thibodeaux-Hwang
emissions model for surface-applied plots.'
  Also,  the  effect of tilling  is similar for  both application
methods. Tilling resulted in a 2- to 10-fold increase in the emission
rates. A  large portion  of the 5-week cumulative emissions from
both plots is due to emissions which occurred within 4 hours after
tilling was performed.  The magnitude of emissions which occur-
red immediately after tilling decreases from tilling event to tilling
event approximately in proportion to  the decrease in emissions
during non-tilling periods.
  Based  on a statistical analysis of  the emissions,  it was found
that the emission rates were significantly affected by the elapsed
time. Soil and air temperatures within the chamber were mar-
ginally significant. None of the other  variables that were mon-
itored  were found to  have a significant effect on the emission
  The information  in this document has been funded wholly or
in part by the U.S.  EPA under contract to Radian Corporation.
It has been  subject to the Agency's peer and administrative re-
view, and it has been approved for publication. Mention of trade-
names or commercial products does not  constitute an endorse-
ment or recommendation for use.
1. Radian Corporation,  "Field Assessment of Volatile Organic Emis-
   sions and Their Control at a Land Treatment  Facility," Draft Re-
   port, DCN 85-222-078-15-05, U.S. EPA  Contract No. 68-02-3850,
   Oct. 1985.
2. Hill, F.B., Aneja, V.P. and Felder, R.M., "A Technique for Meas-
   urements of  Biogenic Sulfur Emission Fluxes," J.  Environ.  Sci.
   Health, AIB(3), 1978, 199-225.
3. Adams, D.F., Pack, M.R., Bamesberger,  W.L. and Sherrard, A.E.,
   "Measurement of Biogenic Sulfur-Containing Gas Emissions from
   Soils and Vegetation," Proc. of 71sl Annual APCA Meeting, Hous-
   ton, TX 1978.
4. Schmidt, C.E., Balfour,  W.D.  and Cox, R.D., "Sampling Tech-
   niques  for Emissions  Measurements  at Hazardous Waste Sites,"
   Proc. of 3rd  National Conference on Management of Uncontrolled
   Waste Sites, Washington, DC, 1982, 334.
5. Brownlee, K.A.,  Statistical Theory and Methodology in Science and
   Engineering, John Wiley & Sons, Inc., New York, NY, 1965.
6. Thibodeaux, L.J. and Hwang, S.T., "A Model for Volatile Chemical
   Emissions to Air from Landfarming of Oily Waste," paper presented
   at the Annual Meeting of the American  Institute of Chemical En-
   gineers, New Orleans, LA, Nov. 1981.

                               Technology  for Remediation  of
                                 Groundwater  Contamination

                                            David  V. Nakles, Ph.D.
                                                James E. Bratina
                                                     ERT, Inc.
                                            Pittsburgh, Pennsylvania
  The AquaDetox* technology is a high-efficiency stripping tech-
nology for the removal of organics from water. The technology
was developed by the Dow Chemical Company for internal use
through considerable process research, and only recently has the
technology become available for general use outside of Dow via a
licensing program.
  Two types  of AquaDetox*  units are available,  air or steam
AquaDetox*. Air AquaDetox* is capable of efficient removal of
chlorinated solvents and other pollutants considered easily strip-
pable from water. Steam AquaDetox* is used for the removal of
low-volatility organic materials from water including many com-
pounds not   considered  strippable   by  any  prior   process
  The technology has a wide range of applications  including
removal of 92 of the 111 organic priority pollutants as well as
other organic materials. It is applicable to the removal of trace
quantities of organics in water intended for drinking water pur-
poses with the capability of producing waters with organic con-
centrations from /ig/1 down to non-detectable levels.
  This paper describes the AquaDetox* technology, describes its
potential applications and presents economic and operating data
from  several operating commercial units.

  Extensive groundwater contamination from both active and in-
active sites  has been identified across the United States. It has
been stated that more than 700 synthetic organic chemicals have
been discovered in groundwater resulting from over 30 different
types  of sources of contamination. The House Public Works Sub-
committee of Investigations and Oversight has estimated from an
inquiry last year that 4,000 city drinking wells were affected by
hazardous waste seepage into groundwater. Hence, the subcom-
mittee has assessed groundwater contamination as potentially one
of the most expensive problems confronting Congress and the
  ERT has been involved extensively in the assessment of ground-
water contamination at a variety of active and inactive sites.
These efforts have focused largely on determining the nature  and
extent of contamination, modeling the migration of contaminants
beyond the site boundaries, assessing risks associated with  this
off-site  migration and preparing conceptual approaches  to
remediating the contamination as a means to reduce risk to accep-
table  levels.  To address the implementation of cost-effective

•Trademark and Service Mark of the Dow Chemical Company
remediation  strategies,  ERT  has licensed a  high efficiency
air/steam stripper known as the AquaDetox™ technology from
Dow Chemical.
  This paper discusses the current regulatory climate governing
groundwater remediation, describes the AquaDetox technology
(including selected performance/economic case studies) and pro-
vides  an overview  of ERT's  approach to offering it  to  the

Regulatory Overview
  In response to the growing concern for groundwater quality,
the U.S. EPA has developed a national groundwater protection
plan which is derived from eight different laws passed by Con-
gress. This plan divides  groundwater into three classifications
based on its use or potential use. There are specific groundwaters
that are irreplaceable drinking water sources and are particularly
vulnerable to contamination; groundwaters that are  currently
used or available for drinking water; and groundwaters that are
not potential sources of drinking  water because they already are
  Under the various statutes, drinking water aquifers will be pro-
tected using  RCRA (and its 1984 amendments) to discourage
pollution from landfills,  the Superfund law to select areas for
urgent cleanup and TSCA to develop  additional restrictions on
the use and storage of potentially threatening chemicals over these
areas. Other laws to be used by the U.S. EPA and state agencies
to control groundwater pollution are: the Clean Water Act under
its construction grants provisions; the Uranium  Mining and Mill
Tailings Reclamation and Control Act (radioactive contamina-
tion); and the Federal Insecticide, Fungicide and Rodenticide Act
(pesticide contamination).  Last,  the  National Environmental
Policy Act's sole source aquifer program and the Safe Drinking
Water Act's provisions for controlling underground  injection will
also be used.
  With regard to the Safe Drinking Water Act, current versions
of the Senate and House both refer to Phase I and Phase II lists of
contaminants which the U.S. EPA published in 1982 and 1983.
The Senate version  of the Act requires the U.S.  EPA to set stan-
dards for 85 contaminants with nine regulations  due within a
year, another 40 within 2 years and the remainder within 3 years.
  In January 1988, the U.S. EPA will be directed to publish a
new list of 25 priority pollutants with a year to propose recom-
mended and maximum contaminant levels and an additional year
to promulgate  final standards. This  process is to repeat  on a
3-year cycle. The House version of the Act addresses 64 con-
taminants and dictates either the promulgation of regulations or a
                                                                                            ON-SITE TREATMENT    133

determination that no rational basis exists to believe the contami-
nant may have adverse health effects. This action must be com-
pleted on 14 volatile organic compounds within a year,  and deci-
sions on  the  remaining  contaminants are due within 3 years.
Beginning in 1988, the bill demands annual priority lists  for possi-
ble regulation, calling for standards within 3 years of publication.
   Under either version of the Act, the U.S. EPA must propose
Filtration  and  disinfection as treatment for drinking water. The
target maximum contaminant levels will be set as close to health-
based recommended maximum levels as possible, considering cost
and available  technology. In both cases,  granular activated car-
bon (GAC) is  suggested, though not mandated, as the "feasible"
technology. Generally speaking, however, GAC is a very expen-
sive treatment technology, and trade organizations such as the
Synthetic Organic Chemical Manufacturer's Association, as part
of a Safe Drinking Water Coalition, have stated that GAC is too
expensive and impractical. It is this same thinking which  led to the
development of the AquaDetox stripping technology by the Dow
Chemical Company and ERT's participation as a Licensee of that

The AquaDetox™ Technology
   The technology was developed by the Dow Chemical Company
for internal use through considerable process research  and  only
recently has it become available for general use outside of Dow
via a licensing program. AquaDetox achieves high stripping effi-
ciency at lower operating and capital costs when compared to
other technologies or unit operations. In many cases, removal of
organic contaminants, including many which have boiling points
in excess of 200°C and are typically considered "not strippable,"
can be achieved by a single, continuous stripping operation.
   The process has a wide range of applications including removal
of 92 of the  111 organic priority pollutants as well  as other
organic materials. It is applicable to the  removal of trace quan-
tities of organics in water  intended for drinking purposes with the
capability of producing waters with organic  concentrations from
jig/1  down to non-detectable levels. The  ability to achieve these
low levels of organics can eliminate the need for a polishing step
using GAC.
   Two types  of AquaDetox units are  available  as shown in
Figures 1 and 2. The air unit is capable of extremely efficient
removal of chlorinated solvents  such as perchloroethylene, tri-
chloroethylene, 1,1,1-trichloroethane, methylene  chloride and
other pollutants  considered  easily strippable from water.  The
steam AquaDetox is used for the removal of low volatility organic
mateials from water including many compounds not considered
strippable by any prior process technology. For example, it has ef-
fectively removed chlorinated phenols (i.e., pentachlorophenol)
and chlorinated  benzenes. The steam unit  also removes easily
strippable pollutants from water when vent gas emissions from the
air process are unacceptable.
                                 VENT TO ATMOSPHERE
        RAW WATER -
                                        «. TREATED WATER
              AIR •
 Primary Advantages
   The benefits afforded by the application of AquaDetox units
 are many. First and  foremost, it is a  fully  developed,  proven
 technology. The Dow Chemical Company has 12 units operating
 at its Michigan Division treating a range of process wastewaters at
 flow rates of 10 to 3,000 gal/min. In addition, four field applica-
 tions of AquaDetox for outside firms have been installed and are
 operating on contaminated groundwater.

                       AQUEOUS CONDENSATE


                                       —	» TREATED WATER
                           Figure 2
                    Steam AquaDetox Unit

  The continuous nature of the technology offers simplicity of
operation  as  compared  to  batch  processes  and  has  low
maintenance requirements and low operating costs (as will be
demonstrated in the following section). The technology provides
maximum flexibility for the sizing and location of units for both
permanent and mobile installations. Finally, its ability to achieve
high-efficiency removal (99 to  9.99%) of organics to achieve at
least /ig/1 levels makes it a viable alternative to GAC, which is ex-
pensive, introduces solids handling equipment and poses disposal
and/or regeneration problems.

  Current  applications of  the  AquaDetox  technology  have
demonstrated its ability to cost-effectively remove organics from
water. Three case studies emphasizing groundwater remediation
are briefly presented to demonstrate this point.

Case Study No. 1
  An air AquaDetox was applied to groundwater contaminated
with 1,1,1-trichloroethane, methylene chloride, trichloroethylene
and  perchloroethylene. The system  was designed to treat a 10
gal/min stream and achieved extremely high removal efficiencies
for all pollutants as shown in Table  1.
  The estimated capital cost for this installation was $80,000 with
an operating cost (including utilities, labor, maintenance,  taxes
and depreciation) of $2.82/1000 gal (or $0.0028/gal). This com-
pared to an estimated GAC operating cost of $20.42/1000 gal for
the same level of pollutant removal.

                          Table 1
               Performance Data: Air AquaDetox
                      Case Study No. 1
                             	Pollutant  levels,  ppm
                             Croundwater	Treated Effluent
                          Figure 1
                     Air AquaDetox Unit

Methylene Chloride



Case Study No. 2
  A steam AquaDetox system was installed to treat groundwater
contaminated with 6,000 mg/1 of a wide variety of organics shown
in Table 2. Typical pollutant  removals from  this 10 gal/min
stream exceeded 99% with several in excess of 99.99%.
  Estimated capital cost for the unit  was  $225,000 with an
operating cost  of $10.00/1000 gal (or $0.010/gal). Estimated
GAC operating costs to achieve these levels of pollutant removal
were $113 to $135/1000 gal of water.

                           Table 2
              Performance Data: Steam AquaDetox
                      Case Study No. 2
Pollutant levels , ppm
Contaminated Treated Percent
Pollutant Groundwater Effluent Removal
Methvlene choride 18
1,1,1-trichloroethane 434
laopropanol 5674
Acetone 772
HER 252
2-butanol 399
MIBK 130
2-hexanol 21
0.0007 99.
0.0019 99.
0.191 99.
2.7 99.
0.01 99.
0.014 99.
0.002 99.
0.059 99.
Case Study No. 3
  An air AquaDetox system was utilized to treat groundwater
contaminated with trace quantities of toluene (22,000 /ig/1) and
benzene (150 /tg/1). The technology achieved an effluent quality
with respect to both pollutants of less than 2 pg/1, the analytical
detection limits.
  This removal was achieved at an estimated operating cost of
$1.35/1000 gal versus an estimated cost for carbon adsorption of
$16.00/1000 gal.

  One of the primary considerations when evaluating the use of
stripper technologies for groundwater contamination is the accep-
tability of the vent gas discharge. In general, this is evaluated on a
case-by-case basis at  the state level, but it is estimated that less
than half-a-dozen states have any specific legislation at this time.
The crux of the matter centers on a cost-benefit analysis where the
trade-offs are between the additional cost to control vent gas
emissions (or avoid them entirely) using GAC and the health risk
to these uncontrolled emissions.
  Since a number of industries already emit volatile organic car-
bon at legal levels much higher than those under consideration for
the groundwater cleanup  situation, the key  question  becomes
whether a reduction of the already low health risk, especially with
the rapid dispersion of pollutants, is worth the price of control-
ling air emissions from groundwater cleanup. To put this issue in-
to perspective, it has been determined in many instances that the
concentrations of pollutants in the vent gas prior to their disper-
sion in the atmosphere do not exceed OSHA workplace standards
for the individual pollutants. Clearly, each application will have
to be reviewed individually with consideration given to the relative
health risk  associated with the lack of control  of such emissions.
  In instances where vent gas control is dictated, the use of GAC
to remove the pollutants from the gas may be considered. Such a
combination of stripping and GAC has been applied on a Super-
fund site in Michigan and has proven cost-competitive with direct
adsorption of the organics from the water.  This competitiveness
has been enhanced by the incorporation of humidity control prior
to the carbon beds since the adsorptive capacity of GAC is re-
duced dramatically when it gets wet.

  As a process wastewater treatment technology, AquaDetox can
be used alone or in combination with GAC or biological treat-
ment systems  to  produce  a  more  cost-effective or  reliable
operating system.
                                                                                                ON-SITE TREATMENT    135

                              Anaerobic  Biological Treatment
                                of  Sanitary  Landfill  Leachate

                                              A.K. Mureebe, P.E.
                                                   D.A. Busch
                                               P.T. Chen, Ph.D.
                                              Wehran Engineering
                                            Middletown, New  York
  The anaerobic treatment of sanitary landfill leachate was in-
vestigated in bench-scale studies at ambient room temperature.
The test runs were conducted in continuous operation for more
than 6 months using the leachate collected directly from the
leachate collection sump installed at a lined landfill site. No  pre-
adjustments of pH or other chemical or physical features were
made. The leachate used in the study showed highly variable
characteristics  with a COD of 2,000 to 35,000 mg/1, BOD5 of
1,800 to 24,000 mg/1 and ammonia nitrogen of 50 to 500 mg/1.
  An anaerobic fixed-flim reactor, with a volume of 22.24 1 and
with a controlled rate of sludge  recycling, was  operated with
upflow feeding at  organic loading between 0.23 and 2.48 kg
COD/mVday and hydraulic detention times of 6.9 to 10.3 days.
  The test resulted in filtered COD removals from 92 to 94%, un-
filtered BOD5 removals  of 90-94% and ammonia removal from
40 to 50%.  Despite the  wide range of loading conditions, the
system showed a high level of stability in terms of COD removal
efficiency which hovered about 93%. With influent COD of 1,700
to 17,000 mg/1, the effluent leaving the system was in the range of
100 to 1,300 mg/1, depending on the organic loading rate. The ef-
fluent can be further purified by an aerobic biotreatment unit to
reduce the COD and BOD5 levels below 250 and 35 mg/1, respec-

  The landfill is located  off Route 17M east of the Wallkill River
in the Town of Goshen,  Orange County, New York (Fig. 1).  The
site was  designed  to protect  surface and groundwater from
leachate contamination by utilizing the following technologies:

• Low permeability on-site glacial till material used as soil liner
  to minimize  the impact of leachate  on adjacent surface  and
• A sloped landfill  bottom to direct leachate toward a collection
• A leachate collection piping system to intercept and collect the
• A collection  well  with level-controlled pumps that transfer the
  leachate collected in the landfill to a geomembrane-lined pond,
  thus maintaining the hydraulic head in the landfill at a low level
  The leachate management issues currently facing  the landfill
are the  establishment  of a feasible and economically  sound
disposal  scheme that will meet the needs of the current operation
and future closure of the landfill, and the acquisition of permits
from state  and/or  local agencies for  disposal of the leachate
                         Figure 1
                Location Map — Turi Landfill
  Wehran Engineering has been retained to conduct a study that
will provide a viable system for leachate treatment and to furnish
the necessary framework for implementation.
  This paper includes a detailed investigation of all relevant
features and data pertaining to leachate treatment. Specifically,
the scope included the following:

• Analysis of leachate characteristics to determine the quality and
  quantity of leachate generated by the site
• Performance of bench-scale treatability tests to confirm the
  technical  feasibility of on-site treatment options
• Development of engineering design criteria

Quantity and Characteristics of Leachate
  An extensive monitoring, sampling and analysis program was
conducted for a period of 1 year at Turi Landfill to determine the
quantity and quality of leachate generated from the operational
cell of the landfill. Weekly and monthly samples were collected
during this period and analyzed in the laboratory.
  The weekly samples were analyzed for the conventional pollu-
tant parameters of chemical oxygen demand (COD), 5-day bio-

chemical oxygen demand (BOD5),  total solids (TS),  ash, total
Kjeldahl nitrogen (TKN), ammonia nitrogen  (NH3-N),  hard-
ness, conductivity and pH.
  The monthly samples were analyzed for iron (Fe),  zinc (Zn),
copper (Cu), cadmium (Cd), lead (Pb), manganese (Mn) and total
phosphate (T-PO4).

Leachate Flow
  The flow of leachate generated in the active landfill operation is
shown in Table 1. The actual monthly average daily leachate flow,
measured in the  operating landfill  area for the  period between
April  1984 and March 1985, ranged from 100 to  18,000 gal/day,
with an  annual average of 6,300 gal/day. The variation in flow
probably is due to  the  absorptive characteristics of the solid

Characteristics of Leachate
   Starting in November 1984, sampling and analysis of the con-
ventional pollutant parameters described above were  performed
as part of the study program. In conjunction with this sampling
and analysis, leachate flows were recorded weekly.
   With the exception of pH, the pattern of changes in the concen-
trations  of  these pollutant parameters appears to  be  in general
correlation with the changes in leachate flows.  The concentration
changes are inversely proportional to the leachate  flow rates
                           Table 2
             Leachate Characteristics - Turi Landfill
Am monia Nitrogen
Organic Nitrogen
Total Solids
180- 36,900 mg/1
200 -45,800 mg/1
5 - 640 m g/1
1 - 140 mg/1
1,005-25,330 mg/1
805- 13,580 mg/1
345 - 9,460 mg/1
900 - 15,500 u mhos/cm
5.30 - 6.00
12,042 mg/1
19,527 mg/1
340 mg/1
46 mg/1
14,367 mg/1
7,037 mg/1
5,647 mg/1
9,360 u mhos/cm
  The monthly results for the metals and total phosphate analyses
are summarized in Table 3. These concentrations are comparable
to reported ranges for sanitary  landfills.4  All  concentrations
observed in this study are  below reported  threshold limits for
biological assimilation in a wastewater treatment facility.''4'8

                          Table 3
              Monthly Analysis of the Concentration
              of Metallic Elements and Phosphates
wmcn may DC aunouiea 10 me percoiaiion or ramiau. i ne exier-
nal water percolation in the landfill will thus act as a dilutant in
carrying leachables out of the landfill bed.
Table 1
Leachate Generation Rates
Measured Values Estimated Values (5)
Ppt(l) Perc (2) Ratio Flow (4) Ppt Perc Ratio Flow
Month (inch) (inch) (3) (gpd) (inch) (inch) (3) (qpd)
April 84 4.65 1.080 0.23 12,840 3.86 2.12 0.55 25,135
May 10.03 — -- — 4.22 1.48 0.35 17,548
June 3.95 0.270 0.07 3,212 3.86 1.35 0.35 16,006
July 6.27 0.430 0.07 4,920 4.01 1.41 0.35 16,717
Aug. 2.61 0.290 0.11 3,367 3.39 1.19 0.35 14,109
Sept. 1.18 -- - - 3.01 1.06 0.35 12,567
Oct. 2.88 — — — 3.03 1.67 0.55 19,800
Nov. 2.58 0.008 0.03 92 2.88 1.59 0.55 18,852
Dec. 2.99 0.540 0.18 6,139 2.46 2.09 0.85 24,780
Jan. 85 1.06 0.127 0.12 1,463 2.46 1.23 0.50 14,584
Feb. 1.99 1.430 0.72 18,192 2.68 1.34 0.50 15,888
March 2.21 0.516 0.23 5,916 3.88 3.30 0.85 39,127
Annual 42.4 4.7+ — -- 39.74 19.83
Average .19 6,239 .50 19,584

(1) Precipitation (ppt) from climatological data, Middletown Station, NY, NOAA.
(2) Calculated percolation (perc) through cover soil and underlying solid wastes based on actually
measured leachate flow within 13.1 acres of operating landfill.
(3) Ratio of perc to ppt.
(4) Actual leachate flow measured.
(5) Calculated leachate flow, based on the estimated values of percolation for 13.1 -acre area.
Elements 10/25/84 12/20/84 1/24/85 2/27/85 4/4/85
Fe 770.00 420.00 1,100.00 1,140.00 840.00
Cu 0.05 0.08 0.05 0.05 .35
Pb 0.05 0.05 0.16 0.05 .14
Mn 190.00 41.00 310.00 200.00 200.00
In 11.00 14.00 33.00 6.30 11.00
Cd 0.01 0.01 0.01 0.01 .01
T-P04 8.40 1.90 3.40 2.50 6.50
Note: All units are in mg/1

The concentrations of priority pollutants which were present in
the leachate at levels above analytical detection limits are given in
Table 4. All organic chemicals on the list are reported to be effec-
tively removed by conventional biological treatment processes at
both POTWs and industrial wastewater treatment facilities.'

Leachate Treatability Evaluation
A review of the literature reveals that leachate may be treated
by aerobic1 '2'3'4'6'9 or anaerobic biological processes7'8, physical
methods and chemical means.4-9 Existing reports discuss mainly
laboratory-scale studies.
Bench-Scale Treatability Tests
In the selection of systems for a treatability study, considera-
tion was given to the processes available at local POTWs, in-
dustrial wastewater treatment facilities and for the purchase of
packaged treatment systems for possible on-site use. The follow-
ing process techniques were chosen for the bench-scale treatability

• Aerobic bio-oxidation system utilizing local POTW activated
sludge for the evaluation of maximum ratio of leachate to
  The range and average concentrations  for  the conventional
pollutant parameters during the study's test period are given in
Table 2. The  ranges are comparable with reported  ranges for
sanitary landfills published in technical literature.4'7
   sewage flow
   Anaerobic biological system for treating raw leachate
   Aerobic treatment for polishing the effluent of the anaerobic
   Physical/chemical treatment for removing solids or COD as a
   pre-treatment or post-treatment process
                                                                                 TREATMENT OF HAZARDOUS WASTES     137

                             Table 4
            Concentrations of Detectable Priority Pollutants
                      in Turi Landfill Leachatc
       Dichlorodifluoro methane
       1,2- Dichloro propane
       Methylene Chloride
       Vinyl Chloride
       Diethyl Phthalate
       Total Cyanide
       Total Phenolics

   The  bench-scale  treatability  tests were performed using the
 following equipment:
 • For anaerobic biological system: A fixed film upflow reactor,
   6 in. diameter  x  5  ft high with hydraulic  detention time of
   7 days, was used  in a plug flow mode with recycle (Fig. 2).
 • For aerobic bio-oxidation  system: A 4-1 reactor with air dif-
   fusers was used  in  a completely mixed mode.  Sludge  was
   wasted directly from the unit or recycled from a settling  unit
   to maintain the desired MLVSS under aeration (Fig. 3).
 • For physical/chemical test: A multiple stirrer,  Phipps-Bird
   stirring apparatus was used to test the effect  of coagulation/
   precipitation with lime, caustic and/or alum.

   In this test, activated sludge  seed  from the Harriman Waste-
 water Treatment Facility was aerated in a bio-oxidation unit. The
 system  was  run  to  a steady state MLVSS with a sewage  and
 nutrient feed prior  to the addition  of raw leachate. Once the
 bench-scale activated sludge system achieved this steady state, raw
 leachate was added in ratios of raw leachate to sewage of 1 to  5%.
   The results of this study are given in Table 5. At dosage ratios
 of raw leachate to raw sewage greater than 2%, the aerobic treat-
 ment system is adversely impacted. The system failed to maintain
 a steady state MLVSS at a 5%  dosage, and higher order organ-
 isms were observed. At the high dosage levels,  increased phos-
 phate addition on the order of 1,000 mg/1 was necessary for ade-
 quate cell development.

r 	 "

\ /

i — >>
"^x j
e ai 3 H <|
H ,

I •'» i a' H

-| 1 	
10 1





                   FEED    FEED     31UDOE   EFFLUENT
                   PUMP    TANK     TANK     TANK
                             Figure 2
     Bench-Scale Anaerobic Fixed Film Biological Treatment System
                                                    TRAP BOTTLE
                                                 Figure 3
                                    Aerobic Bio-Oxidation Treatment Unit
                                                  Table S
                                      POTW Tolerance S(ud> of Aerobic
                                  Bio-Oxidation System Treatment of Leachate
                                   Raw Leichatc,
                                    Raw Sewage
                       •Ai 5°'o ihc system coult) not aliam u steady stale Ml VSS
                                              Average COD
                       ANAEROBIC TREATMENT
                         The  parameters monitored in studying the efficiency of the
                       anaerobic biological treatment  system were  COD,  BOD5 and
                       TVA. The total volatile acid removal (TVA) is an important step
                       in methane fermentation and is therefore included. The results for
                       the removal of these parameters are  summarized in Table 6.
                         In evaluating the anaerobic treatment of landfill leachate, the
                       removal of COD and BOD are the most important parameters to
                       study. As can be seen from the data in Table 6. the efficiency of

                                                          Table 6
                                          Anaerobic Test Results of Leachate Treatment


X Removal

COO (rag/1)

X Removal
(rag acetate/1)
% Removal
              •TVA is total volatile acids.
the removal of these two parameters is quite constant. For Run
No. 1, which had an organic loading rate of 0.23  kg COD/m3
• day, the COD removal was 94%, while with an organic loading
rate of more than 10 times that in Run No. 1, the COD removal
rate (Run No. 5) was 92.5%. The system showed an ability to
operate efficiently under a wide range of COD loading condi-
tions, which is an important characteristic in the treatment of
landfill leachate. In conjunction with the COD removal, the study
of two other system paramters is important: methane and sludge

Methane Production
  One of the major advantages of the anaerobic process is that
organic material in  the wastewater is converted to methane gas.
The methane  gas  evolution rates  and composition for  the
anaerobic treatability study are summarized in Table 7.

                          Table 7
     Gas Production Data for Anaerobic Treatment of Leachate
Gas Evolution
Methane Rate
(at STP)
   As can be seen, methane production rates determined from ex-
 perimental removal of organics are higher than theoretical rates.
 This may be a result of average daily influent conditions actually
 being higher than those detected in the samples analyzed;  this
 would mean that the organics removal rate is higher than the  rate
 used in the calculations. An important finding of this portion of
 the study is that the methane production rate was directly propor-
 tional to the organics  loading rate.

 Solids Production
   Another advantage  of the anaerobic system is the low solids
 production rate. Sludge production from anaerobic biological
 treatment systems  typically ranges from 0.04 to 0.10 g VSS/g
 COD removed. The sludge  production rates as a function of
 organics removal are  summarized in Table 9.  The amounts of
 solids produced were estimated from the increase in volatile solids
 concentrations in the circulation stream.

                          Table 9
   Anaerobic Sludge Production in Treatment of Landfill  Leachate
                                                                    Run     (g/day)

 (g solid/g

  To determine the efficiency of the methane fermentation, the
experimental  gas production  rate  can be  compared to the
theoretical value which is 0.35 1 CH4/g of organics removed. The
methane production rates are shown as a function of the removal
of BOD, COD and TVA in Table 8.
                          Table 8
        Methane Production Rates for Anaerobic Treatment
                     of Landfill Leachate
                 Gas Rate as a Function of Organic Removal
Liters CH4/
g COD removed
Liters CH4/
g BOD removed
Liters CH4/
g TVA removed
  The sludge production of  the  anaerobic  digestion of  the
leachate was slightly higher than normal anaerobic sludge produc-
tion, yet the amount of sludge was still significantly lower than
the amount of sludge produced by a conventional activated sludge

  The aerobic  treatment  of the effluent from  the anaerobic
system was conducted in a fill and draw manner. Anaerobic ef-
fluent was added to activated sludge from the Harriman Sewage
Treatment  Plant, and the system  was aerated.  Samples were
drawn from the system at 12,  24 and 48 hrs  for analysis. The
characteristics of these samples were compared to the anaerobic
effluent that  was added initially. The test was conducted with
COD and BOD values of 1,300 and 200 mg/1. The MLSS, MLVS
and SVI values were 8,100, 3,950 and 135, respectively.
                                                                               TREATMENT OF HAZARDOUS WASTES    139

                         Table 10
    Results of Aerobic Treatment of Anaerobic System Effluent





  The removal of the conventional pollutants (COD, BOD and
NH3-N) was monitored to determine the efficiency of the pro-
cess. The results of these tests are summarized in Table 10.

Removal of BOD and COD
  In discussing the removal of organics by the aerobic system, an
important parameter to consider is the ratio COD/BOD5.
  The general trend is an increase of the COD/BOD ratio with in-
creased  aeration. This  increase  is due  to a rise  in  the non-
biodegradable portion of the leachate over time. The test  for
chemical oxygen demand will detect some organics which are non-
biodegradable and thus will not be measured by the BOD test. As
the aeration time increases,  the biodegradable organics will  be
removed by the system microorganisms, and the non-biodegrad-
able organics will represent an increasing proportion of the re-
maining organics.
  Based on the results of this test, we concluded that a removal
rate of 80% of the COD  and 95% of the BOD is attainable by the
conventional aerobic treatment systems.

Removal of Ammonia
  The  aerobic  system  removes  ammonia by  the  process  of
nitrification; the ammonia is oxidized to  nitrate. The bacteria
responsible for nitrification have a growth rate slower than that of
the bacteria responsible for the removal of organics. Thus, it is ex-
pected that a longer aeration time will be required for the removal
of NH3-N than that for BOD. As  can be seen from the  test
results, the removal of BOD was as high as 90% in only 12 hr,
while an aeration time of 48 hr was needed to remove 96% of the
average sludge production rate of 0.68  g  VSS per g  COD re-
moved. This increased sludge production is attributed to the ex-
tended aeration time and the lack of sludge wasting in the test to
maintain  a  steady state MLVSS. The results do indicate that
under typical conditions  for aerobic biological  treatment, the
sludge production from treatment of the leachate would  be within
the expected range for such systems.

  The effect of chemical coagulation and precipitation on COD
and suspended solids removal also was evaluated in this study.
The test conditions for the study are summarized  in Table 12. In
this test, the samples were dosed with the coagulating chemicals,
rapidly mixed  using a  Phipps-Bird  stirring apparatus  at  50
revs/min  for 30 sec, flocculated at 15 revs/min for 20 min and
then allowed to settle  for 30 min. The supernatants then were
sampled and analyzed for COD and TSS.
                          Table 12
               Chemical Coagulation/Precipitation
                Treatment Operating Conditions
 Waste Stream

Raw Leachate
Anaerobic Effluent
Aerobic Effluent

1.  800-4200
2.  2800
3.  --

1.  1800
2.  --
3.  --
4.  --
                               Cnemical Coagulant Dosages


   The  results  for  the  chemical  coagulation/precipitation
treatability studies on the raw leachate, anaerobic filter effluent
and aerobic polishing unit effluent are presented in Tables 13, 14
and 15, respectively.
Solids Production
  The production of sludge is an important parameter in the design
of the aerobic process. Table 11 contains a summary of the solids
production as a function of the amount of organics removed.
  The reported range of sludge production for activated sludge is
0.25 to 04 g of VSS formed per gram  of COD removed. The
results of  the  bench-scale aerobic polishing  unit indicate  an
                         Table 11
                Aerobic Sludge Production for
            Treatment of Anaerobic System Effluent
(in 9/1)




                          Table 13
           Results of Chemical Coagulation/Precipitation
                  Treatment of Raw Leachate
Chemical Coagulant
Dosages (ng/1)





.. .


                          Table 14
      Results of Chemical Coagulation/Precipitation Treatment
                  of Anaerobic Filter Effluent
                            Table 16
              Operating Conditions and Performance of
                   Bench-Scale Anaerobic Filter


Dosages (mg/1)









P ara m eters




Operating Conditions Removals
Oet. Time

Run 1 7.7
Run 2 7.3
Run 3 10.3
Run 4 10.3
Run 5 6.8

9 V5/
kg COO/ g COD rem.
m 3 day "ercent Solids
Org. Loads COD rem.

0.216 94.0
0.419 94.6
1.302 93.7
2.100 94.0
2.380 92.0







* liters of methane produced/g COD removed


Table 17
of Performance of the

Four Treatment Systems






in Removal of


                          Table 15
      Results of Chemical Coagulation/Precipitation Treatment
               of Aerobic Polishing Unit Effluent



Removal of COD
Raw Leachate
  Of the three tests run, only the addition of caustic and alum
showed any significant reduction in COD, and this was less than
20%. The addition of lime only, and lime and alum showed no
appreciable  decrease in COD from pH elevation or coagulation.
The results of the third run indicate that the  best COD  removal
was obtained with caustic alone, and that alum was ineffective as
a coagulating agent for COD removal.
  The results  obtained  in this study indicate  that  chemical
coagulation and precipitation processes do not appear to be ade-
quate as a pretreatment process for disposal of the leachate either
off-site at a POTW or through an on-site treatment system.

Anaerobic Filter Effluent
  The results of the chemical coagulation and  precipitation of the
anaerobic filter effluent indicate that large dosages of alum are re-
quired to obtain moderate removal efficiencies. The results of the
second and third runs indicate that a dosage of 333 mg/1 of alum
results in a 10-13% removal of COD. By increasing the dosage to
 Anaerobic  system

 Aerobic system

 Alum coagulation

 Integrated efficiency
  94                  49

  80                  95

	25                  --_

99.1                97.5
 1,500 mg/1 in Run No. 4, a removal of only 17.5% is attained.
 These results indicate that an increase in the alum dosage of 500%
 results in only a 5% increase in COD removal. This process is,
 therefore,  not considered effective  for reducing residual COD
 from the anaerobic filter effluent.

 Aerobic Polishing Unit Effluent
   The results presented in Table 15 indicate that very high levels
 of alum (1,500-2,000 mg/1) will remove only 25-30% of the COD
 in the aerobic polishing unit's effluent. This process, therefore, is
 not considered effective  for removing any residual COD in the
 aerobic polishing unit, should such removal be essential to meet
 discharge standards.

 Removal of Suspended Solids
 Raw Leachate
   The test results presented in Table 13  indicate that effective
 suspended solids removal in the raw leachate is achieved only with
 very high dosages of lime (3,330-4,170 mg/1). Such high dosages
 generate proportionally high sludge volumes which would require
 further processing for proper disposal.
   Based on these results, we concluded that this process is not ef-
 fective as a  pretreatment process for disposal of the leachate
 either off-site at a POTW or through an on-site treatment system.

Anaerobic Filter Effluent
   The results presented in Table 14 indicate that effective removal
 of suspended solids in the anaerobic filter effluent is achieved
 only with high levels of alum (2,000 mg/1). At lower dosages
(33-83  mg/1), moderate removals were attained,  indicating that
this process may be viable for solids removal in conjunction with
a  clarification system,  should  subsequent  biological  or
physical/chemical treatment for residual organics and ammonia-
nitrogen require pretreatment for solids removal.
                                                                                 TREATMENT OF HAZARDOUS WASTES    141

 Aerobic Polishing Unit Effluent
    The results presented in Table 15 indicate lhat relatively high
 levels of alum (1,000 mg/1) effectively reduce the suspended solids
 level in  the  aerobic  polishing  unit's  effluent  to  acceptable
 discharge standards (20 mg/1). Full-scale clarification units, with
 tube or plate settlers, may reduce the chemical dosage required to
 achieve this effluent quality. This process is considered viable for
 consideration  as  a polishing treatment  for  suspended  solids

    The laboratory bench scale  study using an anaerobic fixed-film
 reactor   for   the  treatment of   sanitary  landfill  leachatc
 demonstrated the amenability of the process and its  application
  for the stabilization of leachate.
    The landfill raw leachate was rapidly adopted  by the sludge
 used in  the anaerobic fermentation system. Hydraulic detention
 times as short as 7 days and organic loads  up to 2.2 kg COD/m'
 day resulted in a 94% removal of COD.  The gas produced in the
 system was composed of 60-70%  combustible material. Based on
 the rate of gas production, we estimated that more than 99%  of
 the COD removed was converted to CH4 gas.
    The process performance with respect to  COD removal and gas
 production at  various  loading conditions is shown in Table 16.
 The removal of COD and the  rate of solids production were very
 steady.  The  response of process  performance  to changes  in
 organic  or hydraulic  loads was reflected in the rates of methane
 production;  however, the  efficiency of COD  removal remained
 unaffected for the conditions  tested.
    The performance of the activated sludge system for treating the
 effluent  from the anaerobic filter was much better than anaerobic
 treatment alone. With a loading  of 0.90 g  COD/g MLVSS, the
 system attained removals of 80 and 95%  for COD and NH, - N,
 respectively, within 48 hr of aeration. The estimated rate of sludge
 formation in the aerobic system was 0.41  g/VS/g COD removed.
  The series of tests conducted  in the bench-scale treatability
study have confirmed that the landfill leachate can be treated with
the use of conventional biological and physical/chemical pro-
cesses. The effluent qualities achievable are comparable to the ef-
fluent  of wastewater treatment plants with secondary treatment
facilities.  Since  the  processes  tested   are  all  conventional
technology,  the equipment  required  for use is  commercially
  Based on  the results  of each test included  in this treatability
study,  a combination of anaerobic and aerobic treatment in series
alum coagulation could  achieve more than 99% removal of COD
from raw leachate.

 I. Boyle. W.C. and Ham, R.K., "Biological Trealabilily of  Landfill
   Lcachatc." JWPCI. 46. 1974, 860.
 2. Cook, O.M. and force, G.G., "Aerobic Biostabilization of Sanitary
   Landfill Leachate," JWPCI.  46,  1974, 380.
 3. Uloih, V.C. and Max ink, "Aerobic Biotreatment of a High Strength
   Lcachalc," J. Environ.  EnK. Dt\  . ASCE. EE4, 1977, 841.
 4. Chain, E.S.K. and DcWalle,  F.B.. "Evaluation of  Leachate Treat-
   ment." U.S EPA Report, Washington, D.C.
 5. Zapf-Gilzc,  R.. "Temperature Effects on Biostabilization of Leach-
   ate," J. Environ. Eng. Di\'.. ASCE. EE4, 1981. 653.
 6. Ehrig, H.J. and  Slegcman.  R..  "Treatment of Sanitary  Landfill
   Leachate:  Biological Treatment." Waste.
 7. Force, E.G. and Reid, V.H.,  "Anaerobic Biological Stabilization of
   Sanitary Landfill Leachatc," Technical Report URY TR 65-73-CG17,
 8. Wu, V.C.  and Kennedy. J.C.,  "Anaerobic  Treatment of Landfill
   Leachate by an Upfloxx Two-Stage Biological Filter," Proc. Fixed-
   Film Bioiechnol. Seminar, 1984.
 9. Ho, S.,  Boyle. W.C.  and Ham. R.K..  "Chemical Treatment  of
   Leachate from Sanitary Landfill." JWPCF. 46. 1974, 1776.

                            On-Site  Versus  Off-Site Treatment
               of Contaminated  Groundwater —  An Evaluation
                            of Technical  Feasibility and Costs
                                              Kent L. Bainbridge
                                          Daniel  W. Rothman, P.E.
                                              URS Company,  Inc.
                                              Buffalo, New York
  The Pollution Abatement Services, Inc. (PAS) site is an aban-
doned hazardous waste incinerator site  located in the City of
Oswego,  New York. Groundwater contamination has  resulted
due to leakage from a  surface impoundment, buried tanks and
drums—both above and below ground.  The principal  ground-
water contaminants include volatile organic compounds, organic
acids and other low molecular weight organics, and nickel, which
also appears to be organically bound.
  Groundwater recovery and leachate collection  were included in
the site remediation plan. Alternatives considered  for treatment
of the groundwater/leachate were: (1) hauling to an off-site com-
mercial treatment facility, (2) conveyance to the local POTW and
(3) on-site treatment. The relative merits of these alternatives were
evaluated on the basis of the  degree of treatment technically
achievable and the relative  cost.
  Critical to the evaluation was the identification and demonstra-
tion of technology capability for treatment of the groundwater/
leachate. Initial laboratory  treatability studies demonstrated that
biological treatment followed by carbon adsorption was required
for cost-effective treatment, while carbon adsorption alone was
not effective. Further studies demonstrated the feasibility of the
treatment system on an intermittent basis—a necessary require-
ment based on the expected low rate of groundwater/leachate
  In this analysis, the costs and ultimate selection of a  ground-
water/leachate treatment alternative were sensitive to several fac-
tors including: (1) the wastewater character, (2) the wastewater
volume and rate of recovery and (3) the degree  of treatment re-
quired. On-site treatment was  offered a  higher  degree of treat-
ment than would be achievable at the local POTW and a lower
cost than would be expended for hauling to an off-site commer-
cial facility.

  The Pollution Abatement Services, Inc. (PAS) site, located
near the eastern limit of the  City of Oswego, New York (Fig. 1), is
an abandoned hazardous waste facility which is listed among the
top 10 priority sites on  the  U.S. EPA's National Priorities List.
During the period from 1982-84, URS Company, Inc. was con-
tracted by the New York  State Department of Environmental
Conservation (NYSDEC to perform a Remedial Investigation/
Feasibility Study (RI/FS) of the site. One result of the field in-
vestigation was the determination that the groundwatei underly-
ing the site and leachate taken from two on-site drainage ditches
was highly contaminated by organic priority pollutants.
  Hydrogeological data collected during the RI/FS supported the
conclusion that all subsurface contaminant  migration was being
                        Figure 1
            Location of the Pollution Abatement Site
intercepted by two creeks which bounded the site on three sides.
Based on these findings, the RI/FS report recommended specific
remedial measures for subsurface cleanup and closure of the PAS
site. These measures included construction of a leachate collec-
tion system, installation of a groundwater recovery system and
treatment of the groundwater/leachate.1
  As part of the subsequent Remedial Design Study, laboratory
treatability studies  were  conducted to  identify  process  re-
quirements to  determine the level of treatment attainable and to
develop design and cost data. Results of the treatability studies
are reviewed in this paper. Estimates of the cost of on-site treat-
ment also are presented. Finally, the on-site and off-site treatment
methods are compared to identify the most cost-effective treat-
ment alternative.
                                                                           TREATMENT OF HAZARDOUS WASTES    143

  The facility known as Pollution Abatement Services,  Inc.
(PAS) was constructed and put into operation in 1969-70. A high
temperature, liquid chemical waste incinerator was the principal
unit installed by the owners. Throughout its active life, PAS ex-
perienced  continuous operating  problems, numerous air and
water quality violations and mounting public opposition.
  During its operating period (from 1970 through 1977), large
numbers of drums containing various chemical wastes were col-
lected and stored on-site. Liquid wastes also were  stored  in
  Beginning in 1973, a series of incidents including liquid waste
spills and overflowing of lagoon wastes into the adjacent White
Creek led to the involvement of the U.S. Coast Guard, the U.S.
EPA and  the  New York State  Department of Environmental
Conservation  (NYSDEC). This involvement included a number
of limited and temporary remedial actions during the period from
1973-1976. In 1977, PAS was abandoned.
  During the several years immediately following abandonment,
a number of emergency remedial  actions and preliminary in-
vestigations were undertaken at the PAS site including draining
and filling the waste lagoons and removing approximately 3,000
leaking barrels and several waste storage tanks from the site. Late
in 1981, contract documents for the surficial cleanup of PAS were
prepared,  including  the  demolition  and disposal  of on-site
facilities,  the  removal of the  remaining (approximately  8,000)
drums from the site and the drainage and disposal of approx-
imately 80,000 gal of liquid chemical waste from 10 bulk storage
tanks. This surficial cleanup was completed in October 1982.
  In November 1982,  URS Company, Inc. (URS) entered into a
contract with NYSDEC to evaluate alternative remedial measures
for final cleanup of buried wastes, contaminated soil and ground-
water and site closure.

  Alternatives considered for  treatment of the  contaminated
groundwater/leachate were: (1) hauling to an off-site commercial
treatment facility, (2) treatment at a local POTW and (3) on-site
treatment. Two commercial hazardous waste disposal facilities
are  located approximately 180  miles from the PAS site.  The
Oswego East Sewage  Treatment Plant,  which provides primary
and secondary treatment, is located just 1 mile from the site.
  A preliminary estimate indicated that on-site treatment would
be less costly than off-site disposal at a commercial facility. For
this reason, laboratory treatability studies were conducted: (1) to
determine  treatment process  requirements, (2) to measure the
degree of treatment achievable and (3) to develop design and cost

  Available  hydrogeological   data indicated  that  the  strata
underlying the PAS site would be low yielding and, on this basis,
it was estimated that groundwater pumping would  yield only ap-
proximately 24,000 gal per each  1-3 week period. Since a 24,000
gal basin  remains on the site and could be used for pumped
groundwater storage, it was anticipated that the treatment system
would be operated on an intermittent basis with treatment being
performed whenever  the  storage basin was  full. A chemical
analysis of this contaminated groundwater is given in Table 1.
  As  part  of the  PAS  remedial design project,  laboratory
wastewater treatability  studies  were conducted  to  determine
specific treatment process requirements and  capabilities. These
studies have been described previously.'  The  key findings are
summarized below:

• Only a minor fraction ( <1%) of the wastewater total organic
  carbon (TOC) was removed by air stripping. This result indi-
                          Table 1
                    Wutewaler Analysis'
    HetK,W CMoridf
    trjns-I ,?-Olchloroeth«nr
    4-Methyl -?-Penunone
    fot«l I/lenet
    2,4-OI«thylphenol        ,
    Tentatively Identified luo/l)
    [thy] Bentene
    ?-3-Olethyl Oilrine
    Hrtn/l Sen/me
    N.N-Olethyl Forvumldr
    Bulinolc Add
    l-Hethoiy  Cthinol
    f thylbenjene
    I,4-Olmethyl Benrene
    Pennnolc  Acid
    2-tthyl Butinolc Acid
    N.N-Dlnelhyl Benien««lne
    ?-fthyl Heitnolc Acid
    2.3-OlMlhyl Phenol
    Beniene Acetic Acid
    3-Mrlhyl Beniolc Acid
    Convention!! PjrMgte
    «U«llnlty  -
    H«rdnesi K« )

    I .WO

    5. POO

 Samples analyzed were 50:30 mixtures of kachate.
 A complete analysis of the US EPA Contract Lab Protocol Hazardous Substances was per-
 formed. Only the hazardous substances found at concentrations above analytical detection
 limits are reported in this table In addition, the ten nonhazardous lubstance list compounds
 present in each fraction (volatile, acid and base.'neutraf) at greatest concentration were tenta-
 tively identified and quantified. These compounds are also reported in this table.
 All concentrations in jig/ unless otherwise indicated.
  cated that very little of the TOC consisted of volatile organic
  contaminants.  Therefore, air stripping was determined to be
• Nickel was not removed by hydroxide precipitation but was re-
  moved by activated carbon adsorption. From these data, it was
  concluded that nickel was organically bound or complexed.
• Adsorption of the organic contaminants by activated carbon
  was enhanced at acidic pH values.
• Wastewater COD was reduced  from a concentration of 4,000
  mg/1  to approximately 450 mg/1 (89°7o removal) by carbon
  adsorption at pH 3.
• Carbon requirements per 24,000 gal batch of wastewater were
  computed to be  6,030 Ib based on an effluent quality  of 800
  mg/1 COD (i.e., 20% breakthrough of the influent COD con-
  centration). This usage is equivalent  to a carbon loading of
  0.106 Ib COD/lb activated carbon.
• With a period of 1 week acclimation,  biotreatment achieved a
  COD removal of 85%. Rate studies showed that  this level of
  treatment could be achieved in less than 8 hr retention time. A
  final effluent quality of less than 50 mg/1 COD was achieved by
  treating the bioreactor effluent with activated carbon.
  On the  basis of the batch bioreactor studies, we concluded that
a treatment process consisting of biological treatment followed by
activated  carbon  polishing  was feasible, would be more cost-
effective and would produce a much higher quality effluent than
carbon adsorption alone. For these reasons,  further laboratory
studies were subsequently performed to  develop the  necessary
design and cost data for a batch biological process to be operated
intermittently. The results  of  these studies, which have been
reported previously,' are summarized below.

Pretreatment Requirements
• The PAS waste has no toxic or inhibitory effect on biological
  treatment.  Consequently,  pretreatment of the wastewater to
  remove substances (such as iron) potentially inhibitory to the
  biological treatment process is unnecessary.
• Approximately 8 hr of aeration were required to achieve opti-
  mum COD removal at 25 °C.
• No difference in bioreactor performance was observed when
  wastewater feed volumes in the range of 10-70% of the total
  reactor volume were used.
Nutrient Requirements
• The addition of nutrients  (i.e., nitrogen and phosphorous) to
  the wastewater is not required to achieve optimum biological
  removal of COD.
• The addition of nitrogen and  phosphorous actually resulted in
  a higher total suspended solids concentration.
Effect of Extended Lags Between
Bioreactor Operation
• Little difference was observed in the COD removal efficiency
  of  the  batch bioreactors when operated  with idle periods of
  1-3 weeks.
• The rate of COD removal is inversely proportional to the length
  of the idle period.
• The cycle time required to achieve optimum COD removal in-
  creased from 8-11 hr when  the bioreactor was operated at
  1 week idle periods, to 25 hr when the idle period was 3 weeks.
Effects of Temperature on Bioreactor Performance
• During the initial  weeks of bioreactor operation, the effluent
  COD was inversely proportional to the bioreactor temperature.
• After several weeks of intermittent operation (1 week idle per-
  iod), the effluent COD level achieved at the different tempera-
  tures (i.e., 25 °C,  15 °C and 5°C) began to converge.
• Bioreactor performance as measured in terms of COD removal
  rate and efficiency is  adversely impacted by cooler tempera-
  tures. The exact temperature at which performance begins to
  degrade is not known, but lies somewhere between 5°C and
  15 °C.
Effects of Powdered Activated  Carbon
Addition to the Bioreactors
• Powdered  activated carbon (PAC) addition substantially en-
  hances COD removal rates under conditions of extended bio-
  reactor idle periods and low temperature.
• COD removal efficiency increases as the PAC dosage is in-
• The time of  PAC addition during the reactor cycle does not
  affect the effluent  quality  achieved.
• When PAC is added to the bioreactor, the COD removal me-
  chanism involves simultaneous adsorption and biodegradation.
Oxygen Supply Requirements
• Oxygen uptake rates are greatest in the initial phases of the
  reactor cycle  and decrease rapidly with time.
• As  the length of the idle period between bioreactor react per-
  iods is increased, the rate  of  oxygen uptake during the initial
  phase of the react  cycle decreases.
Sludge Production
• The net biological  sludge production in sequenced batch bio-
  reactors operated at 25-50% fill ratios was minimal or negative.
• Total bioreactor sludge production was found primarily to be
  a result of inorganic  precipitates.  This production  averaged
  800 mg/1 solids of wastewater treated in bioreactors operated
  at 50% fill, nonpretreated  and 1 week idle periods.
• The use of PAC  in the bioreactor increased the amount of
  sludge produced in direct proportion to the PAS dosage.
Activated Carbon Column Studies
• Results of batch adsorption  isotherm  studies indicated very
  little difference in adsorption characteristics at pH 3 and 7.5.
  Based  on this result, no pH adjustment of bioreactor effluent
  is necessary to achieve optimum COD removal by carbon col-
  umn polishing.
• The carbon inventory  requirement for a granular  activated
  carbon column is a function of the allowable breakthrough
  level. Effluent from a reactor operated at a 50% fill and 1 week
  idle period was polished to determine carbon requirements.
  For breakthrough levels of 20, 30 and 40% of the carbon col-
  umn influent, carbon requirements  of 2,670, 1,460 and 700 lb/
  24,000 gal of wastewater treated were calculated.
• Approximately  95% COD removal was achieved using a se-
  quenced batch  bioreactor followed by carbon column pol-
• Essentially complete removal of priority  pollutants was
  achieved by biological treatment followed by granular activated
  carbon adsorption.

  The effluent quality obtainable is a function of the raw waste-
water concentration, the operating conditions (length of idle time,
temperature, etc.) and the specific treatment process employed. A
summary of the effluent quality achieved, in terms of commonly
used measures, is  presented in Table 2.

                          Table 2
             Summary of Achievable Effluent Quality

Sample Identification
Raw Wastewater
B1o eactor Effluent ^
1 React cycle/wk
1 React cycle/2 wks
1 React cycle/3 wks
1 react cycle/wk
• 5°C
1 react cycle/wk
• 1000 mg/1 PAC
1 react cycle/wk
t 2000 mg/1 PAC
1 react cycle/wk
( 5000 mg/1 PAC
1 react cycle/wk
• 1 React cycle/wk
GAC column polishing







385-410 ,{



























   20* breakthrough v°'

(1) For bioreactors operated at 50% fill.
(2) Equilibrium COD.
(3) Bioreactor effluent contained 465-580 mg/1 COD.
  Based on the treatability study results, seven treatment alter-
natives were identified for cost analysis. The central process in
each of these schemes is biological treatment in a batch reactor.
  These seven alternatives vary primarily with respect to the size
of the activated carbon column units incorporated and the quan-
tity of PAC used. These alternatives are identified below. For our
purpose, carbon breakthrough was defined as the time at which
the COD concentration in the carbon column effluent reached
100 mg/1.

Treatment Processes
• A batch biological reactor followed by a granulated  activated
  carbon (GAC) system sized for 30 min retention time at a flow
  rate of 100 gal/min (System I-A). (Carbon breakthrough ap-
  proximately every 4th batch  treated.)
                                                                                TREATMENT OF HAZARDOUS WASTES    145

• A batch biological reactor followed by a GAC system sized for
  30 min retention at a flow rate of 27 gal/min  (System I-B).
  (Carbon breakthrough approximately every batch treated.)
• A batch biological reactor followed by a GAC system sized for
  60 min retention at a flow rate of 100 gal/min  (System I-C).
  (Carbon breakthrough approximately every 8th batch treated.)
• A batch biological reactor with 1,000 mg/1 powdered activated
  carbon (PAC) addition and a GAC system sized  for 30 min re-
  tention at a flow rate of 100 gal/min (System II-A). (Carbon
  breakthrough approximately every 9th batch  treated.)
• A batch  biological reactor with 1,000 mg/1 PAC addition and
  a GAC system sized for 30  min retention at a flow rate of
  11  gal/min (System II-B). (Carbon breakthrough approxi-
  mately every batch treated.)
• A batch biological reactor with 1,000 mg/1 PAC addition and a
  GAC system sized for 60 min retention at a  flow  rate of 100
  gal/min  (System  Il-C).  (Carbon breakthrough approximately
  every 18th batch  treated.)
• A batch  biological reactor with 5,000 mg/1 PAC addition and
  no GAC system (System III).

Cost Estimates
  Cost estimates were based on the following assumed operating
• Bioreactor volume (approximately  50,000 gal) is  twice the feed
  volume to accommodate a 50:50 ratio of mixed liquor to  waste-
  water feed.
• One batch (24,000 gal) of wastewater is processed each  week.
• Groundwater/leachate is pumped directly to  the bioreactor so
  that the bioreactor is filled and ready for the reactor cycle to
  begin when the operator arrives.
• Twelve hours aeration per batch.
• Two hours settling time per batch.
• Discharge time varies according to GAC system.
• A treatment system operator is available locally  on an on-call
• Minimum effluent quality required is 100 mg/1 COD.
• GAC column breakthrough is  20<%  of COD influent (500 mg/1
  COD reduced to 100 mg/1 COD)  for systems not  using PAC
  (Systems I-A, I-B and I-C).
• GAC column breakthrough is  40%  of COD influent (250 mg/1
  COD reduced to  100 mg/1 COD) for systems  using  1,000 mg/1
  PAC addition (Systems II-A,  II-B and II-C).
• The treatment  system is housed in a temperature-controlled
  Information regarding reactor size, sludge production, carbon
requirements and oxygen  requirements that affected capital and
operating costs was determined on the basis of the  treatability
study results. Capital costs were derived on the basis of vendor
quotes and annually published cost manuals.4'5

  The cost of contract hauling and disposal  at  a  commercial
facility is a function of the  particular firm  that  is  contracted.
Costs for this alternative were based  on quotes from  three com-
mercial waste management firms located within 180 miles of the
PAS site.
  No basis could be established for deriving the cost of treatment
of PAS groundwater/leachate at the local POTW. The city does
not have  a  surcharge system  for  treatment  of  high-strength
wastewater. Furthermore, this POTW has a history of operating
problems.  To alleviate this problem, the city  has established a
policy that no "scavenger wastes" will  be accepted at the plant.
Due to this circumstance, no further consideration was given to
treatment at the local POTW as an off-site alternative.
  A summary of the capital and annual operation/maintenance
(O&M)  costs for each  on-site and off-site treatment alternative
considered is presented in Table 3. These costs were developed on
the  assumption  that  PAS  groundwater/leachate would  be
recovered and treated at a rate of one 24,000 gal batch per week.
The actual  rate at which the groundwater/leachate will  be
recovered is not known.

                           Table 3
             Cost Summary of Treatment Alternatives
Alternative Description
Biological Treatment &
GAC Column Pol ishing:
B 1-8
Cost (S 1986)(1)

Annual Operating Cost Per
Cost ($ 1986)1Z) Gallons (S 1986)13'


     Biological Treatment with
     1,000 mg/1 PAC Addition J
      GAC Column Polishing:

D           II -A
E           II-B
F           II-C

G    Biological Treatment with
      5,000 mg/1 PAC Addition

H    Contract Haul ing/Of f-si te
                                   499,200 - 1,123,000
                                                 0.40 - 0.90
(1) Cost for facilities to treat wastewater in 24,000 gal batches.
(2) Cost to treat 24,000 gal/week.
(3) Cost/gal based on 1,248 million gal/yr (i.e., 52 wks x 24,000 gal/wk).

  Consideration of the quantity of groundwater/leachate to be
treated with time is essential to a comparison of the costs of the
various treatment alternatives, since these costs are a function of
both the quantity  of wastewater to be treated and the length of
time over which treatment is to be performed. The total quantity
of groundwater/leachate requiring treatment has been estimated
to be 3 million gallons based  on hydrogeological data collected
during the RI/FS and the Remedial Design Study. To address the
various rates at which this wastewater may be recovered,  four
scenarios  have been developed (Table 4).

                           Table 4
Hypothetical Rates of Groundwater/Leachate Recovery and Treatment
                                       Remaining years
1 MG
1 MG
1 MG
1 MG
8.8 days
8.8 days
8.8 days
8.8 days
1 MG
da v*.
(1) Assumes 24,000 gal per cycle.
  A present worth analysis of  each  treatment  alternative as a
function of the four recovery rate scenarios is given in Table 5.
The following observations are based on a review of this analysis:
• All on-site treatment alternatives (A through G) are more cost-
  effective than the off-site alternative (H).
• On-site alternatives which incorporate the smaller sized GAC
  columns  (B,E)  are more cost-effective than alternatives  in-
  corporating larger columns  (A,C,D,F).

                          Table 5
       Present Worth Analysis of Treatment Alternatives
(1) (2)
Scenario Alt.
1 A
2 A
3 A
4 A
1st Yr.
Cost ($xlOJ)
87. 4
Remaining Yrs
Annual Cost
2;. 5
Column 1: Scenario (Refer to Table 7-2)
Column 2: Treatment Alternative (Refer to Table 7-1)
Column 3: Total number of years that a treatment system is operated on-site
Column 7: Calculated on basis of:
  Present worth of capital and annual O&M at year n
  = (P/A, i, n) x (O&M) + Capital

  = (1 + i) n  - 1
              x O&M + Capital
     i (i + i)n
 where i = interest rate per year
     n = number of years
• Biological treatment enhanced by a 5,000 mg/1 dose of PAC
  (Alternative G) is more cost-effective than any of the alterna-
  tives (A through F) which include GAC columns  for effluent

  Based on laboratory treatability studies, biological treatment
with activated carbon polishing was determined to be the most ef-
fective treatment method. These studies  also demonstrated the
feasibility of using a batch  bioreactor operated intermittently.
The disadvantage of lower COD removal rates when the bioreac-
tor was operated at low temperature and increased idle periods
was overcome by the  addition  of powdered activated carbon.
Results of a present worth analysis showed that on-site treatment
will be much more cost-effective than off-site treatment/disposal
at a commercial facility.

1. "Site Investigations and  Remedial Alternative  Evaluations at the
   Pollution Abatement Services (PAS) Site in Oswego, New York,"
   Draft Final Report, prepared for New York State  Department  of
   Environmental Conservation by URS Company, Inc., Jan. 1984.
2. "Engineering Design Report for Site Remedial Measures Pollution
   Abatement Services Site,  Oswego,  New York,"  prepared for New
   York State Department of Environmental Conservation by  URS
   Company, Inc., May 1985.
3. "Evaluation of Alternatives  for Treatment of Groundwater/Leachate
   at the  Pollution Abatement Services (PAS) Site in Oswego, New
   York," Final Report, prepared for New York State Department of
   Environmental Conservation by URS Company, Inc., Oct. 1985.
4. Building Construction Cost Data 1985, published by the Robert Snow
   Means Company.
5. Dodge  Manual for Building Construction, Pricing and Scheduling,
   published by McGraw-Hill Information Systems Co., Vol. 2.
                                                                                    TREATMENT OF HAZARDOUS WASTES    147

                         Optimization of  Free  Liquid  Removal
                Alternatives  in the  Closure  of  Hazardous Waste
                                      Surface  Impoundments
                                                David K. Stevens
                                                 Black & Veatch
                                              Kansas  City, Missouri
  A flowchart is presented to serve as a guide in determining the
optimal alternative for removal of free liquids during the closure
of hazardous waste surface impoundments. The flowchart ad-
dresses each part or element  within a closure alternative as a
discrete decision process. This flowchart allows separate evalua-
tion of each phase of removal such as preclosure engineering,
removal and treatment and ultimate disposal.  Criteria used to
select  the best  option within  an element include technological
feasibility, environmental soundness, regulatory acceptance and
unit cost. Once the best elements for each phase of free liquid
removal are determined,  they are recombined to form the optimal
removal alternative.
  Important factors such as in situ reduction and provisions for
mixed  wastes are discussed. Removal equipment and  treatment
processes are presented as a function of the waste type. Ultimate
disposal options such as discharge to surface water vs. discharge
to sewer, on-site vs. off-site treatment and disposal and process
reuse  and recovery are  included. To illustrate the use of the
flowchart, three  case studies of actual  surface impoundment
closures are  presented. Optimal free liquid removal alternatives
are developed for impoundments containing heavy metals, heavy
metals and degreasers, and oily wastes.

  There are tens of thousands of hazardous waste lagoons, pits or
ponds  in  the United States.' These  types of storage facilities or
surface impoundments are used for continuous or intermittent
storage of process water, stormwater and spent products from the
chemical and manufacturing industries. RCRA  or CERCLA can
require closure of an impoundment when soil and groundwater
contamination is discovered or if the impoundment poses an en-
vironmental  risk. Stricter liner requirements for impoundments
under RCRA are also forcing an increased number of closures.
Given the trend toward greater environmental protection, increased
enforcement  and stricter regulations, it is certain that  impound-
ment closures and their associated remedial actions will continue
at a rapid pace.
  Regardless of the RCRA or CERCLA regulations  governing
closure, removal of free liquids is required as one of the first steps
in an impoundment closure. The removal process can present a
number of technological and regulatory  problems as  a  large
volume of liquid hazardous waste must be removed from the site
which frequently  has no suitable direct outlet. The free liquids
often  have a high solids content, exhibit extremes of pH and
viscosity, have concentration gradients and are stratified mixtures
of both inorganic and organic chemicals. These  technological
problems combined with the regulatory pressures for tight closure
schedules and strict air and water discharge limits require careful
engineering of the free liquid removal processes. Recognizing the
inherent problems in dealing with liquid hazardous wastes in sur-
face impoundments, the flowchart directs the user in the selection
of the optimum free liquid removal alternative.

  In assembling the  flowchart elements into a system  of logical
steps, the universe of processes used for free liquid removal were
identified. Once the available processes were identified, they were
categorized into three elements: preclosure engineering; removal
and treatment; and ultimate disposal.  Figure 1 shows the basic
outline of the flowchart. To determine the optimal removal alter-
native, each process within an element is subjected to a set of deci-
sion criteria developed from recent CERCLA guidance1  and ac-
tual closure experience.  Through evaluation of each process ac-
cording to the developed decision criteria, the best option for that
process can be identified.  By  combining the best options within
each element, the optimal removal alternative is determined.

Flowchart Elements
   The flowchart contains  three  major  elements:   preclosure
engineering; removal and treatment; and ultimate disposal (Fig. 1).
Pre-closure engineering addresses the  processes  for flow diver-
sion, waste characterization and the removal alternative's impact
  •  Flow Div0nion
  •  Waste Characterization
  •  Facility Cloaure Plan
                   REMOVAL AND TREATMENT
                    •  In Situ Reduction
                    •  Pumping and Dredging
                    •  Treatment
                     ULTIMATE DISPOSAL OR
                         •   On Site
                         •   Off Site
                              READY FOR
                            FINAL CLOSURE
                         Figure 1
             Overview of Flowchart for Determining
              Optimal Liquid Removal Alternatives

 on the overall facility closure plan. The removal and treatment
 element is composed of three processes: in situ reduction, liquid
 removal and treatment prior to ultimate disposal. In situ reduc-
 tion considers whether the hazard or volume of the free liquid
 should or can be reduced while in the impoundment before actual
 removal.  The removal section addresses the equipment used in
 removing  the waste, such  as  pump  type  and construction
 material. Treatment alternatives include physical and chemical
 treatment of the free liquids for reduction or elimination of the
 hazard from the waste before ultimate disposal. The treatment
 process and the ultimate disposal option often are synergistic, re-
 quiring concurrent evaluation.  The ultimate  disposal  element
 reviews the options for final deposition of the liquids such as sur-
 face water, publicly owned treatment works or an RCRA facility.
   Careful consideration always should be given to those processes
 which allow reuse or  recovery of the liquids.

 Decision Criteria
   The decision criteria employed for evaluation of each closure
 element are expressed in both cost and non-cost factors. Cost fac-
 tors  include  the  cost  for  present  value of  operating  and
 maintenance costs and the capital expenses. Non-cost factors in-
 clude  technological  feasibility,  environmental  soundness  and
 regulatory acceptance. Technological feasibility requires that the
 technology under consideration be proven at the design scale or
 that pilot-scale results show a good chance for success in the field.
 Environmental soundness is a measure of the exposure of the
 chemical constituents to the environment during the removal
operation. This decision can be a judgmental one as in some cases
environmental exposure occurs only for a short time and does not
persist once the removal is complete. Regulatory acceptance is the
third criterion. This factor indicates whether the closure processes
under consideration are acceptable to the governing  regulatory
agency.  Acceptance or  non-acceptance depends on the pertinent
regulations and the perceived  environmental risk.

  During development  of an RCRA closure plan or a CERCLA
feasibility study,  several preliminary concerns must be addressed
before the  actual free  liquid  removal alternatives can be fully
developed. Among these concerns are options for flow diversion,
requirements for complete waste characterization and the re-
moval alternative's impact on the overall impoundment closure
plan.  In the flowchart, Figure 2, these concerns are taken into
consideration in the preclosure engineering element. Carefully ad-
dressing each of these concerns is required for optimization of the
free liquid removal alternative.

Flow Diversion
  The initial step in implementation of an impoundment closure
plan is to eliminate the source of incoming waste. This is not a
major concern for inactive impoundments or impoundments used
for low-volume intermittent waste disposal. At active impound-
ments receiving  continuous or  semi-continuous waste streams
such as  process water or contaminated stormwater, careful con-
sideration is required as dedicated treatment plants and long-term
operation become potential design factors.
                                                                        GENERATED SLUDGE DISPOSAL


- Biological
• Clu-oncnl
• Physical

i 1
• PhvMC.ll




• Reuw; —
• Recovery
• Land Apply

• Surfiicu WuU-r
• RCRA Facility
                                                           Figure 2
                                                 Free Liquid Removal Flowchart
                                                                                TREATMENT OF HAZARDOUS WASTES    149

Waste Characterization
  Accurate information on the chemical constituents in the im-
poundment is essential, including identification of any variances
in the composition as a function of depth. Each layer can require
special consideration and different removal procedures; for exam-
ple, a floatable oil layer can cover a relatively clear aqueous layer
which in turn covers a gelatinous metal hydroxide sludge. The im-
portant physical parameters include pH, viscosity and solids con-
tent. The pH influences pump selection and treatment processes.
Viscosity and solids  content influence the pump type and the
volume of liquid available for removal by pumping alone. Sludges
with solids content greater than 8-10%  must be removed either by
first mixing with  the  other impoundment  liquids  to  lower the
solids content, or by  leaving the high solids content sludge in
place and dewatering by evaporation or by the  addition  of a
solidifying agent.

Impact on the Overall Closure  Plan
  Consideration of the free liquid removal alternative's impact on
the overall closure plan is imperative for optimizing the impound-
ment closure.  Very often other  remediation  activities at  the facil-
ity can  influence selection of the  free liquid removal alternative.
These considerations  are  site-specific and are best illustrated
through examples. Table 1 shows examples of the impact of the
selected closure plan on the free liquid removal alternative.

                           Table 1
                 Impact of Facility Closure Plan
               on Free Liquid Removal Alternatives
Waste Type
Heavy Metals
Oily Wastes
PCB Oils
Facility Closure Plan
Groundwater Treatment
Reclamation of Wastes
Soil Incineration
Free Liquid
Removal Alternative
Same Treatment
Process Employed
Free Liquids Re-
claimed by Same
Free Liquids
In Situ Reduction
  In situ reduction is a process within the treatment and removal
element (Fig.  2). It involves the reduction of the hazard or the
volume of the waste present in the impoundment before actual
removal. Such techniques can be used to facilitate final removal
and disposal of  the waste.
  In dealing with very corrosive wastes such as spent pickle liquors
with pH C2 or caustic washes with a pH > 12, it is advantageous
to add a neutralizing agent to the wastes in the impoundment. Ad-
dition  of bases also can be used  to remove toxicity such as the
alkaline precipitation of heavy metals at pH 9-10. In small acidic
ponds,  a 50% solution of sodium hydroxide is best as it  is easily
pumped at moderate ambient temperatures and offers a  high
degree of neutralizing power.  For larger acidic ponds or when
cost is an important factor, calcium hydroxide and calcium car-
bonate  should be examined. Multi-stage treatment with several
neutralizing materials has been investigated.' For alkaline ponds,
concentrated sulfuric acid is generally best as it is inexpensive, of-
fers a  high degree of neutralizing ability  and does  not pose  a
serious  material  handling problem. Other possibilities for in situ
hazard  reduction are the use of oil booms, reduction/precipita-
tion techniques  and activated carbon or  polymer addition for
organics adsorption.
  The other in situ alternative is to reduce the volumetric quantity
of free liquids  in the impoundment. The two primary mechanisms
are evaporation and  seepage.  Metal-containing  aqueous wastes
are usually more amenable to evaporation or seepage than organic
wastes. Although a limited number of sites actually experience a
high net  evaporation rate,  judicious scheduling  of the  liquid
removal step during periods of low rainfall and high temperature
can effectively reduce the total volume of free liquids. For exam-
ple, at  a West Coast manufacturing facility the volume of water in
a stormwater retention pond  varied from 3.5 million gallons dur-
ing January  to 2.5 million gallons during August. Determining
whether seepage is a viable method of volume reduction is  a less
straightforward  calculation  than determining  the evaporation
potential.' The use of impoundments as seepage basins for con-
centrating wastes has been documented to be very effective with a
limited impact on ground water quality.'

Free Liquid Removal
  Physical removal of the free liquids  is accomplished best by
pumping. If the waste has a high solids content,  mechanical, hy-
draulic or pneumatic dredging  may be required.  This  paper
discusses only pumping of free liquids.  Selection parameters in-
clude the type, construction materials  and size  of pumps. The
pump  type and construction  materials usually  depend on the
waste  material and its  physical  characteristics. Pump size  is
governed by the closure schedule and treatment system capacities.
Table 2 delineates the generally available pump types and con-
struction materials as a function of the waste type.

                            Table 2
             Pump Types and Construction Materials
                  as a Function of Waste Type
                                                                   Physical Description of Waste
                             Pump T)pe
Vi.scous or Oily
Sludge. 1-5% solids
Sludge. 5-8% solids

Chemical Description
of Waste

Oils, phenols
Heavy metab
Basic pH >I2
Acidic pH < 2
                                                                                                Rotary lobe, progressive cavity
                                                                                                Posiihe displacement

                                                                                                Construction Material
Cast iron
Cast iron, possibly bronze fittings
Bronze fitted cast iron
Stainless steel
SS ASTM 743. CN-7M
SS ASTM 743. CN-7M (3«*o Mo)
   At this stage in the evaluation process, the free liquid treatment
and disposal processes are developed. Selection of the treatment
process  is  based on  both  the waste  characteristics  and  the
availability  of ultimate disposal mechanisms. Relationships  be-
tween the disposal processes and treatment requirements are sum-
marized in Table 3.
                            Table 3
 Relationship Between Disposal Processes and Treatment Requirements
 Disposal Processes

 Off-site disposal
 On-site disposal
 On-sile reuse or recovery

 Publicly owned treatment works
 Process reuse

 Surface water
 Groundwatcr rcinjection
   Treatment Requirements

   Usually requires no treatment
   except suspended solids removal

   Suspended solid removal often

   Treatment to drinking water
   standard required

  Since the treatment process and the ultimate disposal method
represent the largest single capital cost items involved in free liq-
uid removal,  all alternatives need to be critically reviewed. Con-
sideration should be given to alternatives which provide a poten-
tial for reuse or combination with other treatment needs.
  Processes available for the treatment of aqueous hazardous
waste are numerous.6 The major limitation to the use of many of
these processes is that many are not cost-effective or not proven at
the field scale. New technologies and innovative processes are best
explored when a long-term use treatment plant is required, such
as in groundwater remediation programs or surface and process
water treatment systems. In other cases, leasing of mobile treat-
ment units or  low-cost  temporary equipment  should be con-
sidered. Table 4 shows the technologies available for treatment as
a function of the waste type.

                           Table 4
        Treatment Technologies as a Function of Waste Type
Waste Type
Acidic pH<2
Basic pH >12
Heavy metals
Low level
High level
Alkali addition
Acid addition
Ion exchange
Air stripping
Carbon adsorption
alkali addition
Acid addition
Air stripping
Carbon adsorption
alkali addition
Acid addition
Carbon adsorption
Direct Off-Site Disposal
  In direct off-site disposal of free liquids, the liquids are taken
directly from the impoundment to an off-site facility. Potential
off-site facilities  include publicly owned  treatment works and
RCRA treatment facilities. If the concentrations of the hazardous
compounds are within sewer discharge limits, the liquids can be
pumped directly  into the sewer system. If the sewer line is not
within  economical  pumping   distance,  alternative  means  of
transportation to the treatment plant must be explored.  A most
practical means is to employ vacuum trucks or similar tanked
vehicles. Transportation costs  can be as much as $1.50 per
unloaded mile and $3.50 per loaded mile.  Sewer discharge rates
vary widely depending on  waste type,  quantity and treatment
facility. When the pollutant concentrations exceed the influent
pretreatment standards and direct off-site  disposal is a potential
alternative, a list of available facilities should be compiled. Many
available directories list RCRA  facilities, and the RCRA hotline
also can provide a list of currently permitted facilities. Transpor-
tation costs are similar to the amounts given above. Disposal costs
can  be from $0.10/gal  for low level contaminated  aqueous
streams to over  $l/gal for liquid injection of  highly  organic
wastes. Deep well injection, generally for  acids and bases only,
costs approximately $0.20/gal.

Treatment Followed  by Off-Site Discharge
  Employing a pretreatment process expands the number of off-
site disposal alternatives  to include surface water discharge and
groundwater injection. Groundwater injection is predominantly
used only in conjunction with other groundwater remediation
programs. Discharge to surface water requires an NPDES permit
as well as very low contaminant concentrations. For metals, some
type of filtration will be required to attain the discharge limit.

On-Site Disposal
  The alternatives to on-site disposal include construction of an
on-site RCRA permitted facility. For most closures at industrial
manufacturing  facilities,  the closed  system with  reuse and
recovery is the most advantageous. At uncontrolled sites, on-site
disposal of the liquids usually is confined to treatment processes
such as landfarming, evaporation or incineration.

  Final consideration for free liquid alternatives is given to the
disposal of sludges generated  during a treatment process. Sludge
is generated during alkaline precipitation of metals, filtration or
solidification/stabilization processes.  Potential  disposal  alter-
natives include off-site disposal in an RCRA permitted landfill,
on-site disposal in a specially constructed RCRA permitted land-
fill and disposal in the closed impoundment.

  Three case studies  of actual surface impoundment closures are
presented here to illustrate some of the aspects in determining the
optimal alternative for free liquid removal. The following cases
are discussed:
• Stormwater retention pond  containing hexavalent and trivalent
• Process water storage lagoon contaminated with chlorinated
  solvents and nickel
• Earthen pit used for oily waste disposal

Chromium Contamination
  A manufacturing facility employed an earthen pond to store
stormwater runoff from its product storage area. The pond was
being closed because of its ineffective liner and the presence of
chromium in the underlying groundwater.

Preclosure Engineering
  The pond currently was being used for storage of stormwater,
and  an alternative mechanism Tor handling future runoff was
needed. Since manufacturing and storage operations were to con-
tinue at the site and no practical method of eliminating the con-
taminated runoff existed, a permanent treatment and disposal
solution was required. An ion-exchange treatment system with
surface water discharge was determined to be the best alternative
for addressing the runoff. Ion-exchange produced a very low level
concentration of chromium in the effluent and allowed reuse of
the recovered chromium. Ion  exchange also added the advantage
that no hazardous waste sludge was generated during treatment as
would have  been  produced using traditional chemical reduction
and precipitation.
  Results  of the pond sampling activities  showed that approx-
imately 400,000 ft 3  of free  liquids containing  5-10 mg/1 hex-
avalent chromium and 1-5 mg/1 trivalent chromium were in the
impoundment. Underlying the free liquids were 1000 ft3 of a
chemical  floe  sludge containing 1000 to  5000  mg/1 total
chromium. The-low chromium concentrations in the free liquids
permitted  use of the  current ion-exchange system for free liquid
treatment. The remaining sludge could be solar dried and fixed in-
place using a solidification agent.  Use of the on-site treatment
system was an attractive solution  for treatment as there was no
negative impact on the other closure activities and no additional
permitting was required for surface water discharge.
                                                                                 TREATMENT OF HAZARDOUS WASTES    151

Removal,  Treatment and Ultimate Disposal
  Given the overwhelming advantages  in using the on-site ion-
exchange treatment system, the free liquids were removed using
an available contractor's pump, treated to NPDES requirements
and  disposed  via the force main to the nearby surface water
stream. Judicious closure scheduling during the summer months
reduced the volume of free liquids from 400,000 ft3 to 260,000 ft3.
This timing reduced pumping time from 14 days to 9 days and
saved approximately $20,000 in treatment expense.
  During the alternative review phase, contractor proposals were
solicited for removal and treatment of the free liquids. The con-
tractor proposed to treat the liquids using sulfite reduction fol-
lowed by  alkaline  precipitation.  For comparison purposes, the
decision criteria evaluation of the two treatment alternatives  is
shown below:
lowered  below  discharge requirements.  Different  treatment
schemes  were developed and compared using the four decision
criteria. The results are summarized below.
Use of On-site
Process Reuse
More Favorable
Unit Cost
S/1000 gal
Degreasers and Contamination
   A metal plating and fabrication facility used an earthen im-
poundment for disposal of process rinsewaters.  Groundwater
contamination  by the degreasers was forcing closure of the im-

Preclosure Engineering
   Due to the continuation of fabrication operations at the plant,
a permanent treatment solution was needed for the waste process
rinsewaters. Through water-use reductions and process modifica-
tions,  the  rinsewater flow rate  was reduced  from 300  to  50
gal/min. A 50 gal/min treatment system was put on line early dur-
ing the closure plan development. Treated rinsewater was dis-
charged to nearby surface water per the NPDES permit require-
   The impoundment contained approximately 130,000 ft3 of free
liquids, 8,000 ft3 of pumpable sludge and 2,000 ft3 of unpump-
able sludge. The nickel and degreaser concentrations in the free
liquids ranged from 30 to 60 mg/1 and 0.2 to 0.9 mg/1, respective-
ly. Both sludge layers contained higher levels of contaminants
than the free liquids, with the unpumpable sludge layer contain-
ing the bulk of nickel and degreaser.
   Due to the depth of degreaser contamination in the soil, the im-
poundment was closed with contaminated soils in place. Labora-
tory studies were conducted on both sludges to determine if they
were  amenable  to  stabilization.  If adequately  stabilized,  the
sludges could be disposed of by leaving them in  the impound-
ment. Tests showed that  the addition of portland cement effec-
tively formed a stable and solid mixture. Using the stabilization
process, only the free liquids needed to be removed and disposed;
the pumpable and unpumpable sludges were stabilized and left in

Removal, Treatment and Ultimate Disposal
  The volume of free liquids in the impoundment precluded the
use of the process water treatment system; the time required for
treatment using the system would exceed the required 180-day
schedule. An alternative mechanism was required. Due to the low
level of contaminants in the free liquids, consideration was given
to the use of a  temporary on-site system consisting of filtration
and possibly  pH adjustment. Laboratory tests confirmed that
with pH adjustment  and settling, the contaminant levels were
& Settling
Unit Cost
5/1000 gal
  It was decided that gravity sand filtration would produce the
best quality effluent at the lowest cost. To enhance operation of
the filtration equipment, care was taken to maintain a low level of
suspended  solids in  the  free liquids.  A  floating intake was
employed to minimize disturbance of the  underlying sludge.  In
addition, when the suspended solids concentration exceeded 400
mg/1, the pump speed was decreased to further minimize resus-
pension. Backwash  water was  held  on-site, and the settled
material was returned to the impoundment after all the free liq-
uids were removed.

Oily Wasles
  An earthen pit located on a farmer's  property had been used
for the disposal of spent cutting  and hydraulic oils. Closure was
proceeding as a generator-directed and funded mitigation action.
The site functioned with little control, and  many accounts of un-
authorized  dumping had been  recorded.  However, one major
generator had  taken prime responsibility for closure instead of
waiting  for a potential enforcement action.

Preclosure Engineering
  The impoundment was receiving intermittent waste streams only,
and additional waste inputs were eliminated by banning all future
waste disposal practices. Results of the pond sampling activities
showed  that three distinct layers were present: an oil layer of ap-
proximately 10,000  ft 3, an aqueous  layer of 60,000 ft3 and a
viscous oily sludge layer of 4,000 ft3. The aqueous layer contained
between 0.2 and 0.5% total hydrocarbons  and could not be dis-
posed as a non-hazardous waste. However, analytical analysis of
the oily  layer showed that it potentially could be reclaimed.

Removal, Treatment and Ultimate Disposal
  In this closure, several site-specific constraints precluded many
of the normally available options; no surface water or sewer line
was readily available. Due to the uncontaminated nature of the oil
it allowed the potential for recovery. The oil layer was removed by
vacuum trucks and taken to an oil recovery  plant for recox cry and
reuse. Since no outlet  for  the treated  aqueous layer was readily
available,  an  off-site  disposal  alternative  was  required.  The
aqueous layer was removed by vacuum  trucks and taken to the
generator's manufacturing facility where it was treated  in their
own industrial wastewater treatment plant. Although the trucking
costs were substantial, the oil recovery and generator treatment of
the aqueous waste offset a portion of this cost. It was determined
that disposal of the viscous sludge was accomplished best by land-
filling rather than by incineration, since incineration was an order
of magnitude in cost higher than landfilting.

  A flowchart has been presented which guides in determining the
optimal alternative for removal  of free  liquids from  hazardous
waste surface impoundments. The flowchart addresses each ele-
ment of the closure process as a separate  entity, whereas in-

dividual decision criteria are applied to each element before com-
bining the elements to form the optimal alternative. The major
elements include: preclosure engineering; removal and treatment;
and ultimate disposal.
  General trends identified from the case studies presented in this
paper are:

• Identify site-specific constraints early during closure plan  de-
  velopment. The combination of free liquid treatment and dis-
  posal options with soil and groundwater remediation programs
  can substantially reduce the cost for free liquid removal.
• Use recovery and reuse whenever possible; this alternative is at-
  tractive environmentally and to regulatory agencies.
• Employ temporary on-site treatment measures when levels of
  contaminants are low. This process is less expensive than off-
  site disposal at RCRA facilities.
• When  available, use  on-site process water or  groundwater
  waste treatment equipment to treat the free liquids.
1.  Wyss, A.W., et al., "Closure of Hazardous Waste Surface Impound-
   ments," prepared by Acurex Corporation for the U.S. EPA, Sept.
2.  "Guidance on Feasibility Studies Under CERCLA," prepared by JRB
   Associates for the U.S. EPA, June 1985.
3.  Hale, F.D., Murphy, C.B. and Parrat, R.S., "Spent Acid and Plat-
   ing Waste Surface Impoundment  Closure,"  Proc.  Third National
   Conference on Management of Uncontrolled Hazardous Waste Sites,
   Washington, DC, Nov. 1982, 195-201.
4.  McWhorter,  D.B., "Seepage in the Unsaturated Zone: A Review,"
   Proc. Seepage and Leakage from  Dams and  Impoundments, May
   1985, 200-219.
5.  Looney, B.B., "Surface Impoundment Legacy: Field Studies," paper
   presented at the  Hazard  Materials Management Conference, June
6.  Sittig, M., "Pollutant Removal Handbook,"  Noyes Data Corpora-
   tion, Park Ridge, NJ,  1973.
                                                                                 TREATMENT OF HAZARDOUS WASTES     153

                Assessment  of Chemical  Treatment  Technologies
                     and  Their Restrictive Waste  Characteristics

                                             Hamid Rastegar, Ph.D.
                                                 James Lu, Ph.D.
                                                   Chris Conroy
                                        Jacobs Engineering Group, Inc.
                                               Pasadena, California
  The applicability of a particular  treatment technology  to a
given liquid waste stream depends upon the physical, chemical
and biological characteristics  of the waste. Waste stream char-
acteristics which  restrict a waste's  applicability to a particular
method of treatment are presented.  With this information, the
most effective treatment technology for a waste stream can be
determined. The  technologies evaluated include chemical oxida-
tion  and reduction, chlorinolysis, chemical fixation,  ion ex-
change, neutralization and precipitation.

  The land disposal of many hazardous wastes may be banned
within the coming years under the RCRA amendments. For those
wastes which will be banned  from landfill disposal, it is neces-
sary to determine alternative methods of treatment and disposal.
  A  wide variety of treatment technologies exists today which
can be used to treat such wastes.  Treatment technologies can be
classified as either chemical, physical, biological or thermal treat-
ment. In this paper,  the authors evaluate the applicability of
chemical  treatment technologies  for  hazardous liquid  waste
streams based on  the wastes restrictive characteristics.
  This analysis is part of an extensive study by the U.S. EPA to
determine the applicability of alternative waste treatment  tech-
nologies. The most promising types of each chemical treatment
technology are discussed. The major restrictive waste characteris-
tics determining  treatment applicability  for each also  are pre-
sented. Finally, the final product or residue from each treatment
technology and the limitations of each are evaluated.

  Several chemical treatment technologies are  widely used for
waste treatment.  The most effective technologies discussed here

  Chemical Oxidation
  Chemical Reduction
  Ion Exchange

  The major restrictive waste characteristics as well as the type of
applications for each of the above technologies are given below.
A summary of the major restrictive waste characteristics is  given
in Table 1.
Chemical Oxidation
  Chemical oxidation is a process which raises the oxidation state
of chemical species, converting the wastes into less-hazardous or
non-hazardous  forms.  The oxidation  is accomplished by  the
addition of an oxidizing agent which is reduced.
  Chemical oxidation has been used most commonly to treat cya-
nide wastes from electroplating operations. The most widely used
oxidants  for this application are chlorine gas and calcium and
sodium hypochlorite.
                          Table 1
   Chemical Treatment Technologies Restrictive Waste Characteristics
Chemical Treatment
Major Restrictive Waste
Chemical Oxidation

Chemical Reduction



Ion Exchange

Physical form, oil and grease content, sus-
pended solids content, viscosity.
Physical form, oil and grease content, sus-
pended solids content, viscosity.
Total solids content, oxygen, sulfur, and
organic content, water content.
Physical  form,  organic  content,  ionic  com-
position, sulfate content.
Suspended solids content, metal content,
organic content, oxidizing agents, physical
Dissolved solids content, physical form.
Physical form, viscosity, metal solubility.
  Other oxidants for cyanides include ozone, hydrogen peroxide,
potassium permanganate and chlorine dioxide. A wide variety of
organics and inorganics also can be treated by oxidation. The
end products of cyanide oxidation are nitrogen gas and carbon
dioxide. The end products of the oxidation of organics are either
simpler organics or carbon dioxide and water; oxidized metals
form metal hydroxides which can be precipitated out of solution
by pH adjustment.
  The  major restrictive characteristics of a waste which are used
to determine the applicability of chemical  oxidation include the
physical form, oil and grease content, suspended solids content

and viscosity of the waste. Chemical oxidation is most effective
for dilute aqueous wastes, although certain gaseous wastes also
can be treated. Since oxidizing agents will indiscriminantly attack
all oxidizable material, the oil and grease content and the oxidiz-
able  suspended solids content should be low,  preferably below
1%.  For liquid solutions,  viscosities much greater than that of
pure water are restrictive  because of the mixing problems en-
countered with highly viscous solutions.

Chemical Reduction
  Chemical reduction lowers the oxidation state of chemical spe-
cies,  converting  them  into less-hazardous  or non-hazardous
forms. The reduction is accomplished by the addition of a reduc-
ing agent which is oxidized.
  Chemical reduction is used most  commonly to  convert hexa-
valent chromium into trivalent chromium, which then is precip-
itated out of solution as chromium hydroxide. Mercury also can
be removed from solution by reducing mercury (II) to its elemen-
tal state. Sulfur dioxide and sulfite salts are the most widely used
reducing agents  for chromium. Sodium borohydride is an effec-
tive reducing agent for mercury and lead.
  The major restrictive waste characteristics for chemical reduc-
tion include  the wastes' physical  form, oil  and grease content,
suspended solids content and viscosity. These are the  same re-
strictive characteristics as for chemical oxidation, since chemical
oxidation and reduction are similar processes.

  Chlorinolysis involves the reaction between chlorine and liquid
chlorinated hydrocarbons  at temperatures of 500 °C or greater
and high pressure to form carbon tetrachloride, hydrogen chlor-
ide and other by-products. In the Chlorinolysis process, the waste
first is  pretreated to remove all solids  and  any unwanted com-
pounds which could result in the formation of unwanted reac-
tion by-products. Then, the purified waste undergoes reaction
with chlorine, and the reaction products are separated and recov-
ered by distillation.
  The waste to be treated by  Chlorinolysis ideally should con-
tain only liquid chlorinated hydrocarbons. The presence of other
compounds may produce toxic reaction by-products, which may
include carbon tetrachloride, hydrogen chloride and phosgene.
The  restrictive waste characteristics include the solids oxygen,
sulfur organic and water content. As mentioned previously, the
presence of compounds other than  chlorinated hydrocarbons is

  Chemical fixation is a process in  which hazardous substances
are stabilized and solidified in a form which is considered non-
hazardous and safe for ultimate disposal. Various types of ma-
terials can be mixed with a waste to form a solid material which
effectively immobilizes  the waste. The process is intended to be
permanent, allowing the safe disposal of the solidified mass to a
landfill or other permanent disposal facility.
  Fixation is used for both inorganic and organic solids and
sludges. It is possible to treat liquids, although the effectiveness
of this is less certain. Fixation also is used for low-level radioac-
tive waste. Fixation of inorganic wastes  is accomplished by solid-
ifying the waste in a cement, pozzolanic or lime based material.
Fixation of organic wastes usually requires the use of a glassifica-
tion, organic polymer or thermoplastic technique.
  The major restrictive  waste characteristics which determine the
applicability  of  fixation  include  the  wastes'  physical   form,
organic content, ionic composition and  sulfate content. Fixation
is most effective on solids,  and sludges.  The greater the moisture
content of the waste, the more difficult  fixation is, although
liquids can be treated. A high organic content may limit or cause
problems when fixating an inorganic waste, because the organics
may interfere with the solidification of the solid product. Simi-
larly, high ionic compositions and high concentrations of sulfate
interfere with stabilization and solidification.

Ion Exchange
  Ion exchange is the reversible interchange of ions between a
liquid and a solid phase. Ions in a liquid waste stream and ions on
the surface of an ion exchange resin are exchanged, purifying the
waste stream while concentrating the waste constituent on the
  Ion exchange resins are either cationic  or  anionic. Cationic
resins exchange positive ions such as H+  or Na+  for negative
ions in solutions such as SQ-3. Anionic resins  exchange negative
ions such as OH - and Cl -  for positive ions in solution, such as
metal cations.
  After an ion exchange  resin  has  reached its capacity for ex-
changing ions, it is  regenerated so that it can be used again.
Cationic ion  exchangers are regenerated by passing dilute acid
solution over the resin; anionic resins are regenerated with a weak
base. During the regeneration  process,  a solution containing a
high concentration of the  original waste constituent is produced.
This concentrated waste then must be treated or disposed of prop-
  The restrictive  waste characteristics for  ion exchange include
suspended solids, metal content and organic contents, oxidizing
agents concentration and physical  form.  Ion  exchange  is most
effective for very dilute aqueous wastes. High concentrations of
solids, organics and oxidizing agents interfere with the exchange
of ions between the solution and the  resins.

  Neutralization involves  the addition of an acid or  a base to a
solution to adjust the pH, usually to between  6 and 9. Neutral-
ization often is performed  in connection  with other  treatment
processes, such as oxidation or reduction. It can be either a final
treatment or a pretreatment step.
  Alkaline neutralization  is  commonly accomplished using lime,
sodium hydroxide or soda ash. Acidic  neutralization  utilizes
hydrochloric, sulfuric or nitric acid.  Neutralization is most applic-
able for acidic or alkaline aqueous wastes, although it also can be
used  for certain  organic  liquids. Neutralization is  a relatively
simple but important waste  treatment process  and most  often is
followed by precipitation  of metals  from solution. This technol-
ogy can be used  effectively for most aqueous  wastes, although
wastes with very high dissolved solids content may result in the
formation of complexes which are difficult to remove from solu-

  Waste removal  by precipitation involves the addition of a
chemical substance to alter the equilibrium affecting dissolved
and suspended solids solubility. Precipitation  normally requires
pH adjustment, followed by the addition of a precipitating agent,
resulting in flocculation and  coagulation of the solids in solution.
  Precipitation is used most commonly to remove heavy metals
from aqueous wastes. The precipitation process produces a sludge
composed of metal hydroxides, metal carbonates or metal sul-
fides  as well  as the precipitating agent used. In some instances,
precipitation can be used for organic-based liquids, although this
application is very limited due to sedimentation problems in vis-
cous media.
                                                                                  TREATMENT OF HAZARDOUS WASTES    155

   The major restrictive waste characteristics limiting the use of
precipitation include voscosity, physical form and metal solubility
constant. As mentioned before, highly viscous  solutions, such as
many organic solvents, may  not be applicable for treatment by
precipitation because of difficulties in  settling the metals or other
particles out of solution. For highly soluble species, pH adjust-
ment is needed to minimize solubility and enhance precipitation.
  The applicability of seven chemical treatment technologies for
hazardous  liquid waste treatment has  been examined.  It was
found that  the  proper choice of treatment depends  upon  an
understanding of  the  restrictive  waste  characteristics  for each
technology. Once the waste characteristics which restrict the use
of a particular treatment technology are  known, proper pretreat-
ment or other treatment methods can be chosen.

                Treatment Technologies  for Hazardous  Materials
                                                 Jeffrey M. Thomas
                                                    DETOX, Inc.
                                             San Francisco,  California
                                                  Phillip T. Jarboe
                                                    U.S. Ecology
                                                Louisville, Kentucky
  The nature and quantities of hazardous materials and the
various technologies for treating them have grown rapidly in re-
cent years. The efficiency and cost-effectiveness of technologies
such  as  physical/chemical  treatment,  metal  precipitation,
biological treatment  and carbon adsorption, vary enormously
depending on the application requirements. In most instances,
these treatment processes have been utilized to treat large volumes
of mixed matrix wastes.
  Transferring these technologies for  effective use in ground-
water treatment, with relatively low flow volume and a range 'of
contaminants from a single substance to a complex mixture, re-
quires special considerations in the design and operation of  a
treatment program. The process  design for these  remedial pro-
grams must utilize the optimum mix of available technologies to
generate the highest degree of treatment at a reasonable cost.
  This paper evaluates the process design  requirements and
economic considerations for application of these processes to
groundwater  treatment. It  also examines the unique design and
operating conditions encountered in groundwater treatment. The
concept of "life-cycle"  design is addressed to demonstrate the ef-
ficiency of optimizing  the  particular attributes  of the different
technologies available.

  There is a  broad range of potentially applicable technologies
for the elimination of  hazardous materials.  However, for each
material a specific review must be performed to select the most
appropriate form of treatment. During this review, consideration
must be given to the type of the contamination, organic versus in-
organic; the extent of the contamination, surface soil contamina-
tion alone, vadose zone, groundwater  or any  combination of
these; also to be considered are the physical and chemical char-
acteristics of the contaminants, the relative concentration of the
contaminants and the interactive effects the contaminants may
have with each other. In most instances after careful evaluation, it
will be found that a mix of appropriate technologies can be imple-
mented to produce the most efficient, cost-effective solution to
treat the contamination.
  Most of the conventional treatment technologies are intended
for use on waterborne contaminants present in fairly constant
concentrations. One of the characteristics of most groundwater
contamination problems is their relatively short term nature (2-5
year treatment programs). These treatment programs initially will
be treating a relatively high concentration of contaminants. With
time,  however, the concentration  of  these contaminants will
decrease.  This concentration decrease is due to the removal of
chemicals as a result of the treatment, as well as the dilution effect
of clean groundwater passing through the zones of contamination.
  As contaminated water is withdrawn from the ground (in either
the aquifer or vadose zone) clean water will either be drawn in
from the surrounding areas of the aquifer or, in the case of the
vadose zone, clean water used for flushing the contaminants will
displace the chemical compounds present in the interstitial spaces.
The net effect is a continuous, sometimes dramatic, decrease in
contaminant concentrations. This changing chemical composition
creates difficult design requirements for remedial actions.  The
treatment programs implemented must be flexible enough to ac-
commodate the  high concentrations  initially encountered as well
as successfully treat the lower concentrations near the completion
of the program. In fact, most programs will see the highest con-
centrations of contaminants withdrawn over a relatively short
time period during the initiation of the treatment program. This is
a most important  consideration  as the efficiency  and  operating
costs of the treatment process will be seriously affected.

  There are several types of treatment processes which can be ap-
plied to the remediation of groundwater contamination. These
can be divided into three categories:  physical/chemical treatment
of inorganic constituents; physical/chemical treatment of organic
materials and biological treatment of organic compounds.  It is
important to understand the difference between the use of in situ
treatment as a technique for decontaminating groundwater and
soils and the use of process systems for the same purpose. When
using in  situ  techniques,  materials  (oxidants, microorganisms,
etc.) are introduced into the treatment area via wells, borings or
percolation beds. Under these conditions, there is  relatively little
control over the reactions or the environmental conditions under
which they are occurring. Analysis of the performance of these
techniques is difficult to ascertain and most certainly involves a
time lag between the start of a program and any result which may
be reached.
  In our approach to treating contaminated groundwaters and
soils, the contaminants and contaminated waters  are withdrawn
from the ground and treated in surface reactors. This is the case
for all the technologies discussed  in this paper. In most instances,
there is a requirement for pumping liquid into  or out  of the
ground. Since the surface reactors allow  a much greater degree of
control over the  process and better measurement of the results, we
find this to be much more appropriate for the treatment of  con-
taminated materials. In some cases, the combination of both tech-
niques will yield the most benefit and should be employed.

  The  inorganic compounds  most  frequently encountered in
groundwater contamination are  heavy  metals. These materials
                                                                              TREATMENT OF HAZARDOUS WASTES    157

usually  are not  particularly mobile in the environment. Most
metals exist as a cationic species and have a tendency to adsorb
onto soil particles. The characteristics of some soils also allow an
"ion exchange" process  to occur, thus  removing the metals from
groundwater passing through a formation.

Hydroxide Precipitation
  The metals which are found in extracted groundwater can be
treated most effectively through precipitation processes.  Most of
the heavy metals are insoluble as a metal hydroxide or as a metal
sulfide.  If they are not in a chelated form, the metals can quite
readily  be removed by  using pH adjustment and  creating the
metal hydroxide. Typically, the minimum  solubility of most
heavy metals is in the pH range of 7.5 to  9.5.
  The steps necessary to determine metal removals involve ex-
perimentally determining the solubility of a metal (or metals) at
various  pH levels by neutralizing a sample of the contaminated
water using lime or caustic soda. An insoluble precipitate will be
formed  which can be physically removed, and the soluble concen-
trations of the metal(s) can be directly measured at each pH level.
  In  a  full-scale treatment process,  the  groundwater  will  be
withdrawn and neutralized to the optimum pH level. The metal
hydroxide precipitate will be removed from the liquid stream via
flocculation and sedimentation. The clarified, metal-free effluent
then can be discharged.  In many instances, this process  must be
performed prior to processing for organic compound removal.

Sulfide  Precipitation
  The process of sulfide precipitation of metals works in a similar
fashion. The minimum  solubility of any  given species will vary
somewhat from that  exhibited  during hydroxide precipitation;
however, the soluble concentration of the metals  after sulfide
precipitation will be considerably lower, as much as two  or three
orders of magnitude lower than with hydroxides. For most treat-
ment program requirements, the levels  which can be achieved via
hydroxide precipitation are more than  adequate.

Chrome Removal
  The process design for the removal  of  hexavalent chrome re-
quires an additional step  prior to the  precipitation process.
Chrome,  in  the hexavalent  form, will  not  precipitate as a metal
hydroxide. It must first  be reduced to  the trivalent state using a
strong reducing agent such  as sulfur dioxide or sodium bisulfite
under acidic pH conditions. The pH  then can  be  raised using
either lime or  sodium hydroxide and precipitation of the metal
will be effected.

System Design
  The design criteria  used in these  treatment programs usually
will follow those which have  long been  applied  in the  metal
plating and processing industries. The systems applied to ground-
water treatment should be designed with ease of installation and
operation as major considerations. Once the treatment process is
developed and implemented, there is generally little in the way of
process variables which need to be controlled. Automatic pH con-
trol systems can be used to control neutralization; fiocculant feed
rates need to be monitored and adjusted  periodically, but once
these systems are on line, there usually is  little which can be ad-
justed. The use of lamella separators for  liquid-solid separation
can substantially reduce the space required for this operation and
allow easy installation and operation of the system.
  The use of metal precipitation is the first step in processing con-
taminated waters containing both heavy metals and organic com-
pounds.  It is important to remove  the metals at  this point to
achieve  and  maintain high  degrees of treatment efficiencies in
subsequent groundwater treatment processes.
Downhole Recovery
  The category of physical/chemical treatment of organic com-
pounds is a broad one. The characteristics of groundwater con-
tamination are such  that in many instances  the compounds are
present in very high concentrations or as a pure material floating
on or settling below the groundwater. The first type of treatment
process to investigate would be recovery of pure product. Every
effort should be made to remove as much pure product as pos-
sible  using  either  downhole  recovery  systems  or  surface
separators. The nature of most  organic materials renders them
only  partially soluble  in  water,  and  many  compounds will
physically separate from water quite readily. There are a number
of downhole systems now available for recovering hydrocarbon
products from both the top and the bottom of a contaminated
  The use of simple separation equipment at the surface also can
be implemented. This may consist of rope-type skimmers or even
separator tanks using tubes and plates. The cost of the treatment
program is substantially reduced when relatively pure product can
be recovered as opposed to having to process the material. In
many instances, the recovered product will have some economic
value and may partially offset some of the treatment costs.
  There are also physical/chemical treatment processes for use on
low concentrations of organics in the liquid stream. Most notable
among these are air stripping,  carbon adsorption, peroxide oxida-
tion  and UV/ozone  treatment. These are the traditional  tech-
niques for  removing  (or  destroying) organic constituents  from
contaminated groundwater.

Air Stripping
  Air  stripping  of  organic  compounds  uses the  volatility
characteristics of certain  organic  species, such as  chlorinated
solvents, to eliminate them from an  aqueous stream. The con-
taminated groundwater is pumped to the top of a column which
contains an inert packing. The water flow is downward from the
top of the column against a countercurrent air flow from the bot-
tom of the unit. The air exhausted from the column contains the
volatile organic compounds and is simply exhausted from the air
stripping column via  exhaust  ports at the top of the column.
  Air stripping is  an  effective method for removing low concen-
trations of highly volatile compounds. The units are quite inex-
pensive to operate and require a minimum of operating attention.
There are, however, potential problems associated with their use
in groundwater treatment programs. Certain  compounds do not
have particularly high volatility characteristics and do not readily
strip from water.  In  some instances, this problem can be over-
come by increasing the temperature of the contaminated water,
thus increasing the volatility  of the  organics, but heating can
become quite expensive.
  Another consideration is the potential for fouling the internal
surfaces of the air strippers. The groundwaters which are heated
in the  strippers  generally will  contain  bacteria capable of
degrading the compounds present. Under the environmental con-
ditions found in an air stripper, it is possible to create substantial
biological growth  on  the internal packing surfaces as well as on
the mist eliminating packing at the top of the  towers. In some in-
stances,  this growth  can  lead  to excessive  maintenance  re-
quirements for the operating units.
  Another  major consideration when evaluating the use of air
stripping is the ultimate fate of the stripped compounds.  In the
past,  these  compounds have merely been  vented  to the  at-
mosphere. This is not truly a treatment technique for the elimina-
tion  of the compounds but rather a relocation of the problem
from  the  water to the air. There currently is a move by the

regulatory agencies in many areas to substantially reduce or elim-
inate the air emissions from these units. This control by air pollu-
tion authorities will  mean  implementing  complex and  expensive
emission control technology on the air streams exiting these units.

Activated Carbon
  Carbon adsorption has been used to purify water for over 100
years. It is based upon the natural process of adsorption—that is
the natural affinity of a gas or liquid to be attracted to and held at
the surface of a solid. This attractive characteristic is due to the
surface tension of the solid holding the molecules in water at its
surface. Activated carbon particles are characterized by very high
surface area to volume ratios. The carbon particle has a structure
with a series of pores whose diameters decrease as the pores ex-
tend into the particle so that the small molecules will migrate fur-
ther into a carbon particle  than a large molecule.
  The adsorption of the molecule onto the carbon surface is a
physical process consisting of three steps: (1) diffusion of the
molecules through the liquid phase to the carbon particle; (2) dif-
fusion of the molecules through the macropores to the adsorption
site; and (3) adsorption  of the molecule  to  the site. The
characteristics of the adsorbate molecule will determine the rate
of adsorption and, ultimately, the efficiency of the process.
  Carbon adsorption has gained a very favorable reputation as a
means of removing  a broad range of organic compounds  from
water streams. It is quite effective at removing low level concen-
trations  of  organics, particularly  chlorinated solvents. The  ef-
ficiency of this process varies enormously, from as low as 3<7o to
as high as 25% utilization  efficiency. Activated carbon is an ex-
tremely effective method for polishing treated effluents to remove
the last traces of contaminants.

Hydrogen Peroxide
  Hydrogen peroxide is used in waste treatment applications as
an  oxidizing agent. It is a moderately strong oxidizer of organic
compounds and is  quite capable of oxidizing a broad range of
compounds at almost any concentration. The reaction involves
the formation of a free radical hydroxyl group. Hydrogen perox-
ide is a highly reactive chemical which reacts quite readily with
most organic materials (except saturated alkanes). Most reactions
involving the use of peroxides require the use  of an iron catalyst
to effect the reaction.
  Peroxides can be  difficult to handle from a safety perspective
and are by nature unstable. Also, the reactions involved are quite
exothermic and can  create  significant hazards when used to treat
higher concentrations (>500 mg/1) of organic concentrations.
The most appropriate use for peroxide oxidation may be for the
treatment of highly biologically refractory materials, highly inter-
mittent waste flows  or extremely low concentrations of organics
in a polishing operation.

UV/Ozone Treatment
  The last major category of physical/chemical treatment tech-
niques  for organic compounds is UV/ozone. UV/ozone  treat-
ment technology seems to  offer a great deal of potential for the
treatment of  low  concentrations  of halogenated solvents. The
ozone is the source of a highly reactive free radical oxygen species.
This highly  reactive  free radical readily cleaves the double bonds
found in aromatic and unsaturated aliphatic compounds.
  UV  light  increases the reactivity of the target molecules, par-
ticularly those bearing carbon-halogen bonds. The wavelength of
the ultraviolet light is in the absorption wavelength of this bond
and excites  the bond to the point of breaking and cleaving the
halide from the carbon atom.
  Some of  the  limitations of this technology are:  the cost  in
generating ozone; developing efficient UV light systems; reducing
the necessary contact times in the reactors and thus the number of
lamps  and the electricity to operate them; and improving the
transfer of the generated ozone gas into the water stream.
  The  combination of UV and ozone is an excellent method for
treating extremely low levels of organics in water. This technology
is still  in its development phase and has not been used in more
than a few pilot applications.  One outstanding characteristic of
this technology and of the biological processes is that they both
are destructive technologies, converting hazardous organic com-
pounds into non-hazardous mineralized products of carbon di-
oxide and water.

  The  use of the biological  process  in waste treatment involves
the creation  of a growth environment in which  bacteria are
capable of metabolizing  organic compounds for energy for cell
maintenance  and reproduction.  Essentially, this process is the
conversion of a soluble organic compound into  an  insoluble
organic (cell protein) which can be  physically  removed from  a
water stream. This technology is truly a destructive process, as the
end products of the biological process are carbon dioxide, water
and additional bacterial cells (sludge).
  In order to accomplish this reaction, it is important to provide a
proper  growth environment with  consideration of inorganic
nutrients (nitrogen and phosphorous), pH, dissolved oxygen (for
aerobic processes) and several other process  variables. The design
and operating conditions for a system of this type can be quite ex-
pensive using conventional activated sludge technology. Recent
developments in bioreactor and  process  technology have given
this process the opportunity to be one of the most  useful tech-
niques  for the treatment of hazardous materials.
  The use of biological treatment for degrading organic chemicals
has been in practice for some decades now.  The use  of this tech-
nology on hazardous materials, and in particular on groundwater
contaminants, has  not been  particularly  extensive until  very
Activated Sludge
  The  process  of activated  sludge  treatment  is an operator-
intensive  process which is  sensitive to variations   in  influent
characteristics and operating conditions and requires close, careful
operator attention. One of the key elements in activated sludge
treatment is  the  concentration  of   large  numbers  of
microorganisms in a reaction vessel that effectively use an organic
substance  for a  food source. In  order to concentrate these
organisms, liquid-solid separation must be  effectively and care-
fully controlled. The use  of fixed-film biological reactors instead
of the suspended growth activated sludge process has eliminated
the requirement for extensive operator attention and reduced the
critical nature of liquid-solid separation as  part of the process.

Fixed Films
  The  use of a fixed  medium surface allows  the  microbes  to
become  attached to  a  surface  over which their   food source
(substrate) passes. The organisms form a biological  film into
which  their substrate,  nutrients and oxygen  diffuse. In this
fashion, the organisms merely grow to a film thickness consistent
with the availability of substrate.  As the  concentration  of
substrate diminishes, the rate of  microbe growth decreases.
  In a submerged fixed-film bioreactor (rotating biological con-
tactor), the release  of  air in a  diffusion  system  beneath the
medium in the tank provides the means of  both oxygen transfer
and mixing energy. This mixing  energy creates a relatively high
velocity of air and  water moving past the  biological film. This
movement provides a means for shearing  the accumulated bio-
logical growth from the fixed-film and for preventing bridging of
biological growth and plugging within the medium.
                                                                                 TREATMENT OF HAZARDOUS WASTES     159


* \

                              ANAEROBIC LAYER
                  — INCREASING GROWTH	»

                           Figure 1
              Biological Film Growth and Sloughing
  This mechanism of sloughing (Fig. 1) of the biomass is essen-
tially self-regulating.  As the film becomes  thicker due  to  the
growth  of the microbes,  its ability  to  remain attached  to  the
medium is reduced. Since the velocity of the air and water mixture
remains constant,  the shear forces are constant. When the film
becomes  thick enough, the shear forces are greater than  the
adhesive forces and the film sloughs  off the  medium surface.
  The effective concentration of microorganisms within the reactor
tank would be the equivalent of 6,000-8,000 mg/1 of mixed liquor
suspended solids in an activated sludge system. These biotreaters
can be placed within a fairly small reactor volume and are quite
readily preassembled and installed at a treatment location.
  A minimum of operator expertise is required for these systems,
and  they are  becoming  increasingly  popular as  methods  of
destroying the hazardous materials. The biological process tends
to complete the degradation reaction, yielding carbon dioxide.
Perhaps 99%  of all known organics  can be effectively degraded
using biological treatment.

Slightly Contaminated Streams
  Recent  developments in  biological treatment technology now
enable low concentration waste streams to be biologically treated.
In the past, all biological reactors were designed using the growth
characteristics of the biological populations as the limiting factors
to their effectiveness.  It  was not considered possible to  use
biological processing on many wastes at a concentration contain-
ing less than 50 mg/1.  This limitation  is due to the nature of
biological growth dynamics, the effect of biological washout from
the reactors, loss of settling properties and other engineering con-
siderations. Low concentration reactors now have been developed
using acclimated populations for removal of low concentration
( <$ mg/1) contaminants to low fig/\ (1-3 /ig/1) effluent  levels.
These reactors are portable, require almost no operator attention
and appear capable of operating several months on the low in-
fluent concentrations before needing to be  reacclimated. This
technology probably will displace activated  carbon  in many in-
stances for the polishing of low level organic compounds.

Soil Treatment
  The use of  the biological process also has been  successfully
adapted to the treatment of contaminated soils. It is possible, using
the proper engineering designs, to create a growth environment
for microorganisms in a soil environment. This process requires the
same environmental conditions as the treatment of a wastewater
stream, but they can be created and maintained quite readily.
  The availability of commercial bacterial inocula has substan-
tially increased the ease with which the soils can be  treated,  and
these techniques offer substantial potential for eliminating a wide
range of soil contaminants such as pentachlorophenol, creosote,
hydrocarbons, pesticides and  many other types of compounds.
The economics of soil treatment are also very favorable, and once
again this type of treatment is truly a destructive technology.
  The current state-of-the-art  in groundwater  treatment technol-
ogies allows us to use these technologies now. However, the use of
these technologies as part of a remedial program must be carefully
scrutinized to ensure that they are consistent with the economics
and operating considerations of each individual program. For in-
stance, the use of a carbon adsorption  process for treatment of
organics  at  a  very high  concentration  (>100 mg/1) can create
more problems than it will solve. There are instances where highly
biodegradable compounds  present in  a  carbon system have cre-
ated high degrees of biological activity within the columns and
have subsequently generated substantial amounts of sulfides in
the carbon column effluent. These sulfides then required an addi-
tional peroxide process to eliminate the odor and corrosion prob-
lems associated with them.

  The economics of the various treatment technologies cover a
considerable range. The cost of any particular process can also vary
considerably depending on the degree of treatment required and the
time frame of the treatment program. The technologies ranked by
cost in descending order from most to least expensive are: peroxide
oxidation, activated carbon, biological and air stripping.
  The costs vary as a function of contaminant concentration; as a
rule the cost will be higher  for higher  concentrations (on a $/gal
treated basis); the efficiency of each process (i.e., an easily stripped
 compound) will cost  less to treat per gallon than a less volatile
compound. The duration of the treatment program also will have
a significant effect on the economics of treatment since a longer
term program will allow for amortization of the equipment over a
longer lifespai. Also,  some process equipment will have a higher
residual value ai the end of the program and can be sold or reused
in  other  applications.  The  following  example compares the
economics of treating a waste stream via  three different processes:
air stripping, carbon adsorption and biological/carbon adsorption.
Design Basis
    Flow = 5 gal/min
    Organic concentration =  300 mg/1 methyl ethyl ketone
    Operating period =  2  years

  The economic estimates  are based on the process achieving 1
mg/1 MEK and 100 ng/\ total other organics in the effluent. The
cost for all capital equipment, operations and maintenance are in-
cluded in this analysis.

    Carbon adsorption  =  $0.065/gal
    Biological/carbon = $0.042/gal
    Air stripping = $0.032/gal

  The economics involved  in the treatment of any  material can
vary enormously depending on the specific requirements  of the
application but,  in general, the above  figures are relatively ac-
curate. As the concentration of the contaminants to be treated in-
creases, the biological process becomes more economical. As the
volatility of the materials increases, the cost of air stripping im-
proves somewhat;  however, this cost  is difficult  to lower much
  Other economic  analyses have been performed on a  $/lb re-
moved basis and have  been reported to be as low as $0.40-0.50/lb
removed  biologically to as high as $7.50-$10.00/lb removed with
carbon adsorption. A true economic evaluation of the actual costs
can be performed only on a specific case-by-case basis.

  The most important consideration when  considering the ap-
propriate technology is to  optimize the cost advantage of each
process. As can be seen from the economics, there are many
possible combinations of technologies which can be put together
to produce a highly efficient, cost-effective treatment process.

  The treatment requirements  for  hazardous  materials  in  a
groundwater contamination problem are  unique  for  several

• Relatively low and/or variable flow conditions
• Varying influent concentrations
• Short time frame projects
• Ease of operation and simplicity of design
• Stringent discharge regulations
The design and operating requirements for  these treatment
systems have to reconcile all these factors and at the same time of-
fer an efficient cost-effective system. The most important con-
sideration in designing these systems is to insure the systems will
provide the necessary degree of treatment throughout the lifetime
of the project while adjusting to the changing influent conditions.
  Initially, the flow volume in these programs is generally quite
low. They rarely exceed 100-150 gal/min and quite often are in the
10-20 gal/min range. The actual flow over the course of the treat-
ment program may vary considerably due to  seasonal conditions,
the transmissivity characteristics of the aquifer and  percolation
rates of the overlying soils.
  The concentration of the contaminants will  also vary with a
general trend to decreasing concentrations. There  periodically
may be increased contaminant levels  as pockets of material are
freed, but the concentration will decrease over the life  of the treat-
ment program. It is important to  use a process design that will
cover the entire life-cycle of the project.
  The time frame  of most groundwater cleanup programs  is
generally between 2-5 years. Most  treatment systems for perma-
nent operation are  designed on a lifespan of 10-20  years. It  is
much easier to justify the costs of installing a major process on a
long-term system than on a relatively short-term system. It is most
important to minimize the amount of capital required to install a
short-term remedial project.
  The systems  being  used in these  treatment  situations are
generally unfamiliar to many of the industries and individuals in-
volved in  cleanup programs. There are many industries which,
until recently, merely discharged their wastewaters to public treat-
ment facilities and,  as a result, have limited  experience in waste-
water processing and the actual operation of wastewater treat-
ment facilities. In many instances,  there is a  minimal amount of
manpower available for operating a treatment system,  or the loca-
tions of the treatment facilities are  remote  and not frequently
  Taking all these factors into account, it becomes  apparent that
the design of a treatment facility to be used in these situations  is
indeed quite different than  a typical wastewater treatment plant.
It is important to optimize  the benefits of each technology. For
instance, it can be quite costly to provide a carbon adsorption
system to treat 50 gal/min of groundwater contaminated with 300
mg/1 of adsorbable solvent which will ultimately contain only a
few mg/1 of organics near the end  of the treatment.
  Preferable is a combined system using either air stripping or
biological  treatment before the carbon system to  reduce the
organic loading on the carbon and hence reduce the amount of
carbon which will be used. Both air stripping and biological treat-
ment are less expensive than the use of activated carbon  for the
removal of high concentrations of organics.
  In most  cases, using modular, life-cycle designs will tend to
lower the costs and increase the efficiency of the tretment process.
For instance, the use of two fixed-film bioreactors to handle a
given flow and loading is preferable over using one large unit. As
the concentrations decrease, one of the units can be taken off line
and the operating costs  and maintenance requirements will  be
decreased substantially. Similarly,  a carbon tank which is sized
for a certain flow and organic  concentration may become an
operational problem if the flow and concentrations in the influent
drop substantially. Problems with carbon treatment include: (1)
channeling  and carbon bed  may develop biological growth which
can cause anaerobic conditions and subsequent sulfide problems.
When properly designed for  life-cycle  operation, the treatment
systems can provide a highly efficient, cost-effective process.

Phenol Contaminated Groundwater Treatment
  An example of life-cycle design is the following treatment pro-
gram on the U.S. Gulf Coast. At the initiation of treatment, the
influent organic  concentration was approximately 1,300 mg/1 as
TOC with phenol concentration at 400 mg/1. Over the lifespan of
the treatment program, it is known that the concentration of the
influent stream will decrease substantially until  it is below 100
mg/1 TOC. It would  be  quite difficult to design and operate  a
system that functions as effectively on 1,300 mg/1 as on 100 mg/1.
Obviously,  activated  carbon  treatment  alone  would be pro-
hibitively expensive and would still require the disposal of large
quantities of spent carbon. The  use of an air stripping process
also is not possible since the organic compounds present are not
readily strippable. In order to meet all the needs of this particular
program, the  life-cycle design as  shown  in Figure  2 was im-

                       I 300 ma/I TOC

Aeration Aeration—

                         Blowers I
                       < 900 mq/1 TOC
                       < 300 mo/1 TOC
                       « 100 ma/1 TQC
                           Figure 2
  Life-Cycle Design for Phenol Contaminated Oroundwater Treatment
                                                                                 TREATMENT OF HAZARDOUS WASTES    161

  In this installation, the aeration tanks were converted storage
tanks which, when coupled with a clarifier, were used as an ac-
tivated sludge process. As the concentrations decreased only one
tank was used for the treatment program. Eventually, that tank
was also removed from service and the fixed-film biotreater was
the remaining biological unit in operation. Throughout the dura-
tion of this program, the process design and equipment used were
changed to meet  the prevailing conditions.
  The cost of operating the system and the attention required for
its operation also diminished with time. Throughout the design of
the treatment process, those technologies used were applied to
provide the most  economical operation. The activated sludge pro-
cess (using existing equipment) provided bulk organic removal;
the fixed-film unit provided additional treatment (below about
300 mg/1 TOC the activated sludge  process would not  be  par-
ticularly effective) to reduce the organic loading to the carbon ad-
sorption system;  and, ultimately, the carbon was used to remove
the lowest level of contaminants.
  The nature of this program required that the effluent  be treated
to a TOC concentration less than 18 mg/1 so that it could be used
for reinjection to force additional contaminated water toward the
withdrawal wells. Once biological activity was established in the
system, TOC was consistently biologically reduced by 75-85%.
Effluent phenol  levels were consistently  reduced by 99%. This
biological oxidation of organics substantially reduced the  amount
of carbon required to attain the 18 mg/1 TOC  effluent  require-
ment. The influent to the system has decreased to  less than 800
mg/1 TOC, and  parts of the activated sludge system have been
taken off line. There is  now a requirement for the removal of
arsenic from this particular waste. An arsenic precipitation system
is being added so  that the arsenic can be co-precipitated with iron.
This system is being  added at a minimal cost.
  The economics involved in this particular case were based on
the use of carbon adsorption alone as compared to the cost of us-
ing biological processing. Assuming  23,000 gal/day flow at a
TOC of 1,300 mg/1,  249  pounds of organics need to be removed
from the water. Using a carbon cost of $0.75/lb and an operating
efficiency of 1 Ib organics removed/10 Ib carbon, the  operating
cost for just carbon is $l,875/day. This cost does not include the
cost of capitalizing the equipment or other costs of operation such
as electricity and  personnel.
  The anticipated cost for the operation of a biological system
alone, including capitalization of the equipment, electricity, but
not personnel, is approximately $0.46/lb  of organic material
removed. This is  $115/day plus the cost of carbon  for polishing
the effluent (assuming 50 Ib/day TOC is removed via carbon pol-
ishing) of $370/day.  The total costs of operating this system are
$485/day. This system has been in operation for over two years,
resulting in a cost savings of $834,000. The actual costs of daily
operation should  continue to diminish over the remainder of the

Chemical Disposal Facility Leachale
  The optimization of several technologies in combination has
been investigated  to treat the leachate from a chemical disposal
facility.  It  has been  determined that a plume of high TOC
material was migrating off-site and required treatment. A  process
design using carbon  adsorption  alone and in conjunction with
biological treatment was evaluated in a pilot program.
  Initially, 300 gal of sample were collected from the well at the
site. Analysis revealed 2,000 mg/1 TC (to compensate  for  in-
organic  carbon, this  value was reduced by 10% to compute a
TOC value). N,N- dimethyl formamide and propionic acid were
specifically identified.
   A series of four carbon columns, each designed for a surface
loading of 2 gal/min ft2, processed the waste liquid. The utiliza-
tion efficiency of the carbon was found to be 25% (1 Ib organics
removed/4 Ib activated carbon used). The apparent attainable ef-
fluent level was 90 mg/1 as TC. The carbon was highly efficient in
its removal capabilities in this instance and was quite able to lower
the contaminant levels in the effluent. Even at this high efficien-
cy, the  cost for treatment in replacement carbon costs alone is
about $4/lb organics removed.
   A second  series of evaluation was performed to determine the
viability of using biological treatment in conjunction with the ac-
tivated  carbon in order to attain similar effluent levels at  a  re-
duced cost and, in addition, provide a destructive treatment of the
contaminant  compounds. A simple biological reactor  was de-
signed to provide degradation of the bulk of the organics present.
This reactor was followed by carbon adsorption for final polish-
ing and  removal of refractory organic compounds. The biological
reactor  was able to remove an average of 76% of the organic con-
centration of the  liquid. The carbon adsorption system removed
the remainder of the TOC down to a level of 40-70 mg/1 with the
carbon  system providing a utilization efficiency of about 16%.
The economics for the treatment of this waste stream break down
as follows:
Treatment System
Activated Carbon
   1 gal/min
  2 gal/min
Biological/Activated Carbon
   1 gal/min
  2 gal/min
(S/lb TOC removed)

These costs include both capital and operating expenses for the
treatment program.
  On the basis of this economic analysis and the needs of this par-
ticular  treatment program, a combination of  biological/carbon
treatment likely will be used at this facility.
  The adaption  of conventional treatment technologies for the
treatment of hazardous materials in groundwater situtations is
clearly possible. However, the use of these technologies requires a
design basis unique to each application as well as an economic
justification unique to each application. Each application must be
evaluated independently to  determine which technology or com-
bination  of technologies  is most appropriate for  producing the
desired level of treatment on a cost-effective basis.
Envirosphcre Co.,  "Evaluation  of systems  to stabilize waste piles or
deposits," Dec. 1983.
Nyer,  E.K., Groundwater Treatment Technology', Van Nostrand Rein-
hold Company, New York, NY,  1985.
Wagner, K.  and Kosin, Z., "In Situ Treatment," Proc. of the 6th Na-
tional  Conference on Management of Controlled Hazardous Waste Sites,
Washington, DC, Nov. 1985, 221-230.

                                  Update on the  PACT Process

                                                Harry W.  Heath, Jr.
                                  E.I. du Pont de Nemours & Company,  Inc.
                                                  Chambers Works
                                               Deepwater, New Jersey
  This paper is an updated discussion of a 40 million gal/day
plant which uses the  du Pont developed PACT* process to pro-
vide combined secondary/tertiary treatment to industrial waste-
  Reasons for selecting the PACT process are discussed. Special
features of the plant are reviewed including a 40 ton/day multiple
hearth furnace to regenerate powdered activated carbon and an
on-site, double-lined,  secure hazardous  waste landfill  where
primary sludge  from the plant is deposited. Operating data are
presented which demonstrate the effectiveness of the PACT pro-
cess including data on removal of priority pollutants.
  Because of under-utilization of the plant in recent years, a suc-
cessful outside wastewater treatment business has developed. This
aspect is discussed as an example of the financial, technical and
environmental advantages of using  a large,  advanced  treatment
facility to treat  industrial wastewaters from a variety of sources.

  This paper is an updated case study of a 40 million gal/day
design industrial wastewater treatment plant (WWTP); several
papers discussing the initial years of operation of this plant were
presented in the early 1980s. This WWTP was built to treat the
waste from du Font's large Chambers Works site located on the
Delaware River estuary in southern New Jersey. The Chambers
Works site produces  a large variety of organic intermediates,
mainly substituted aromatic compounds; tetraethyl lead; Freon
fluorocarbons;  textile  treating  chemicals;  fuel additives;  and
miscellaneous other products. During the initial years of WWTP
operation, the site also made dyes, dye intermediates,  sulfuric
acid and isocyanates. In addition, the WWTP treated acidic waste
from  du Font's  adjacent Carney's Point nitrocellulose plant,
since shut down.
  Because the WWTP was greatly under-utilized by 1983, the site
began  to actively  seek outside  aqueous wastes for treatment to
help reduce the  fixed cost burden of the WWTP on the remaining
operations on the site. The last section of the discussion reviews
various aspects  of this very successful venture.
  The PACT process was  invented  specifically to  treat the
original Chambers Works-Carney's Point waste stream. It added
powdered  activated carbon to a conventional  activated sludge
aerator and achieved a higher degree of treatment than could be
obtained with activated sludge alone. Table 1  summarizes the
composition of the waste stream expected to be treated during the
design stages of the WWTP.

*PACT is an acronym for Powdered Activated Carbon Treatment. The
PACT technology is now owned by Zimpro, Inc., which is making the
process available under license.  "PACT System" is  a registered
trademark of Zimpro, Inc.
                          Table 1
       Design Load (Average) for Chambers Works WWTP
Soluble BOD
Dissolved Organic Carbon
Total Dissolved Solids
Total Suspended Solids
26,400 gal/min
88,700 Ib/day (280 mg/1)
1000 APHA
454,000 Ib/day* (1430 mg/1)
65,000 Ib/day (205 mg/1)
2000 to 5000 mg/1
80,000 Ib/day** (258 mg/1)
*Acidity expressed as CaCo^ equivalent throughout this paper.
**lncluded 56,000 Ib/day by product solids from lime neutralization.

  This wastewater was salty, acidic  (pH between 1  and 2) and
highly colored. In terms of organic concentration, the wastewater
was only medium strength; because of the high flow, the absolute
pounds loading was very large. Similarly, the total amount of acid
to be neutralized was high. The loading of dissolved organics was
highly variable, both in terms of composition and amount, in part
due to the batch nature of many Chambers Works processes. The
organic compounds to  be  treated  included many not normally
amenable to biological treatment.
  The first and third sections of the discussion review the factors
leading to selection of the PACT process and the subsequent per-
formance of the plant. The second section is a description of the
equipment, including the multiple hearth regeneration furnace for
activated carbon, to our knowledge the only use of this equipment
for powdered carbon regeneration in the world; the double-lined,
secure hazardous waste landfill for storage of primary sludge is
also described.
  The PACT process developed from technical studies of treat-
ment processes for Chambers Works-Carney's Point wastewater.
Compared to a conventional activated sludge process, the PACT
process  with  a dosage  of  150 mg/1 virgin carbon at an 8-day
sludge age:
• Gave  consistent BOP removals of over 95%
• Increased DOC removal  from 62% to 85%
• Reduced color
• Minimized  foaming during aeration
• Improved sludge settling and filtration properties
• Protected the microorganisms from shock loadings of organics
  Data from typical continuous laboratory unit tests are shown in
Table 2. Table 3 shows similar data demonstrating the effect of
increasing carbon dose on system performance. The carbon used
in these tests  had an iodine number between 800 and 1,000. Sig-
                                                                              TREATMENT OF HAZARDOUS WASTES    163

nificantly different performance would have been obtained with
carbon of different quality,  although  any  PACT unit generally
will outperform a parallel activated sludge  unit. Note in Table 2
that color actually was increased by activated sludge treatment.
  By  1970, when design work on  the Chambers Works WWTP
started, it had become obvious that proposed limits on DOC and
color  would require some form of tertiary  treatment. Review of
available technology showed the  most promising alternatives to
PACT was granular carbon columns located either before or after
a conventional biological reactor.  These options were studied in

                           Table 2
          Typical Laboratory PACT Unit Performance*

Feed Concentration
  BOD5                     296 mg/l
  DOC                      203 mg/l
  Color                     1960 APHA
Effluent	        PACTJJnll       Biological Unit

BOD5: Cone.
BOD5: Removal
DOC: Cone.
DOC: Removal
Color: Cone.
Color: Removal
Operating Conditions
Sludge Age, days              8.0             8.0
Carbon Dose, mg/l          148                0
Temp. Range, °C           15-24            15-24
Aerator Hydraulic
  Res. Time (hr)              7.4             7.3
* Data for time period 1/22/77  thru 5/6/77.
                           Table 3
         Laboratory Data Showing Effect of Carbon Dose

11.5 mg/l
29.1 mg/l
395 APHA
26.1 mg/l
91. 2 mg/l
2280 APHA

Soluble BOD5, mg/l
BOD Removal
DOC, mg/l
DOC Removal
Color, APHA
Color Removal
25 mg/
100 mg/l
Data represent three months' operation of continuous, laboratory 7 5 I units. Temperature range
for all units was 18 to 25°C, and hydraulic resident lime was 8 hr
  In laboratory studies, neither process with carbon columns per-
formed as well as PACT. When carbon columns were placed after
the activated sludge unit, in the more conventional arrangement,
the biological  system constantly experienced upsets caused  by
shock loads of organics. With carbon columns before the biologi-
cal  unit, the  carbon was used inefficiently. Much of its com-
paratively expensive adsorption capacity was occupied by easily
biodegradable  compounds.  Further, as the composition of the
feed changed, it had a chromatographic effect on the carbon col-
umn. Thus, there  were  surges in the  concentration of specific
organics in the effluent from the column, and the biological unit
did not receive the expected protection from organic shocks.
  Investment costs  for the PACT system were  lower than for
either of the carbon column systems. Very generally, the invest-
ment difference was caused by the added expense for separate car-
bon columns as well  as the expense for twice the clarification
capacity in the biological portion of the carbon column systems.
Because of the superior settling properties of PACT  sludge, the
PACT system requires less  clarifier capacity than conventional ac-
tivated sludge systems.
  PACT process economics were compared with a conventional
activated sludge process for the  Chambers Works, even though
the latter was not a viable alternate in this case. Cost estimates for
PACT secondary/tertiary treatment vs. secondary treatment by a
conventional activated sludge plant  were within 10% of  each
other. The extra  PACT  investment for carbon handling and
regeneration facilities was mostly offset by the additional clarifier
capacity  and sludge disposal  facility needed  for the activated
sludge plant.
  In effect, the PACT process at Chambers Works was expected
to provide tertiary treatment quality effluent at a cost close to that
for secondary treatment alone. In  one piece of equipment, the
aerator, the PACT process destroyed most of the organic wastes
relatively cheaply by biological action, while using expensive car-
bon efficiently to remove only difficult to biodegrade or normally
non-biodegradable substances.
  Figure 1 is a flow diagram of the WWTP. It is easiest to con-
sider the plant  in two sections:  (1) primary section, where the
acidic wastewater is neutralized, most heavy metal removal occurs
and primary solids are settled,  filtered and disposed of in the se-
cure landfill; and (2) secondary/tertiary PACT treatment section
where color  and dissolved organics are removed.

Primary Treatment
  The acidic water is  neutralized with lime in a single stage
neutralization   using   three,   stirred  200,000-gaI  reactors
operated in  parallel. The WWTP receives powdered lime by
rail car or truck. Powdered lime is stored in 4 large silos and is
slaked to 8-10% concentration as needed. Lime slurry is fed to
the neutralizes by an automatic control system that maintains
pH at any preset level. The WWTP  currently consumes ap-
proximately  60  tons/day of dry lime to treat 100 tons/day
acid. The plant has enough capacity to treat 240 tons/day acid.
The highest monthly average load treated to date is 220 tons/-
  The neutralizes overflow to four primary clarifiers, each 1
million gallons  in size. The rectangular clarifiers operate in
parallel and are 230 ft  long, 55 ft wide and 12 ft deep. Solids in
the wastewater feed, as well as byproduct solids formed during
neutralization, settle and are removed as a 6-10% slurry in the
clarifier underflow. After filtration to a 45 to 50% solids cake,
the solids are hauled to the  secure landfill. Two or  three
(depending on the load) high-pressure (210 lb/in2 at end of cy-
cle) large recessed chamber filter presses make approximately 7
tons of wet press cake per 30 to 60  min cycle.
  The WWTP  currently  generates from  100,000 to 200,000
Ib/day (dry basis)  solids,  equivalent to 43,000 to 85,000
ydVyear of  landfill volume. The solids are mainly inorganic,
primarily calcium and magnesium salts, as well as silica com-
pounds  from river water  silt.  These solids also contain small
amounts of heavy metals and a variety of organic compounds.
The  solids are an inevitable byproduct of the primary treat-
ment process, and because they are primarily inorganic com-
pounds, the  only feasible disposal method is landfilling.

                            Figure 1
       Du Font's Chambers Works Wastewater Treatment Plant
                                  RR TRACK
                             11,  iRRrrc.Ki, i.
                             111111111.1 h
                         AB =
                         AS =
                          F -
                         FT -
                        LST =
                         NE -
                        RAS =
                       1°ST =
Acid Storage
Carbon Slurry Tank
Carbon Acid Wash Tank
Filter Press
Flow Splitter
Fuel Storage Tank
200 PSIO Filter Feed Pump
Lime Storage Silo
Lime Slurry Tank
Neutralization Tank
8,000 GPM Waste Water Feed Pump
Recycle Activated Sludge Pump
Primary Sludge Hold Tank
                                        1,000,000 GAL
  A diagram of the landfill is shown in Figure 2. It is located on
the Chambers Works site, and primary sludge is hauled there by
truck.  A  double  liner  of  a  chlorosulfonated  polyethylene
material, Hypalon, covers the entire bottom and part of the sides
of the landfill. Collection pipes between the liners serve as leak
detectors and drain to sumps outside the landfill. The sumps are
sampled on a regular basis for any signs of contamination. Above
the top liner is a similar collection system for leachate which is
pumped back to the WWTP. The landfill ultimately will become a
70-ft high pyramid with a 15-acre base.
  As outer sections of the landfill are filled, the edges and top are
covered with a 2-ft layer of essentially impervious, permeability of
less than 1  X 10-7 cm/sec, clay and 12 in. of top soil. Newer sec-
tions of the landfill also may be required to have a Hypalon liner
under the clay cap.
  A controlling factor in the  operation of the primary filter
presses is that the filter cake must meet rigid soil stability criteria.
These solids have  to  support the weight of heavy earth-moving
equipment. Further, the allowable slope of the landfill sides,  and
hence the volumetric  capacity of the landfill,  is closely regulated
to assure a large safety factor against any possible slippage along
the slopes.
  As backup protection,  there are 26 monitor wells located
around the  landfill; and in the very unlikely event  of  a double
liner failure, interceptor wells would be utilized to  prevent  any
harm to the environment while the problem was corrected.
PACT Secondary/Tertiary Treatment
  PACT  treatment starts  at the  flow  splitter  which  feeds
neutralized primary effluent to the aerators. The activated carbon
is added here as a 1 Ib/gal aqueous  slurry.  The  three, 185-ft
diameter, 4,000,000 gal aerators operate in parallel and provide 5
to 9 hr hydraulic resident time.
  Air  is used  to  suspend the mixed  liquor suspended solids
(MLSS) which are an approximate 50/50 mix of biomass and ac-
tivated carbon.  Initially,  there were approximately 1150 static
mixers in each aerator to improve  oxygen transfer. These aerators
have since been removed, leaving  just air holes on the bottom of
the  injection air laterals. The BOD loading to the WWTP was
never high enough to require the  added gas transfer claimed for
the  mixers. At low BOD loadings more air was required than
needed for the bacterial metabolism alone; the extra air flow was
needed to keep the MLSS well mixed,  as the mixers appeared to
hinder gross solids circulation.
  After aeration, water  is  fed  through two  parallel, 172-ft
diameter,  2,500,000 gal  clarifiers.  Feed  enters through a
centerwell, flows through a flocculating zone and then beneath a
ring into  the annular area bounded by the clarifier  wall. The
treated wastewater overflows  into a circumferential trough and
then to the plant effluent trench.  It goes to a settling basin where
it mixes with approximately 1 Vi times  its volume of non-contact
cooling water before discharge to the Delaware River estuary.
                                                                                   TREATMENT OF HAZARDOUS WASTES    165

                                                                                 V TOP SOIL

                           TOP LINER
                                 BOTTOM LINER
• c

                                                          Figure 2
                                                       Secure Landfill
   Rim drive bridges support rakes which push settled solids to the
 center of the clarifiers; from there the solids flow via underground
 lines to  the recycle activated sludge (RAS)  pumps. Two 78-in.
 diameter screw pumps of 10,000 to 12,000 gal/min capacity each
 are provided to pump the clarifier underflow or RAS back to the
 aerators. Solids may be wasted from either the RAS stream or the
 more dilute aerator discharge.
   The WWTP can regenerate activated carbon from the wasted
 MLSS. When the plant  is  regenerating carbon, the wasted slurry
 is pumped through a thickener. Underflow from the thickener  at
 5 to  12%  solids  concentration is pumped to a 250 lb/in2 filter
 press. The 64-in.  diameter press has 112 recessed chambers 1 in.
 deep. Each press  cycle produces approximately 12,000 Ib of 38  to
 45%  solids press cake. This press is  similar  to the two other
 presses used in primary service. In fact, one  of the other two
 presses is a "swing" press which also can be  used to filter PACT
 sludge. When not regenerating carbon, all three presses may be
 used in primary service.  The filtered PACT sludge is conveyed by
 a series of drag flite, screw and belt conveyors to the top of a 26-ft
 diameter, 40-ft high multiple hearth regeneration furnace.
   There are five hearths in the furnace. Hot gases flow up the fur-
 nace  countercurrent to  the  sludge. The furnace has a rotating
 centershaft with  two  rabble arms extending over  each  hearth.
 Filter cake drops  onto the top hearth and is raked from hearth to
 hearth down  through the furnace.  From  the fifth (and last)
 hearth, the incandescent regenerated carbon drops into a water-
 filled quench tank.
  Some of the furnace's unique design features reduce upward
gas velocity and minimize  entrainment of carbon particles. The
general principle of operation is that water is evaporated on the
first and  second hearths at  gas temperatures of 480 to 700°C. On
the third hearth,  biological solids  and adsorbed organics arc
volatilized by 750 to 870 °C gas. On the bottom two hearths, at gas
temperatures of 870 to  1020°C, the powdered carbon is thermally
regenerated in the presence of water vapor.
  Off-gas from the  first hearth passes through a two-stage water
scrubbing system and then to an  oil-fired  afterburner which
destroys odors and any organics at temperatures of 650 to 760°C.
  Continuous water flow to the product quench tank is adjusted
to give a product slurry concentration of approximately 5%. The
regenerated carbon  is washed in acid to remove  inorganic  ash.
Washings from this  step are  returned to primary treatment where
removed ash precipitates and is trapped in the primary sludge. The
washed,  regenerated carbon  slurry is transferred  to a  60,000-gal
storage tank prior to recycle to the aerators. Three of these tanks
are provided to hold virgin and/or regenerative carbon.
  In recent years, organic loading to the WWTP  has decreased;
therefore, less activated carbon is needed. It has become more
economical to use only virgin carbon on a throw away basis. Cur-
rently, waste  MLSS are returned to the influent steam to primary
treatment. The carbon and  biomass are removed in underflow
from the primary clarifiers, making up about  10% of the filtered
solids deposited  in the  secure landfill.
  During the first years of operation, the PACT process exceeded
expectations in terms of BOD and color removals and approx-
imately equalled design in DOC removal. Data for early years are
shown in Table 4; in this period, the system was  most heavily
loaded.  Performance in subsequent years has been comparable,
but inlet DOC and BOD concentration have been significantly
lower. The  flow rate also has decreased to approximately 20,000

                           Table 4
             PACT System Performance, 1978 & 1979
Flow (gal/min)
 Inlet Soluble BOD5, mg/1
 Effluent Soluble BOD5, mg/1
 Removal (%)
 Inlet Color, APHA
 Effluent Color, APHA
 Removal (%)
 Aerator MLSS (mg/1)
 Sludge Age (days)
 Carbon Dose (mg/1)
  Virgin (°7o)
  Regenerated (%)







 *The low ratio of regenerated to virgin carbon reflects start up problems with the filtration and
 regeneration systems. In 1980 and 1981, the ratio was 70% regenerated/30% virgin at similar
 total carbon dose.
   The initial BOD load was less than expected because the mix of
 organics to be treated shifted and became less biodegradable than
 anticipated. The design ratio of BOD/DOC was 1.3 to 1.4, which
 was the then current ratio in Chambers Works waste. However,
 extensive waste reduction  efforts prior to WWTP  startup re-
 moved a disproportionate amount of biodegradables. As can be
 seen in Table 4, the BOD/DOC ratio in 1978 and 1979 was ap-
 proximately 1.0.
   The activated carbon in the aerator allowed the WWTP to treat
 this difficult waste. One result of the lower BOD load was that the
 minimum air flow to the aerators was set by  the amount needed to
 keep the biomass and carbon well suspended. This was about 13
 ftVmin per static mixer.  It had been expected that minimum air
 flow would be set by the need to maintain  a minimum  dissolved
 oxygen concentration  of 0.5 mg/1 in the aerator.  As was men-
 tioned earlier, this was the reason the static  mixers were removed
 from the aerators in 1981. This allowed the air flow to be reduced
 from 15,000 ftVmin per aerator to 8,000 ftVmin or lower, while
 still maintaining 0.5 mg/1 oxygen levels.
   The influent was more highly colored than was assumed during
 design. However, color removal efficiency was better than antici-
 pated, and the site was able to meet permit standards for color.
   Average DOC removal was equal to design, so effluent quality
 was slighly better than predicted, reflecting  the lower  influent
   In terms of meeting NPDES permit limits, variation in perfor-
 mance is often more critical than long term average. Most permits
 specify a daily maximum limit as well as a monthly average, and
 consistently meeting the  daily maximum is often the more de-
 manding task for the process. The consistency of the PACT pro-
 cess at Chambers Works is  outstanding.  Figures 3 and 4 are
 histograms showing the distribution of feed and effluent DOC
 and soluble BOD respectively in 1978 and 1979. Note the narrow
 distribution of effluent concentration, particularly of BOD.
  Analyses for the removal of U.S.  EPA-designated  priority
pollutants are in Table 5. Volatile organics  and acid  extractable
compounds generally  are removed very well; base  neutral com-
pounds are removed with a lesser amount of  success. Some metals
removal occurs across PACT treatment, although the process is
not designed for this purpose.
  N        = 721
  Avg      = 170
  Std D6V  =  506
  Max      = 278
  Mm      =  80
                                                                                       90% CL
                                                                                     Log Normal
110      160
  mg/llter DOC


, 90% CL
/Log Normal
y /
k /
i 	 1 	 1 	
10 60 110 160
mg/liter DOC
stci Dev

— i 	

= 721
= 312
= 99
= 106
= 108

	 1 	 1

                 68%      76%      84%      92%     100%
                  % DOC Removal

                         Figure 3
            DOC Histogram (1978 and 1979 Data)
                                      A major reason for selection of PACT was the protection af-
                                    forded the biosystem from upsets caused by shock loadings  of
                                    organic wastes. Because of the large number of batch processes at
                                    Chambers Works, it was not impossible for the WWTP to see
                                    20,000 to 50,000-lb slugs of various organics in a 1- to 4-hr period.
                                    The biosystem handled documented spills  to  the WWTP  of
                                    50,000 Ibs of an aromatic diamine and  30,000 Ib of o-toluidine
                                    with no ill effects. No equalization is provided before the WWTP
                                    because pilot plant studies showed no significant improvement in
                                    performance with equalization.
                                      The demise of dyes manufacture on the  site  ended a major
                                    source of organic shock loads. However, the increasing amounts
                                    of trucked in outside wastes actually have made the overall prob-
                                    lem more severe. A 5,000 gal truck is discharged into the 20,000
                                    gal/min influent stream within 30 min. The PACT process has
                                    tolerated the resultant shock loads when wastes with relative high
                                    DOC content are received.
                                      Initially, the WWTP  operated  at higher sludge ages, 20 to  60
                                    days, than the 8 to 15 days specified in design. Organic removals
                                    were  improved at higher sludge  age,  with the  result that the
                                    WWTP operated at an average carbon dose of 120 mg/1.  This
                                    compares with a design  dose  of 170 mg/1. Because  of lower
                                    organic loadings, since 1983 the plant has operated at sludge ages
                                    of 8 to 30 days with virgin carbon doses of 10 to 50 mg/1.
                                      Because of the initial high solids concentrations in the aerators,
                                    10,000 to 30,000  mg/1,  the secondary clarifiers were  heavily
                                    overloaded. They were designed for a solids flux rate of 63 Ib/day
                                                                                 TREATMENT OF HAZARDOUS WASTES     167

                   200      300      400
                      mg/llter BOD
             90% CL
           Log Normal
                      Std  Dev
200      300      400
  mg/llter BOD

Std Dev

2 60
= 543
= 961
= 36
= 999
= 66


                                90% CL
                   % BOD Removal
£. , ** ~ LJ HI JL U
                           Figure 4
              BOD Histogram (1978 and 1979 Data)
                                                                                           Table 5
                                                                          Priority Pollutant Removals Across PACT System
Carbon Tetrachlorlde
Methyl Chloride
1,1,1-Trlchloroe thane




                                              The values in Table 5 arc averages of data whtch sho» significant day-to-day variations in
                                             removals. Results were obtained by GC MS analyses of plant sample* and are subject to analytical
                                             error. The figures represent only estimates of average percent remcnab for these materials.
ft2  at an  underflow  concentration  of 3.6%.  They  actually
operated at solids  flux  rates from  200 to 350  Ib/day  ft2 at
underflow concentrations of 4 to 8%. The carbon in the sludge
greatly enhanced the thickening abilities  of the clarifiers.
  Part of the reason for such heavy solic4, loading were initial
problems with the PACT sludge filter prf ^s and, surprisingly, the
conveying system to carry  filtered slurge into the regeneration
furnace.  During 1978  and  1979,  the inability  to  remove solids
from the liquid train by normal wastage methods forced solids to
exit the system as total suspended solids (TSS) in the treated ef-
fluent overflowing the secondary clarifiers. Effluent TSS  aver-
aged about 100 mg/1 vs.  a  design level of 30 mg/1. These solids
settled in the polishing basin and ultimately were dredged to the
primary treatment section of the WWTP.
  From 1978 through March 1982 when  the regeneration furnace
was  shut down, there was a steady series  of mechanical im-
provements in the PACT  sludge filtration and conveying systems.
Not all portions of the system reached design rates, and some added
and/or redesigned equipment had to be installed.
  The expected improvement in sludge filterability with the addi-
tion  of carbon to the  biomass was achieved. Filter cake solids
averaged 45% vs. a predicted 35% to 38% level.  Press capacity
was below forecast because  of: (1) lower dosage of carbon needed
with longer sludge age operation, and hence less  carbon in the
sludge to aid filtration, and (2) operating problems with the filter
press, particularly with cloth life, cloth  changing procedure and
high pressure feed pumps. The original steel press plates are being
                                              replaced with lighter weight fiberglas or  plastic plates that are
                                              easier to dress and are expected to dramatically improve PACT
                                              sludge filtration if the furnace resumes operation.
                                                As anticipated, the regeneration furnace incurred  relatively
                                              high  maintenance costs.  There have been several failures of  he
                                              brick hearths.  The combination of the above, plus numerous
                                              design problems  of the type often associated with  a pioneering
                                              chemical processing venture, limited furnace in time to just over
                                              50%  in the first two years of operation. In time of furnace opera-
                                              tion improved from 55% in 1978-79 to 66% in 1980-81. Much of
                                              the furnace downtime was caused by problems with the feed con-
                                              veyors and not with the multiple hearth furnace itself.
                                                When operating, the furnace produced  an  average of 24,000
                                              Ib/day regenerated carbon in 1978-79. In 1980-81, production was
                                              raised to 38,000 Ib/day.  Recovery of virgin  carbon properties
                                              averaged just over 60% for the high activity carbons needed for
                                              Chambers Works waste.  Weight  yield recovery was over 80%
                                              after acid washing. As expected, the acid washing step was needed
                                              to prevent buildup of inorganic ash on the carbon with successive
                                              regenerations. However, during the last months of furnace opera-
                                              tion, ash levels increased in regenerated carbon. Revisions to the
                                              acid  washing system  to better remove iron and silica were sche-
                                              duled when the decision  was made to shut down the furnace.
                                                Despite  significant equipment improvements,  it  appears the
                                              solids train will  continue  to incur relatively  high  maintenance
                                              costs. As a result, alternate technologies  have been considered.

One which shows great theoretical promise is wet air oxidation. It
would eliminate both filtration and sludge handling and replace
the furnace with a high pressure reactor. Whether this high pres-
sure/lower temperature  liquid phase treatment can  regenerate
Chambers Works spent carbon to the quality necessary to treat
this type of waste is not certain. Experience at other locations sug-
gests it may be able to do so, however there has not been suffi-
cient justification yet to do the large scale pilot studies necessary
to test the process for Chambers Works.

  Operating experience at the Chambers Works WWTP over the
past 4 years has demonstrated  the  advantages of transporting
aqueous wastes to a large, central, advanced wastewater facility
for treatment.  In the early 1980s,  du  Pont began aggressively
seeking outside wastes for treatment as it became apparent that
the WWTP was going to be under-utilized because of declining
production  at  the site.  A  marketing study showed this  large
PACT plant had unique advantages  in treating outside wastes.
  Because of the great volume of wastewater being treated, the
WWTP had essentially  an  unlimited  volumetric  capacity for
trucked waste. Four tankers simultaneously discharging 5,000 gal
each in only 20 min increase influent flow less than 5%. Highly
acidic wastes are no problem. The WWTP has treated 2,500 gal of
almost pure chlorosulfonic acid in only 4 hr by direct discharge in-
to the influent trench with no release of fumes.
  The tremendous dilution provided by over 20,000 gal/min flow
is a major protection against organic shock loads from a tank
truck  discharge.  The  PACT system was  selected specifically
because of its ability to handle varying loads of industrial waste.
Therefore, this plant  is almost  ideally  suited to treat batch
discharges of aqueous wastes containing organics, especially those
which would pass through or be inhibitory to a conventional
biological treatment system.
  The Chambers Works  already had logistics in place to handle a
large  number of truck deliveries  each day and excellent interstate
highways leading to the site. This was vital because of the volume
of incoming tank trucks. Deliveries of over 50 trucks a day are
   Because of other activities on the  site, there were "in-house"
environmental  and technical groups to support this activity.  A
major technical effort is needed to evaluate wastes proposed for
treatment both in terms  of their suitability to the PACT process
and the estimated costs  of treatment. A strict protocol must be
established to approve a waste for  treatment,  and subsequent
deliveries  must be monitored closely to  assure that the waste ac-
tually received is what was contracted for. Because of the amount
of hazardous waste treated, there are frequent contacts with state
and federal regulatory  authorities regarding operation of the
WWTP. A state inspector is at the site every 1 to 2 weeks, and it is
important that a competent environmental staff is available  to
represent the plant.
  From a regulator's point of view, it  is advantageous to treat
wastes at a central site. The disposition  of a large number of in-
dividual wastes can be monitored conveniently by a single inspec-
tor. The single, large plant usually will be better controlled than a
variety of small separate facilities, often operated by those less
than expert in waste treatment. The quality of the treated effluent
can be assured  by a detailed analytical evaluation of a single ef-
fluent stream; such analyses might be impractical and prohibitive-
ly costly for a large number of individual sites.
  Another advantage of  this site is  the receiving water. Chambers
Works main outfall discharges into an estuary, so salt content of
the stream is not critical. There  are no downstream withdrawals
for drinking water purposes. The volume of flow at the Delaware
River estuary is immense, and so for a given absolute quantity of
residual waste the effluent concentration will be very low. Any
aqueous  waste treatment facility  will discharge  some  residual
material. It is an obvious advantage if the receiving water flow is
high; then the resultant concentrations of any residual pollutants
are low, and there is no adverse impact on downstream water use.
  Since 1980, the outside waste business has grown rapidly. Cur-
rently, over 15% of the organic load to the WWTP is from non-
Chambers  Works  wastes.  Over  50% of the outside  waste
originates in  New  Jersey; 35%  comes  from  Pennsylvania,
Delaware and Maryland. However, it has been cost-effective to
ship  aqueous wastes surprisingly long distances. There are routine
contracts originating  as  far away as  Maine, Florida and Col-
orado. The freight-logical area has been extended by offering rail
car service to transport wastes, and the potential exists for barge
  Wastes from a variety of sources have been treated. Some are
listed in Table 6. No one type of waste makes up more than 20%
of the total.

                           Table 6
     Typical Industrial  Wastewaters Treated at Chambers Works

• Tank Truck & Tank Car Washings
• Pharmaceutical Wastes
• Water from Oil-Water Separation Processes
• Textile Treating Wastes
• Metals Treating Wastes
• Bio-sludges from Industrial Wastewater Treatment Plants
• Electronics Industry Waste
• Landfill Leachates
• Lagoon Cleanups
• Miscellaneous Chemical Process Wastes
• Latex Wastes
• Chemical Cleaning Rinse Waters
• Waste Acids or Bases
• Food Processing Wastes
• Paint & Dye Wastes
  Initially, treatment of outside waste was limited to streams
which could be discharged directly into the WWTP influent as
rapidly as they flowed from the truck. This limitation precluded
certain aqueous wastes. For instance, wastes containing sulfide or
cyanide could not be accepted because of the possibility of releas-
ing toxic H2S or HCN gas when the waste mixed into the acid in-
  The WWTP has  limits on most of the  heavy metals. It  is
capable of removing those present in the original wastewater feed
to meet  NPDES permit limits. Initially, some outside wastes
which were evaluated had to be  rejected because they contained
metals in amounts such that permit limits might have been ex-
ceeded had the waste been added directly to the WWTP influent.
  Recently,  heavy metals pretreatment facilities  have been in-
stalled and are now being expanded to allow the WWTP to accept
aqueous  waste streams containing higher  amounts of  heavy
metals. Similarly, the plant is developing treating techniques for
waste containing higher concentrations of sulfide  and cyanide.
  Facilities also are being installed for oil-water separation, with
the aqueous phase to be discharged to the WWTP and the oil to
be handled separately. Other expansions are being studied which
involve   separate  pretreatments  which  themselves  generate
aqueous  waste that could be handled by the WWTP. This might
give du Pont  a competitive advantage over other stand alone
                                                                                 TREATMENT OF HAZARDOUS WASTES    169

  This outside waste treatment has been successful enough that
du Pont is expanding the business beyond the Chambers Works
WWTP.  In  recognition  of their  expanded charter,  the outside
waste business group was recently  renamed du Pont Environmen-
tal Services and is actively seeking other waste treatment oppor-


• PACT is an excellent, advanced treatment process for many
  wastewaters containing normally non-biodegradable or slowly
  biodegradable materials. The process can treat highly colored
  wastes. At du Font's  Chambers Works chemical  complex, a
  PACT plant removes over 96% of the BOD and over 80% of
  the DOC from 30 to 40 million gal/day of wastewater.
 • A PACT process is resistant to toxic upset, can tolerate shock
  loads of organic compounds and is tolerant of rapid changes in
  waste composition.
 • The presence of carbon in the mixed  liquor suspended solids in
  a  PACT system dramatically increases settling in  the clarifier
  and improves filtration rate. Solids concentrations of approxi-
  mately 40% have been achieved with PACT sludge in a conven-
  tional pressure  filter.
• Conventional wastewater treatment equipment can be used for
  a PACT system with a minimum amount of design changes.
• Powdered activated carbon has  been successfully regenerated
  from PACT sludges in a multiple hearth regeneration furnace
  at a rate of 20 tons/day. Maintenance costs for the entire fil-
  tration, conveying and regeneration unit are high, and to date
  the best system in time has been about 65%.
• The Chambers Works PACT plant is successfully treating a
  large variety of off plant  wastes delivered by tank  truck or
  railroad car. Over 30% of the current organic load to this plant
  is from outside wastes. Particular, demonstrated advantages of
  a large central wastewater treatment are:
  (1) An advanced treatment process is used.
  (2) High  quality "in house"  high quality technical and en-
     vironmental staffs can be afforded.
  (3) It is sited on optimum receiving waters.
  (4) It is  easier for  regulatory  authorities to monitor the  dis-
     posal of the wastes.
  (5) It can  treat  aqueous, organic waste  streams  which  might
     upset conventional municipal plants.

                         Treatment  of  PCB-Contaminated  Soil
                               in a  Circulating Bed Combustor

                                               D.D. Jensen, Ph.D.
                                                    D.T. Young
                                              GA Technologies Inc.
                                      Process & Energy Systems Division
                                              San Diego, California
  A trial burn of PCB-contaminated soils was completed in GA
Technologies' 16-in. Pilot Plant  Circulating  Bed Combustor
(CBC).  More than 4000 Ib  of soil  containing 1% PCB were
treated in three identical 4-hr runs at  1800°F. The results showed
excellent compliance with TSCA requirements. Destruction and
removal efficiencies  (DREs) were  greater  than 99.9999%, and
PCB in combustor ash  was  less than 200 ppb. No chlorinated
dioxins or furans were detected in the stack gas, bed ash or fly
ash. In addition, no significant concentrations of other Products
of Incomplete Combustion (PICs) were detected.  Combustion
efficiencies were  greater than  99.9%,  with CO concentrations
less than 50 ppm and NOX concentrations less than 75 ppm. Par-
ticulate emissions were generally below 0.08 grain/dscf, and HC1
emissions were maintained below 4.0 Ib/hr  by introducing lime-
stone directly into the combustor.
  These results demonstrate that the CBC is an environmentally
acceptable means of treating contaminated  soil containing PCB
and other organic wastes. In addition, the high thermal efficiency
of the  CBC, the absence of afterburners or scrubbers  and  the
use of simple feed systems make CBC treatment competitive with
soil removal and transport to landfills and other potential treat-
ment/disposal options.

  Of the many hazardous chemicals found in waste sites around
the country, perhaps none has received the scrutiny conferred on
polychlorinated biphenyls (PCBs). This group of  209 synthetic
chlorinated organic compounds found wide use as a dielectric
fluid in utility transformers and capacitors and as a high-tempera-
ture heat transfer medium.' However, because of their exception-
al resistance to degradation in the biosphere and apparent toxic-
ity, the manufacture and sale of PCBs  were banned in 1976  for
virtually all purposes. The control,  treatment and disposal of
PCBs  were  mandated   by  TSCA and currently  are  handled
through the U.S. EPA's Office of Toxic Substances.
  Until recently, it has been common practice to remove PCB-
contaminated soils for burial in a secured landfill. However, this
option  is becoming less desirable since landfill  costs are escalat-
ing, the number of available landfill sites has decreased and gen-
erators or potential responsible parties (PRPs) retain the liability
associated with the contaminated soil, even in a secured landfill.
Treatment of PCB-contaminated soil by incineration in the CBC
can eliminate or significantly reduce the potential liability of gen-
erators or PRPs at a cost competitive with current landfill prices.
  CBC units are designed to burn a wide variety of fuels such as
coal, peat, wood, municipal wastes  and oil while treating haz-
ardous  wastes. Over 25 units are operating or under construction
worldwide. Three units are currently in operation in the United
States. In 1983 GA began concentrating its efforts on the applica-
tion of GBC technology to  incineration of hazardous wastes.
Table 1 presents examples of wastes that have been burned in the
CBC.  It was the successful treatment of this diversity of wastes
that provided assurance that PCBs could be destroyed in a CBC
at a lower temperature than used in conventional incinerators.

                         Table 1
 Circulating Bed Test Results1 for the Destruction of Hazardous Wastes
Carbon Tetrachlorlde
Aromatic Nltrlle
Efficiency, J
solid >99.9999
Capture, %



1 .7

1 .7
1 Results obtained in GA pilot plant CBC.

  The CBC is a new generation of incinerator that uses high
velocity air to entrain circulating solids in a highly turbulent com-
bustion  zone. This design allows combustion along the entire
length of the reaction zone. Because of its high thermal efficien-
cy, the CBC is ideally suited to treat low heat content feed, in-
cluding contaminated soil. Figure 1 shows the major components
of a CBC for soil treatment.
  Soil is introduced into the  combustor loop at the loop seal
where it immediately contacts  hot recirculating soil from the hot
cyclone. Hazardous materials  adhering to soil are rapidly heated
when introduced into the loop  and continue to be exposed to high
temperatures throughout their residence time in the CBC. Upon
entering the combustor, high velocity air (14 to 20 ft/sec) en-
trains the circulating soil which travels upward through the com-
bustor into the hot cyclone. Retention  times in the combustor
range from 2 sec for gases to approximately 30 min  for larger
feed materials (^1.0 in.).
  The cyclone separates the combustion gases from the hot solids
which are returned to the  combustion chamber via  a  proprie-
tary non-mechanical seal. Hot flue gases and fly ash pass through
a convective gas cooler  and on to a baghouse filter where fly
ash is removed. Filtered flue gas then exhausts to the atmosphere.
Heavier  particles of purified soil remaining in the combustor
lower bed are removed slowly by a water-cooled ash conveyor
                                                                            TREATMENT OF HAZARDOUS WASTES    171

                                                          Figure I
                                     Schematic Flow Diagram of Circulating Bed Combustor for
                                                       Soil Treatment
  As a consequence of the highly turbulent combustion zone,
temperatures around the entire combustion loop (combustion
chamber, hot cyclone, return leg) are uniform to within  ±SO°F.
The uniform low temperature and high solids turbulence in the
CBC also help avoid ash slagging encountered in other types of
  Acid gases formed during destruction reactions are rapidly cap-
tured by limestone added directly into the combustor. The reac-
tion of limestone and HC1, released during PCB incineration,
forms dry, benign calcium chloride. The rapid combustion and
quick neutralization of the  acid gases within the combustion
chamber eliminate the need for afterburners and add-on scrub-
bers to complete destruction and acid gas capture. Emissions of
CO and NO, are controlled to low levels by excellent mixing, rela-
tively low temperatures (14SO to 1800°F) and staged combustion
achieved by injecting secondary air at higher locations in the com-
bustor.  Because of its efficient combustion and highly turbulent
mixing, the CBC is capable of attaining required DREs for  both
hazardous  wastes (99.99%) and  toxic  wastes (99.9999%) at
temperatures below those used  in  conventional  incinerators
(typically > 2000 °F).

  A variety of requirements are imposed  prior to and during a
PCB trial burn.' The key target of a trial burn is to ensure that
PCB DREs are ^99.9999%  at the operating conditions chosen
for the incinerator. In addition, the concentration of PCB in ash
from the unit must not exceed 2 ppm. The potential formation of
PICs also is carefully evaluated, with particular attention given
to polychlorinated dibenzo-p-dioxins (PCDDs)  and polychlori-
nated dibenzo-p-furans (PCDFs). The combustion efficiency of
the unit must be » 99.9fo,  and particulate emission must not ex-
ceed 0.08 grain/dscf.
                     Figure 2
Pilot Plant CBC and Feed Preparation/Handling Equipment

   The CBC trial burn was carried out in GA's 16-in. pilot plant
 unit shown in Figure 2. This is the smallest CBC offered by GA
 for commercial application. The glovebox in the foreground was
 used to prepare contaminated soil prior to transport to the feed
 system and CBC shown behind the glovebox.  Soil treated in the
 test was obtained from a former chemical processing site known
 to contain pockets  of PCB up to 6000 ppm as well as other
 organic and inorganic wastes. To ensure that the CBC would be
 permitted to treat all likely site concentrations of PCB,  uncon-
 taminated soil from the site was "spiked" with liquid  PCB to
 10,000 ppm. Spiking was carried out by blending a 50:50 com-
 mercial mixture of PCB "1248" and trichlorobenzene with a rib-
 bon blender in 1000 Ib lots. Approximately 4000 Ib of soil  were
 spiked for the three  burns  required by  the  TSCA trial burn
   While the CBC was maintained at 1800T using natural gas
 as the auxiliary fuel, several barrels of clean  soil from  the site
 were treated in the combustion system prior to the addition of
 spiked soil. During this time, all  operating parameters and sys-
 tem components  were confirmed to be in the required operating
 ranges. Process parameters monitored included:
 • Temperature around the loop
 • Pressure drop across the loop
 • Soil feed rate
 • Primary air flow
 • Secondary air flow
 • Loop seal air flow
 • Total air flow
 • Methane flow
 • CO concentration
 • CO2 concentration
 • Excess oxygen level
 • NOX concentration

   Spiked soil was pneumatically  transported  to a bunker  and
 screw  feeder. Soil  feeding,  limestone addition and stack gas
 monitoring were  started simultaneously. A U.S. EPA Modified
 Method 5 sample train3 was used to sample stack gas emissions.

                           Table 2
          PCB Trial Burn Operational Data and Test Results
Teat Duration, hr
Operating Temperature, °F
Soil Feed Rate, Ib/hr
Total Soil Feed, Ib
PCB Concentration in Feed, ppm
PCB Concentration
' Bed Ash, ppm
Fly Ash, ppm
Dioxin/Furan Concentration
Stack Gas, pp
Bed Ash, ppm
Fly Ash, ppm
Combustion Efficiency, J
Acid Gas Release, Ib/hr
Particulate Emissions,
Excess Oxygen, J
CO , ppm
C02. >
Requirement 1
-1 1
11 ,000
>99.9999 99.999995

<2 0.0035
<2 0.066

>99.9 99.91
<1.0 0.16
<0.08 0.095(b)
>3.0 7.9
Test Number
99.999981 99






 In addition, a separate Volatile Organic Sampling Train (VOST)4
 was used to sample for volatile organic PICs. Feed, bed ash and
 fly ash samples also were gathered throughout the test (see Fig. 1
 for sample port locations). Three identical tests of spiked soil (4
 hrs each) were carried out over two days in late May 1985. Each
 test was observed and/or audited by U.S. EPA personnel or rep-
 resentatives. All feed, ash and stack gas samples were subsequent-
 ly analyzed for PCBs, PCDDs and  PCDFs. Fly ash, bed ash and
 stack gas samples also were analyzed for other PICs (both vola-
 tile and semivolatile). Stack gases were analyzed for fly ash and
 chloride release as well.

  Table 2 presents a summary of the trial burn operational data
 and test  results gathered  during the  tests. Near-identical con-
 ditions were maintained for each test. In each case, PCB DREs
 were well in excess of the  U.S. EPA-required 99.9999%.  PCB
 concentration  in the bed ash and the fly ash  was well below the
 required 2 ppm and, in fact, did not exceed 200 ppb. No PCDDs
 or PCDFs were detected in the stack gas, bed ash or fly ash. Com-
 bustion efficiencies were greater than 99.9%, and acid gas release
 was well below the required 4 Ib/hr.  Paniculate emissions  were
 generally less than the required 0.08 grain/dscf. Only the grain
loading from the  first test, obtained from a 2-hr makeup test
 after the completion of Tests 1 through 3, showed a value slight-
 ly higher  than  the limit. This higher than normal value is attrib-
uted to off-normal process conditions for the baghouse (i.e.,
excessive blowback air pressures along with  a higher-than-normal
number of blowback cycles). Nitrogen oxides and CO levels re-
mained low as a result of the staged combustion utilized in the
CBC and  the  relatively low combustion temperature (1800T).
                                            FLUE GAS
                                                                 ASH REMOVAL
ND  Not detected.
(^Derived from 2-hr makeup test.
                         Figure 3
      Isometric of Site-Assembled Circulating Bed Combustor
                                                                               TREATMENT OF HAZARDOUS WASTES    173

These results demonstrate that the CBC is an effective means of
destroying PCBs contained in a soil matrix without the need for
high temperatures, afterburners or wet scrubbers. In particular,
the absence of undesirable combustion byproducts (i.e., dioxins
and furans) is important in ensuring that  effective treatment of
soil can be accomplished in an environmentally acceptable man-
   These results  confirm the design of GA's transportable CBC
shown in Figure 3.  The combustion and all other plant  com-
ponents are designed as modular units which can be transported
by truck  or rail. These units are assembled at the site into an
operating unit  in 4 to 6 weeks. The major components of this
CBC plant include the combustor loop, feed system, pollution
control and air  induction equipment. GA's 36-in. transportable
CBC is capable  of processing up to 18,000 Ib/hr of dry soil on a
24-hr basis requiring an operating crew of only two persons per
shift. Soil treatment costs may be as low as  $100/ton at a large
site. For  smaller sites or sites having unique treatment require-
ments, costs may approach $400/ton.
   The results of the PCB soil trial burn in GA's CBC demonstrate
 compliance with  TSCA requirements: stack  emissions are well
 within regulatory requirements and bed ash and fly ash contain
 PCBs  well below  the  regulatory  requirements. The superior
 thermal efficiency,  high throughput and small  staffing require-
 ments  of the CBC  provide a soil treatment option that is cost-
 competitive with landfill disposal while at the same time reducing
 overall liability of the generator or PRP.

I.  SCS Engineers, Inc.,  "PCB Disposal Manual," Electric  Power Re-
   search Institute, Palo Alto, CA, Report No. CS-4098, June 1985.
2.  "Polychlorinated Biphenyls (PCBs) Manufacture, Processing, Distri-
   bution in  Commerce and Use Prohibition," 40 CFR 761.70.
3.  "Test Methods  for Evaluating Solid  Waste," U.S. EPA Report
   SW-846, 2nd Edition,  1984.
4.  "Proposed Sampling  and Analytical  Methodologies for Addition to
   Test  Methods for Evaluating Solid Waste: Physical/Chemical
   Methods  (SW-846,  2nd  Edition)." U.S. EPA Report PB85-103026,

                        Treatment of Hazardous  Waste  Leachate

                                                  Judy L. McArdle
                                               Michael M. Arozarena
                                            William E.  Gallagher, P.E.
                                                PEI Associates, Inc.
                                                  Cincinnati, Ohio
                                                Edward J.  Opatken
                                     U.S. Environmental Protection Agency
                              Hazardous Waste Engineering Research Laboratory
                                                  Cincinnati, Ohio
  The nature of hazardous waste leachate varies greatly from site
to site and over time. Thus, leachate treatment systems must be
designed on a case-by-case basis. This paper presents profiles of
seven unit treatment operations which can be applied in varying
combinations  to effect the  desired degree of contaminant re-
moval: (1) equalization (to dampen influent flow and concentra-
tion fluctuations); (2) air stripping (to remove ammonia and vola-
tile organic compounds); (3) precipitation/flocculation/sedimen-
tation (to remove soluble heavy metals and  suspended solids);
(4)  neutralization (to adjust pH); (5) activated sludge treatment
(to remove aerobically biodegradable organic matter); (6) carbon
adsorption (to remove dissolved organics and certain inorgan-
ics); and (7) reverse osmosis  (to remove dissolved inorganics and
high molecular weight organics). These technology profiles em-
phasize process  applicability and limitations, treatment effective-
ness and capital and operating costs.

  Leachate is generated by the percolation of water (from precip-
itation, groundwater flow or liquid wastes) through a waste dis-
posal site. The leachate can be expected to contain soluble toxic
components of the waste as well as soluble chemical and biochem-
ical reaction products. Leachate collection systems are designed
to collect this contaminated  liquid and to channel it away from
the disposal site before it can contaminate the surrounding soil,
groundwater or surface water. The collected leachate then can be
treated to reduce the level of contaminants prior to its discharge
into a stream, to groundwater  recharge or to a municipal  or in-
dustrial wastewater treatment system.
  The composition and volume of leachate vary highly from site
to site and over time. Factors contributing to this variability in-
clude: the nature of the waste;  the age of the land disposal unit;
the amount of precipitation; and the porosity, permeability and
adsorption characteristics of the soil. Table 1 presents leachate
characterization data from three land disposal sites.
  The high strength and widely varying nature of hazardous
waste leachate complicate its treatment.  Under contract to  the
U.S. EPA's Hazardous Waste Engineering Research Laboratory,
PEI Associates, Inc., Cincinnati, Ohio, is preparing a compre-
hensive handbook to guide in the planning and design of  leach-
ate treatment systems. Technology profiles emphasizing process
applicability and limitations, treatment effectiveness  and capital
and operating costs are being developed for both demonstrated
and potential technologies. Six highly developed unit treatment
operations (equalization, air stripping, precipitation/flocculation/
sedimentation, neutralization, activated sludge tretment and car-
bon adsorption) and one potentially applicable technology (re-
verse osmosis) were selected for presentation here from the more
than 20 technologies to be profiled in the handbook.
                          Table 1
    Leachate Characterization Data: Concentration of Contaminants
Biochemical oxygen demand
Chemical oxygen demand (COD)
Total organic carbon
Total suspended solids
Total dissolved solids
pH, s.u.
Alkalinity (as CaCOj)
Hardness (as CaCOj)
Total Kjeldahl nitrogen
Total phosphorus (as P)
Phosphate (as P)
Acid Pits
<0. 1-14.0
<0. 1-3.0
<0. 001-0. 012
aReference 1.
bReference 2.
cReference 3.
dNR = not reported.
eAll data except pH are in mg/L
                                                                             TREATMENT OF HAZARDOUS WASTES   175

                                     FLOATING AERATOR
                          Figure 1
                  Equalization Basin Geometry


  The objective of equalization is to dampen influent flow and
concentration fluctuations  to improve performance of  down-
stream processes.  Because the composition and volume of leach-
ate from a given landfill fluctuate with time, equalization should
be  considered in  the planning and design of all leachate treat-
ment facilities.
  Equalization basins may be of concrete, steel or (lined) earthen
construction.  With in-line  equalization, the entire  daily flow
passes through the basin; with off-line  equalization, only  that
portion of the flow outside of an acceptable concentration range
is diverted into a basin. When the  goal is equalization  of the
pollutant mass loading (as in the treatment of hazardous waste
leachate), the in-line arrangement must  be  used. Typical basin
geometry is illustrated in Figure 1.
  Mixing and aeration of equalization basins should be provided
to prevent both the deposition of solids  and the onset of septic
conditions.  As with any process  using  aeration, volatile com-
ponents can be stripped from the wastewater, which results in air
emissions of potentially toxic substances.
  Table 2 presents capital and operating and maintenance (O&M)
equalization costs. These costs are based on the use of a concrete
basin with a 24-hour retention time. In general,  capital costs of
equalization systems are moderately high because of the size of
the basins. For large basins, power costs for operating the aera-
tion system contribute significantly to the  overall O&M costs.

                           Table 2
                   Leachate Treatment Costs*
                        (1985 dollars)
Unit operation
Air stripping
Neutral 1zat1on
Activated sludge
Carbon adsorption
Reverse osmosis
25 gal/mln
50 gal/m1n
100 gal/min
bS/lOOOgal based on 300 day/yr operation.
Air Stripping
  Air stripping is a mass-transfer operation that uses forced air to
remove pollutants from a liquid phase. Applications of this pro-
cess include the removal of ammonia from wastewater and, more
recently, the removal  of volatile organic contaminants  from
drinking water. In the treatment  of leachate, air stripping can
be used ahead of biological processes (to reduce toxic concen-
trations of ammonia) or ahead of adsorption processes (to reduce
the organic loading and thereby extend the life of the sorbent).
  Among the various aeration  devices that can  be used in air
stripping are diffused aerators,  mechanical aerators, spray tow-
ers and countercurrent  flow packed towers. Packed towers pro-
vide  the greatest gas-liquid interfacial area for mass transfer but
the shortest contact time. Removal efficiencies are dependent on
the height of the tower and the air-to-liquid ratio.
  Figure 2  is a schematic diagram of a packed-tower air strip-
ping unit. The unit consists of a vertical column  with randomly
dumped packing on  a  packing  support. Common packings in-
clude the Raschig ring,  Berl saddle, Intalox saddle and Pall ring.
Liquid is distributed  uniformly on the top of the packing with
sprays  or distribution trays and  flows downward by gravity. Air
is blown upward through  the packing and flows countercurrently
to the  descending liquid. Depending on the concentration and
volatility of the contaminants, air-to-water volumetric ratios may
range from 10 to 1  up to  300 to I.' Air stripping  of ammonia is
carried out at elevated pH  (above 11).
                                      AIR PLUS VOLATILES






                                                                                          Figure 2
                                                                 Schematic Diagram of Air Stripping in a Countercurrent Packed Tower
  The volatile organics or ammonia stripped from the leachate
during treatment in a packed tower subsequently may be removed
from the off-gas to prevent air emissions of toxic compounds
or, possibly, for economic recovery. It also may be necessary to
pretreat the leachate for removal of suspended solids and dis-
solved metals that are oxidized to an insoluble form when aerated
to prevent fouling of the packed bed.
  Capital and O&M costs for air stripping, presented in Table 2,
are based on the use of a countercurrent packed tower and in-
clude influent pumping and an ammonia recovery section. Costs
for labor, materials and power comprise the O&M costs of this
system.  Because air requirements,  and hence  power  require-
ments, increase at  lower temperatures,  operation of packed
towers at temperatures below freezing may be cost-prohibitive.

   The most common method of removing suspended solids and
 soluble  heavy metals  (including arsenic,  lead,  cadmium  and
 chromium) from leachate is by combined precipitation/floccula-
 tion/sedimentation. Precipitation involves the addition of chem-
 icals to leachate to form insoluble precipitates from soluble spe-
 cies. Flocculation promotes agglomeration of suspended solids,
 which makes them easier to remove.  Sedimentation is the removal
 of suspended particles by gravity settling.
   The processes of precipitation, flocculation and sedimentation
 can be carried out in separate basins, as shown in Figure 3, or in a
 single basin (e.g., an upflow solids-contact reactor-clarifier) with
 separate zones for each process. Precipitation requires rapid mix-
 ing  (10 to 60 sec) to disperse the chemical, whereas flocculation
 requires slow and gentle mixing (15 to 30 min) to promote particle
 contact. Ancillary equipment requirements for these processes in-
 clude mixers/paddles, chemical storage tanks and chemical feed
 pumps. Packaged plants are available for low flow rates (10,000
 gal/day to 2 million gal/day), but  these may require extensive
 modifications to enable them to handle the high level  of precipi-
 tated solids characteristic of hazardous waste leachates.
                            M    M





                         FLOCCULATION CHAMBER
                                           SEDIMENTATION TANK
                           Figure 3
    Schematic Diagram of Precipitation/Flocculation/Sedimentation
   Metals can be precipitated from leachate as hydroxides or sul-
 fides through the addition of lime or ferric sulfide. Hydroxide
 precipitation with lime at high pH is the most commonly used
 method. The presence of complexing agents such as ammonia or
 cyanide  in the  leachate, however, will  inhibit precipitation  of
 some heavy metals as hydroxides. Flocculants such as alum, ferric
 chloride or polyelectrolytes often are added to the rapid-mix tank
 along with the  chemical precipitant  to  reduce repulsive forces
 between particles and  bring  about  particle  aggregation and
   Precipitation/flocculation produces  large amounts  of wet
 sludge that may be hazardous because of its heavy metals con-
 tents. This sludge, which is settled in  a sedimentation unit, must
 be treated further (e.g.,  by dewatering or fixation/stabilization)
 before it is iandfilled.
   Fluctuating leachate quality requires frequent jar  testing to de-
 termine appropriate  chemicaraosages and removal efficiencies.
 Available data indicate that precipitation/flocculation/sedimen-
 tation can provide good removal of suspended solids (80 to 90%)
 and  moderate removal of BOD5 (40  to  70%) and  COD (30  to
 60%),' The additional treatment required will depend on effluent
 discharge limitations.
  Table  2 presents costs of a packaged precipitation/floccula-
tion/sedimentation system that includes a granular filtration unit.
Operation of these systems is labor-intensive, as reflected by the
moderately high  O&M costs.
   Neutralization involves the addition of an acid or a base to an
 aqueous waste stream to adjust its pH to the desired level (usual-
 ly between 6.0 and 9.0). Neutralization may be required as a pre-
 treatment step to optimize the performance of downstream treat-
 ment, to protect  pH-sensitive processes (particularly biological
 teatment operations) or as a final step to meet effluent criteria.
 The technology is inexpensive, highly developed and widely used
 in the treatment of hazardous waste leachate.
   Typically, neutralization is carried out in completely mixed
 corrosion-resistant tanks as shown in Figure 4. The tanks may be
 operated in batch or continuous mode; however, the latter is only
 suitable for flow rates greater than about 100,000 gal/day. The
 neutralization process usually is controlled automatically by feed-
 back, feedforward or multimode controllers. Common bases used
 for neutralization include lime, calcium hydroxide, caustic,  soda
 ash  and ammonium hydroxide; common  acids include sulfuric
 acid, hydrochloric acid and nitric acid. Reagent selection is based
 on cost, speed of reaction and the reaction byproducts formed.
   Equalization usually precedes neutralization in the treatment
 process train to control fluctuations in influent flow and  con-
 centration. The neutralization process can produce heat and toxic
 gases (ammonia, hydrogen sulfide and hydrogen cyanide) if the
 chemicals are not properly added to and mixed with the waste.
 Treatment to remove precipitated solids (flocculation/sedimen-
 tation or filtration)  often is required after neutralization. If the
 removed solids are hazardous, secure disposal will be required.
   Table 2 presents costs of a neutralization system consisting  of
 an agitated tank with a 3-min retention time, a metering pump
 for acid or caustic and a pH control loop and valve. Both capital
 and O&M costs are low in comparison with other unit treatment
                     pH CONTROLLER
                                                                                                                • MIXER
                                                                                                                 	" EFFLUENT
                              MIXING TANK
                           Figure 4
               Schematic Diagram of Neutralization

Activated Sludge
  Activated sludge is a biological treatment process in which an
active mass of microorganisms (bacteria,  protozoa, rotifers and
fungi) is used to  stabilize biodegradable organic matter under
aerobic conditions. Activated sludge can reduce concentrations of
a wide variety of organic  compounds, including many toxic and
hazardous compounds. It is widely used in municipal and indus-
trial wastewater treatment applications (typical BODj concentra-
tion of 200 mg/1) and can effectively treat leachate with organic
concentrations one to two  orders  of magnitude higher (up to
10,000 mg/1 BOD5."
  Conventional activated  sludge systems,  illustrated in Figure 5,
include  two stages. The first stage  involves aeration of the waste
in an open tank and maintenance of an active biomass. Aeration
is accomplished by mechanical-surface, diffused-air or sparged-
turbine aerators. The second stage entails separation of the solids
in a secondary clarifier. A  portion of the solids is recycled back to
the aeration basin to maintain the desired concentration of organ-
isms, and the remaining portion is wasted.
                                                                                 TREATMENT OF HAZARDOUS WASTES     177


, 	 I
                           Figure 5
        Schematic Diagram of Conventional Activated Sludge
  Variations of the conventional activated  sludge system have
been developed to provide greater tolerance  for shock loadings,
to improve  sludge settling characteristics and to achieve higher
BOD5  removals.  Process modifications  include complete-mix,
step aeration, extended aeration, contact stabilization and the use
of pure oxygen.
  Because of the sensitivity of biological systems,  pretreatment
requirements for high-strength leachate are extensive and may in-
clude equalization to buffer hydraulic and  organic load varia-
tions;  sedimentation/flotation to remove suspended solids, oil
and grease;  neutralization to adjust the pH to near neutral; and
nutrient  addition to provide adequate levels of nitrogen, phos-
phorus and  trace elements. Acclimation of the biological system
to the influent waste stream is necessary prior to full-scale opera-
tion of the  process. The presence of refractory or biologically
toxic compounds (e.g., ammonia or heavy metals) in the leachate
may necessitate the use of other physical/chemical processes in
conjunction with biological treatment.
  Process residuals from activated sludge treatment of leachate
include waste activated sludge, which may be high in metals and
refractory organics,  and air  emissions of volatile organic com-
pounds that are stripped from the waste during aeration.
  In general, biological processes such as activated  sludge treat-
ment  are the most cost-effective means  for removing organics
from high-strength leachates. Table  2 presents costs of a conven-
tional activated sludge system consisting of an aeration tank with
a 6-hr retention time and a mixed liquor suspended solids con-
centration of 2000 mg/1, a secondary  clarifier and a  sludge re-
cycle pump. The O&M costs of activated sludge treatment  are
comprised principally of labor and materials.

Carbon Adsorption
  Carbon adsorption is a separation technique for removal of dis-
solved organics and certain inorganics (e.g., cyanide and chrom-
ium) from aqueous waste streams.  This  well-developed process
has numerous  full-scale applications including treatment  of
domestic and industrial wastewaters,  cleanup of spilled  haz-
ardous wastes. In leachate treatment applications, carbon adsorp-
tion can  be  used  in the pretreatment,  intermediate or polishing
  Treatment of leachate by carbon adsorption  involves passing
the waste stream through beds of granular activated carbon. Acti-
vated  carbon is an amorphous form of carbon  characterized by
a large internal surface area. Contaminants are adsorbed from the
waste onto the carbon  surface by physical and  chemical forces.
When the adsorptive capacity of the carbon has been reached,
regeneration or disposal of the spent carbon is required.
  Most carbon adsorption systems use cylindrical pressure vessels
arranged in series or parallel and operated in  a downflow, up-
flow or pulsed-bed mode. The downflow mode can handle a high-
er concentration of influent suspended solids than can the other
modes; however, frequent backwashing of the carbon beds may
be required to prevent excessive pressure drop. A schematic dia-
gram of an alternating, two-column, downflow system is pre-
sented in Figure 6.
                                                  BACKWASH PLUS
                           Figure 6
     Schematic Diagram of Alternating, Two-Column, Downflow,
                   Carbon Adsorption System
  Carbon adsorption  is applicable  to  the treatment of many
toxic and refractory organics; consequently, it often is used in
combination with biological treatment. Because the carbon beds
are susceptible to clogging, leachate must be pretreated  to re-
move high concentrations of suspended solids, oil and grease. In
general, influent concentrations  are limited to 10,000 mg/1 total
organic carbon,  1,000  mg/1 total inorganics,  2,000 mg/1  sus-
pended solids (downflow),  50 mg/1  suspended solids (upflow)
and 10 mg/1 oil and grease.' With proper pretreatment, removal
efficiencies greater than 99"% can be obtained.'
  Table 2 presents capital and O&M costs of a carbon adsorp-
tion system. The  O&M costs of this system are governed by the
organic loading, which is assumed to be 10 Ib organics per 100 Ib
of carbon. These costs do not include the costs of disposal or re-
generation of the spent carbon.  For plants using less than 200
Ib/day of carbon (or,  equivalently, for plants treating less than
800,000  gal/day  of leachate),  on-site  regeneration  of carbon
probably  is not economical.' Most leachate treatment facilities
will fall into this range.

Reverse Osmosis
   Reverse osmosis is  a membrane separation  technique which
primarily is utilized for the demineralization of water. Poten-
tially, this technique could be used as the final polishing step in
the treatment of leachate to separate dissolved solids [inorganic
salts and high (>120) molecular weight organics] from  secon-
dary treatment effluent. With current technology, reverse osmosis
is not practical for the  treatment of raw leachate because of rapid
fouling of the membrane and poor selectivity for low molecular
weight organics.'
   In reverse osmosis, water is separated  from dissolved solids in
solution by filtering it through  a semipermeable membrane at
pressures of 200  to 1200 Ib/in1.' Basic components of a reverse
osmosis unit are  the membrane,  a membrane support structure,
a containing vessel and a high-pressure pump; these are illus-
trated schematically in Figure 7.
   All  commercially available membranes are structured asym-
metrically, with a  thin (0.1 to 1.0 /un)  dense  surface ("skin")
supported by a porous substructure.  This design promotes high

water transport across the membrane.8 Cellulose acetate and aro-
matic polyamides are common membrane materials.
  Membrane support structures are of four types: tubular, spiral-
wound, hollow-fiber and plate-and-frame. The tubular configur-
ation, which consists of a rigid-walled porous tube lined on the
inside with  a membrane, is recommended for use  with waste-
water effluents.6 Reverse osmosis units can be arranged in parallel
(to increase hydraulic capacity) or in series (to effect the desired
degree of separation).
  For optimal performance, feed water to the reverse osmosis
unit must be pretreated  to remove gross amounts of solids and to
prevent fouling by scaling or biological growth inside the device.
In addition, compounds that are incompatible with the mem-
brane must be removed during pretreatment. Because the com-
position and strength of hazardous waste leachates vary widely,
pilot testing will be necessary to determine pretreatment require-
ments and separation performance. A major disadvantage of this
technology is that it produces a highly concentrated reject stream
that requires further treatment or disposal.
  Table 2 presents costs of a reverse osmosis unit operating at
800 lb/in2. The O&M costs of this unit do not include the costs of
disposing of the concentrated reject  stream. Regardless, reverse
osmosis is an extremely expensive technology and should be con-
sidered  only for leachate treatment applications when an effluent
of extremely high quality is required.

                            PRESSURE  VESSEL
                                                   -»• EFFLUENT
                                        'SEMI PERMEABLE
                     REJECT STREAM
                           Figure 7
              Schematic Diagram of Reverse Osmosis
  The highly variable, complex nature of hazardous waste leach-
ate often requires that several  unit treatment operations be
applied  in combination to effect  the  desired degree of contami-
nant removal. Equalization is used to buffer influent flow and
concentration fluctuations  and to improve the performance of
more costly downstream operations. Air  stripping,  neutraliza-
tion and chemical precipitation often are used ahead of biolog-
ical treatment (activated sludge) to remove biologically toxic com-
pounds  such as ammonia and heavy metals and to adjust the pH
to near  neutral. Carbon adsorption and reverse osmosis are used
as polishing steps to remove refractory organics and to meet final
effluent criteria. Unit treatment costs range from $0.07/1000 gal
for neutralization to $2.10/1000 gal for reverse osmosis.
1.  Shuckrow, A.J.,  Pajak, A.P. and Touhill, C.J.,  "Management of
   Hazardous Waste Leachate," U.S.  EPA  Publication No. SW-871,
2.  Copa,  W.M., et al.,  "Powdered Activated  Carbon  Treatment
   (PACT™) of Leachate From the Stringfellow Quarry," Proc. of the
   Eleventh Annual  Research Symposium on Land Disposal, Remedial
   Action, Incineration and Treatment of  Hazardous  Waste, EPA-
   600/9-85-028, Sept. 1985, 52-65.
3.  Steiner, R.L., Keenan, J.D.  and Fungaroli, A.A., "Demonstrating
   Leachate Treatment: Report on a Full-Scale Operating Plant,"  U.S.
   EPA Publication No. SW-758, 1979.
4.  U.S. EPA, Handbook, Remedial Action at  Waste Disposal Sites,
   EPA-625/6-82-006, 1982.
5.  Metcalf & Eddy,  Inc., Briefing: "Technologies Applicable  to  Haz-
   ardous Waste," prepared for the U.S. EPA Hazardous Waste Engi-
   neering Research Laboratory, Cincinnati, OH, May 1985.
6.  Metcalf & Eddy, Inc., Waste-water Engineering: Treatment, Disposal,
   Reuse, 2nded., McGraw-Hill, Inc., New York, NY, 1979.
7.  Chian, E.S.K. and DeWalle, F.B.,  "Evaluation of Leachate Treat-
   ment, Vol. II: Biological and Physical-Chemical Processes," EPA-
   600/2-77-186b, 1977.
8.  Applegate, L.E., "Membrane Separation Processes," Chemical Engi-
   neering, 71, 1984,64-89.
                                                                                  TREATMENT OF HAZARDOUS WASTES     179

             Manuscript Withdrawn at the Request
                    of The PQ Corporation

Manuscript Withdrawn at the Request
       of The PQ Corporation
                            BARRIERS & WASTE SOLIDIFICATION  181

             Manuscript Withdrawn at the Request
                    of The PQ Corporation

Manuscript Withdrawn at the Request
       of The PQ Corporation
                            BARRIERS & WASTE SOLIDIFICATION  183

             Manuscript Withdrawn at the Request
                    of The PQ Corporation

Manuscript Withdrawn at the Request
       of The PQ Corporation
                             BARRIERS & WASTE SOLIDIFICATION   18~5

                      Innovative  Techniques for  the  Evaluation
                        of Solidified Hazardous  Waste  Systems

                                                 Harvill C. Eaton
                                               Marty E. Tittlebaum
                                               Frank K. Cartledge
                                       Hazardous Waste Research  Center
                                           Louisiana State University
                                             Baton  Rouge, Louisiana
  Hazardous  wastes  solidified  or stabilized  by mixing with
cementing materials are described as complex composiie materials
with microslruclurcs and microchemisiries which determine their
properties. The behaxior of these materials and their possible
long-term effect on the environment can be determined  by  ex-
amining them  with the tools common to  studies  in materials
science and engineering. Several of the microscopic and  micro-
analytical techniques useful in these investigations are described
and selected results are shown. These techniques, together with
measurements  of bulk physical properties, currently are being
utilized by the present authors to determine the mechanisms of
waste stabilization. The results  are being used to design effec-
tive methods  to handle real-world organic and  inorganic haz-
ardous industrial wastes.

  Certain types of hazardous wastes  can be converted  into a
manageable form by solidification. For  example, aqueous sludges
containing inorganic metal ions can be combined  with various
cementing mixtures to form solids which are often disposable in
a secure landfill. Several advantages are potentially afforded by
solidification, including little or no free-standing  liquids and re-
duced hydraulic pressures which might cause failure of the liner
system. If the  permeability of the solidified waste is low, then
the leachate from monolithic structures may be free of toxic ions
after their initial depletion from exposed surfaces.
  In the most desirable situation, the metal ions react chemically
with the matrix thereby forming a complex which is insoluble in
aqueous media and which is stable when leached by solutions with
a pH near those common in the environment or in a landfill.
Therefore, under the ideal situation, the waste is rendered effec-
tively non-hazardous. Consequently, solidification/stabilization
is potentially an important management alternative which may be
attractive for many hazardous wastes.
  Unfortunately, little is known about  the mechanisms of waste
solidification and even less about stabilization.  Until recently,
system design was essentially empirical and based on  industrial re-
search which usually remained proprietary. The relevant informa-
tion in the open literature dealt with nuclear waste stabilization'
or cement admixtures2 and, although useful, it more often than
not concerned  the interaction of cementing matrices with com-
pounds which  are not important to hazardous chemical waste
  Recently, however, it has been recognized that advances can be
made toward understanding the mechanisms of solidification and
stabilization by regarding the products as "complex composite
materials" with a  "microstructure".1  Mechanistic information
which might result from this viewpoint is essential to establishing
u fundamental basis for the a priori design of a solidified or stabil-
i/cd waslc.
  A material science point  of view suggests that the measurable
properties (e.g., unconfined compressive strength and toxic ion
leachability) of the waste/matrix systems depend on the micro-
structure and  the microchemistry. The  experimental techniques
commonly used by  the materials scientist offer ways to examine
the waste/matrix  systems which may provide information about
the mechanisms of solidification and/or stabilization.
  In this paper, the authors discuss the techniques that they have
found useful in their recent  studies of solidified inorganic sludges
and of solidified organics.' The methods  can  be  conveniently
grouped into those which provide morphological information and
those  which provide chemical information. They are briefly de-
scribed and representative results are presented.

  The authors' investigations primarily have involved mixtures of
hazardous chemical wastes and Type I Portland cement. Portland
cement was chosen  because it is a  common component  of many
solidification formulae which have been proposed and  of many
which currently are being  used by industry. Mixtures of pure
Portland cement and hazardous wastes are unlikely to be environ-
mentally, technically or economically practical, but they do offer
some of the laboratory control that is important  to a basic study.
Other, more complicated, matrices presently are  being examined
in the authors' laboratories as  part of a study of  interference
mechanisms. The work is funded by the U.S. EPA and  the U.S.
Army Corps of Engineers.
  A second application of the methods is part of an investigation
of test methods for evaluating real solidified wastes. The latter is a
cooperative study involving the U.S. EPA, Environment Canada
and several vendors in  the United States,  Canada and abroad.
  Weighed amounts of waste and water were mixed with 10 g of
Portland cement, contained in capped vials and stored at room
temperature in a darkened room to avoid photochemical reaction
of the organics. Samples were allowed to cure for various times
ranging from 1 week to over 1 year.
  Three organics  have been investigated:  p-bromophenol,  p-
chlorophenol  and ethylene  glycol. The  reasons for  selection of
these compounds and  detailed descriptions  of their morphology,
microchemistry and leaching behavior have been described else-

  The microstructure of a solidified waste depends on the follow-
ing factors:

• The chemical composition of the waste
• The chemical composition of the solidifying matrix compon-
  ents (cement, fly ash, clay, etc.)
• Relative amounts of waste and matrix components
• Amount of water
• Extent of mixing
• Temperature of mixing
• Time and temperature of set
  In the studies  performed by  the authors, these  factors have
been considered and maintained constant when appropriate, even
though in actual  practice most of the factors can vary consider-
ably.  It will be important,  therefore, for researchers to estab-
lish how strongly these factors affect the gross properties of the
final product.

Optical Microscopy
  Microstructure can be observed at several microscopic levels.
For example, entrained air can form bubbles ranging in size from
0.1 to 1 mm in diameter. These bubbles may be important in the
development of mechanical strength in the solidified product. Air
bubbles also can be the source of carbon dioxide to the matrix.
Carbonation can affect the hydration of cement paste by promot-
ing the formation of calcium carbonate which then decomposes to
a mixture of calcium carbonate and hydrated silica.7 Microstruc-
tural features of this size are  resolved easily with the optical
   There are several types of optical microscopes. They are con-
veniently categorized according to the branch of science which
they  serve  (e.g., the biological  microscope,  the  petrographic
microscope and the metallographic microscope).  The petro-
graphic microscope and, to a lesser extent, the metallographic
microscopes are the most useful for examining microstructures in
solidified and stabilized wastes.
                                                     /"^   -..;
                                                 100^, "Vf
                                                 	; «>w
                                                       '  O
                          Figure 1
   An Optical Micrograph of a Solidified P-Bromophenol Sample. A
               Bimodal Grain Structure is Revealed.
  Thin sections can be prepared, using methods standard to geo-
logical practice, so that light can be transmitted directly through
a small piece of solidified material. Using this method, the grain
structure can  be observed. Figure  1 is  an optical micrograph
which reveals the polycrystalline structure of solidified p-bromo-
phenol.  The Portland cement matrix material is apparent in the
microstructure with small grain-like hydration  phases.  There is
an equiaxed grain structure which is clearly visible at 100X mag-
nification, and the presence of several mineral phases is apparent.
If polarized light is used, the crystalline and noncrystalline com-
ponents of the  solidified system can be distinguished.  It also is
possible to do phase analysis of the crystalline components us-
ing index of refraction liquids. It is, therefore, in principle, pos-
sible to determine the compounds formed due to the presence of
the hazardous waste. In practice, this determination can be very
difficult and the researcher likely will more successfully identify
than locate the waste containing phases.  This identification is
illustrated in Figure 2 where the arrow indicates a mineral, i.e.,
crystalline, phase which contains hazardous p-bromophenol.
                           Figure 2
  A Higher Magnification Optical Micrograph of the Material Shown in
    Figure 1. The Arrow Points to a Phase Which is Concentrated in

  When only morphological information is required, such as in
studies of air entrainment, the metallurgical microscope is use-
ful. With the metallurgical microscope,  light is directed down the
microscope tube and is reflected off of the specimen  surface.
Reflected light travels back up the tube and through the ocular to
be detected by the operator's eye or to be recorded on a photo-
graphic plate or film. The  chief advantage  in  the use of the
metallurgical microscope over the petrographic microscope is in
specimen preparation. Since light is not transmitted through the
specimen, only a polished surface is required.

Scanning Electron Microscopy
  Microstructure on a smaller scale is apparent when the several
minerals  formed upon hydration of Portland cement are  con-
sidered. For example, calcium hydroxide crystallites are formed
and often are observed as hexagonal plates several microns in
extent. Even smaller crystallites of ettringite are  observed in
young cement pastes. Ettringite usually appears as rod or needle
shaped, single-crystals. They contain sulphur and are only a few
microns long.  The ettringite group  of minerals form the AFt
(A  = Al, F = Fe and t =  three moles of CaO.SO3)  phases.'
Both the calcium hydroxide and the AFt phases are observed best
by scanning electron microscopy (SEM).
  In the scanning electron microscope, a primary electron beam
is focused onto a specimen surface and rastered over a small area.
Primary electrons interact with the waste and matrix atoms con-
tained in the specimen producing additional electrons. These elec-
trons are released from the specimen atoms themselves and can be
                                                                                  BARRIERS & WASTE SOLIDIFICATION    187

released through several  mechanisms. The two electron genera-
tion mechanisms particularly useful for waste studies are secon-
dary electrons and backscattered electrons. A secondary electron
detector or a backscattered electron detector placed near the spec-
imen can be used to produce an image of the variations in their
generation as  the primary beam is rastered across  the specimen
surface. The utility of the SEM is related to the following char-
acteristics of its use and image contrast mechanisms:
•  Magnification variable from 10X to ca. 100X with a resolution
   of up to 5 nm or less
•  Magnification variable
•  Large depth of field
•  East of specimen preparation
   Specimens can be prepared by grinding and polishing in a man-
ner identical to that used for the preparation of specimens for
the petrographic microscope or for the metallurgical microscope.
The authors, in fact, frequently use the same specimen for both
types of examination. If three mutually orthogonal sections are
examined, the researcher usually can eliminate the  possibility of
mistaking textured for uniform structure. Polished sections must
be carefully studied, however, to eliminate artifacts caused by the
preferred  removal of certain phases during grinding and polish-
ing. Fractured surfaces can be useful in assuring that this is not
occurring. Figure 3 is an example of a scanning electron micro-
graph of solidified ethylene glycol.
                          Figure 3
    A Scanning Electron Micrograph of a Solidified Ethylene Glycol
Transmission Electron Microscopy
  When Portland cement is a component of the solidification or
stabilization matrix, a major hydration product is calcium sili-
cate hydrate (CSH). CSH is a poorly crystalline material which,
like the AFt phases, is comprised of several different compounds
which are mineralogically  similar.  When viewed by the optical
microscope or the SEM, CSH-appears as the dominant or matrix
phase of a solidified hazardous waste. The SEM sometimes  is
able to resolve Its morphology but  frequently reveals it as a fea-
tureless, but sometimes nodular  or grainy, material in which well
developed crystals like calcium hydroxide or AFt grow.
  The transmission electron microscope (TEM) is the instrument
of choice  for studying the morphology of CSH and other  very
small structures. The  TEM  is capable of resolving features as
small as a few angstrom units in cementitous materials. Unfor-
tunately, specimen preparation usually is more difficult than for
either optical or scanning electron microscopy.
   A stationary electron beam is focused onto a specimen which
intersects the optical axis of the instrument. The transmitted
beam is scattered and produces an image with contrast which is
primarily a function of the mass-thickness of the noncrystalline
components.  Crystalline  components scatter  the  specimen in
more complex ways due to diffraction. Figure 4 is a transmission
electron  micrograph of a small  crystalline phase in a solidified
                                                                                            Figure 4
                                                                   A Transmission Electron Micrograph of a Ca(OH)j Crystal (the large
                                                                  dark structure at (he center of (he micrograph) in a Solidified Ethylene
                                                                                         Glycol Sample
  Bulk  analytical instruments are an important and necessary
part of any investigation of the solidification and stabilization of
a hazardous chemical waste. If mechanistic information is desir-
able, it  is useful to have chemical analyses of the small phases
present  in the solidified material. Consequently, special methods
must be used.

Electron Diffraction
  The high resolution images  obtainable from the transmission
electron microscope and the ability to photograph diffraction pit-
terns make the TEM very useful for solidification/stabilization
research. Chemical  identification of mineral components of the
solidified waste is possible by analysis of the electron diffraction
  Figure 5 is an example of the results which can be obtained
from examination of a solidified hazardous waste by transmission
electron microscopy. The dark structure indicated by the arrow is
a calcium hydroxide crystal as revealed by the diffraction pattern.

Energy Dispersive and Wavelength Dispersive
X-Ray Analysis
  Specimens examined in the SEM emit x-rays in addition to elec-
trons. The x-ray energies are characteristic of the elements from
which they come. This means that they can be used to determine
the elemental composition of the specimen. The method is known
as energy dispersive x-ray analysis (EDS). Most EDS spectromet-
ers  will  detect  elements with atomic number 11 (Na) or greater.

                           Figure 5
    A Transmission Electron Micrograph of a Ca (OH)2 Crystal in a
 Solidified P-Bromophenol Sample. The Diffraction Pattern was Used to
             Confirm the Composition of the Crystal.
Those spectrometers equipped with a windowless detector can de-
tect even lighter elements and are particularly useful for carbon
and oxygen determination. A similar technique is commonly asso-
ciated with electron probe microanalysis (EPMA) and is known as
wavelength dispersive x-ray analysis (WDS).
  It is obvious that EDS and WDS would be useful in studies of
inorganic metal ion stabilization, but in certain cases they also can
be used in studies of organic waste solidification and stabiliza-
tion. The authors have used the techniques to study the partition-
ing of selected organics to  the hydration products of Portland
cement. Two organics, p-bromophenol and p-chlorophenol, were
selected because they were molecules which contained a heavy ele-
ment  which could be detected by EDS and WDS. The crystalline
phase shown  by the arrow in Figure 2 was determined by WDS to
contain large amounts of Br and, therefore, to be rich in organic.

X-Ray Power Diffraction
  X-ray powder diffraction (XRD) is very useful for studies of
waste solidification. A bulk analytical technique, XRD often can
be used  to identify the exact chemical compounds present in the
solidified system. This technique is generally  familiar, so it will
not be discussed in this paper.

  Studies of hazardous waste solidification/stabilization are com-
plicated by the fact that cementing mixtures are used as cementing
matrices.  Upon  hydration, a complex composite material is
formed. The microstructure of this material has structure at levels
which vary from that visible to the eye to that which can only be
resolved by electron microscopy. The properties of the solidified
system are determined by the microstructure and the microchem-
  Incorporation of hazardous wastes into the matrix by mixing
prior to hydration introduces the possibility that the wastes them-
selves are chemically  involved with the hydration reactions and
therefore are stabilized by the process. In addition, there is a pos-
sibility that the wastes are physically contained by the matrix in
pores or interstitial spaces. These sites of physical entrapment
may be so small that they are difficult to locate even by micro-
scopic methods.
  These characteristics suggest the use of microscopic and micro-
analytical methods for studying the mechanisms of waste solidifi-
cation and stabilization. The various methods used by the present
authors have been discussed. Some of the results of these studies
have been shown in the form of photomicrographs which reveal
selected characteristics  of cement/organic waste  solidification

  The authors would like to thank the U.S. EPA,  Environment
Canada, the  U.S. Army Corps of Engineers and the LSU Haz-
ardous Waste Research Center for financial support of this work.
Special  thanks go to  the following individuals for their sugges-
tions,  conversations  and guidance:  Carlton  Wiles,   Charles
Meshni, Phil Malone, Jerry Jones, Trevor Bridle, Pierre Cote and
Roger Seals. Special  thanks  go to  our students: Marie Walsh,
Donna Skipper, Devi Chalasani and Amitava Roy.

1. Cooper, J.A., Cousens, D.R., Lewis, R.A., Myhra, S., Segall, R.L.,
  Smart, R. St. C.,  Turner, P.S. and White, T.J., "Microstructural
  Characterization of Synroc C and E by Electron Microscopy,"  /.
  Am. CeramicSoc., 68, 1985,64-70.
2. Ramachandran, V.S., Concrete Admixtures Handbook: Properties,
  Science, and Technology, Noyes Publications, Park Ridge, NJ, 1984.
3. Cartledge,  F.K., Chalasani, D., Eaton,  H.C., Tittlebaum, M.E.
  and Walsh, M., "Combined Techniques to Probe Chemical Inter-
  actions of Organics with Solid Matrices," Proc. of the Meeting of the
  American Chemical Society, Miami, FL, May 1985.
4. Eaton, H.C., Walsh, M.B., Tittlebaum, M.E., Cartledge,  F.K. and
  Chalasani,  D., "Microscopic Characterization of the  Solidification/
  Stabilization of Organic Hazardous Wastes," Proc.  of the Energy-
  Sources and Technology Conference and Exhibition, Dallas, TX,
  Feb.  1985,  American Society of Mechanical Engineers Paper, No.
5. Walsh, M.B., Eaton, H.C., Tittlebaum, M.E., Cartledge,  F.K. and
  Chalasani,  D., "The Effect of Two Organic Compounds on a Port-
  land Cement-Based Stabilization Matrix," See these Proceedings.
6. Chalasani,  D., Cartledge, F.K., Eaton, H.C., Tittlebaum,  M.E. and
  Walsh, M.B., "The  Effects of Ethylene Glycol on a Cement-Based
  Solidification Process," See these Proceedings.
7. Suzuki,  K., Nishikawa, T. and Ito, S., "Formation and  Carbona-
  tion of C-S-H in Water," Cement and Concrete Research,  15, 1985,
8. Ramachandran, V.S., "Cement Science," in Concrete Admixtures
  Handbook: Properties,  Science, and Technology, V.S. Ramachan-
  dran, Ed., Noyes Publications, Park Ridge, NJ, 1984, 8.
                                                                                   BARRIERS & WASTE SOLIDIFICATION     189

                  Site  Characteristics  and  the  Structural  Integrity
                           of  Dikes  for  Surface Impoundments

                                         Jey K. Jeyapalan,  Ph.D.,  P.E.
                                                  Ernest R. Hanna
                              Wisconsin Hazardous Waste Management Center
                                             University of Wisconsin
                                                Madison,  Wisconsin
  Before a land disposal  facility  is developed for hazardous
wastes, the site characteristics need to be determined. The infor-
mation obtained for the geological and hydrogeological features
is  carefully used in the design  of the  surface impoundment.
Special design provisions need to be chosen if the site conditions
are poor.
  The damage due to the failure of dikes containing the wastes in
surface impoundments poses severe risk of contamination of sur-
rounding areas. The requirements of RCRA seek to ensure the
structural  integrity  of dikes by requiring certification  of  these
structures  by registered engineers.  The evaluation of site  char-
acteristics and the structural integrity of dikes is a complex sub-
ject. For the benefit of those involved in dike certification and
review of permit applications, the essential steps involved in these
evaluations are summarized in this paper.

  A surface impoundment design requires adequate data on the
site characteristics,  geometry of the dike and properties of the
soils forming the dike. In addition, information on groundwater
conditions is required to complete the design. Thus, for proper
water disposal techniques, the identification of various features at
the site is an important step. Subsequent  to the determination of
site characteristics, a detailed laboratory testing program needs to
be undertaken to determine the remaining engineering properties
of soils.
  The geometry of the problem is defined based on specifications
or field survey results,  and the  structural integrity of dikes is
determined using the appropriate analysis procedure. The struc-
tural adequacy needs to be ensured for different stages of the life
of the surface impoundment and shear  strength properties ap-
propriate for these stages need to be used in the analyses. Thus,
adequate understanding and recognition of all fundamental prin-
ciples of geotechnical engineering and hydrogeology are necessary
to successfully complete the above steps of engineering analyses.

  The first step is the site reconnaissance in which all pertinent in-
formation is obtained. An efficient and economical way to  begin
the site investigation is to review all available literature. This pro-
cess involves a review of past and present land uses, review of site
aerial photography  and  assessment of  the  site topography,
cultural landmarks, surface hydrology and surficial geology. The
location, accessibility, general area topography, past and present
ownership  and/or use  of a site are determined from maps ac-
quired from the local assessor's office and/or from the U.S. Geo-
logical Survey (USGS). The USGS and the Soil Conservation Ser-
vice (SCS) may be utilized as sources of preliminary information
regarding  the surficial geology of the site under investigation.
Site-study reports,  as well as previous plans and specifications,
may be available from area contractors, consultants and govern-
ment  agencies.  Any  construction activity  in  the project  area
should be visited. A site survey should be performed to familiarize
the engineer with the area. High water marks and vegetation can
indicate the nature of the soil. Recent changes in topography also
can be helpful in design procedures.
  Geological conditions  and project objectives will dictate the
type of drilling needed to meet the job requirements. It is not
always possible or cost-effective to use a single drilling method to
complete a project. This is one reason why, prior to specifying a
drilling method, an analysis  of the project requirements is done.
Selection of a drilling  method most suited to the particular job
can be based on the following factors:
• Type of formation
• Depth of drilling required
• Availability of drilling  equipment
• Special and/or other specific  requirements

  The use of a backhoe or shovel  is a very simple and quick
means of determining  shallow  soil conditions  and  shallow
groundwater conditions.  Advantages are the low cost, mobility
and ease of operation. The main disadvantages are the depth
limitation and sample disturbance.

Surface Geophysical Techniques
  The application of geophysics to hazardous waste management
combines proven technology with  new technology to form an in-
tegrated systems  approach.  This  process is efficient and  cost-
effective. Several types of geophysical exploration techniques are
now available for a rapid evaluation  of subsoil characteristics.
Definite interpretation of the results requires  correlation with
standard subsurface exploration methods. Geophysical methods
can be divided into two categories: (1) surface techniques and (2)
downhole  techniques. A  few of these surface geophysical tech-
niques are described below.

Ground Penetrating Radar
  Ground penetrating radar (GPR)  uses high  frequency  radio
waves to  acquire subsurface information. Energy  is radiated
downward from a small antenna which is moved slowly across the
surface of the ground. The energy is reflected back to the receiv-
ing antenna,  where variations in the return signal are continuously
recorded. Radar responds to changes such as bedding,  cementa-
tion, moisture, clay content, voids, fractures, intrusions and man-
made objects. Depth of penetration  is site-specific.

  GPR is limited in depth by attenuation. Better overall penetra-
tion is achieved in dry, sandy or rocky areas. Poor results are ob-
tained in moist,  clayey or conductive soils. Radar penetration to
30 ft is common.

Electromagnetic Measurements
  The electromagnetic method provides a means of measuring the
electrical conductivity of subsurface soil, rock and groundwater.
Electrical conductivity is a function of the type of minerals in the
formation, its porosity and permeability and the fluids which fill
the pore space. Often the  conductivity of the pore  fluids
dominates the measurements.
  Natural variations in subsurface conductivity may be caused by
changes in soil moisture content, pore fluid conductance, depth
of soil and thickness of formation layers. The absolute values are
not necessarily diagnostic; it is the variations that are significant.

Seismic Refraction and Reflection
  Seismic refraction and reflection techniques are used to deter-
mine  the thickness and depth of geologic layers by using the
velocity of seismic waves which is an indication of the soil and
rock's engineering properties. Reflection can be used effectively
to depths ranging from about 50 ft to  several thousand feet.
Structural and formation boundaries can be detected.  Refraction
commonly is applied to shallow investigations up to a few hun-
dred feet. Depth to bedrock and the layer thickness of the overly-
ing unconsolidated material can be determined.
  Seismic waves transmitted  into the subsurface travel at dif-
ferent velocities in various types of soil and rock. The waves travel
in all directions  from the source and are  reflected and refracted
when they impinge on an interface. The combination of pattern
and velocity will affect travel time. An array of geophones on the
surface measures the travel times of the seismic waves from the
source to the geophones. The refraction and reflection techniques
use the travel times  of the waves to provide data on  subsurface
layers, including the number of layers, the thickness of the layers
and their depths.

Resistivity Method
  The resistivity method is used to measure the electrical resistiv-
ity of the section which includes the soil, rock and groundwater.
The method may be used to determine lateral changes and vertical
  The procedure requires that an electrical current be injected into
the ground by a  pair of surface electrodes. The resulting potential
field  is  measured at the surface  between a second pair of elec-
trodes.  The subsurface resistivity can be  calculated by knowing
the separation and geometry of the electrode positions, applied
current and measured voltage.
  In  general, most soil and rock minerals are  highly resistive,
hence the flow  of current is conducted  primarily through the
moisture-filled pore  spaces within the soil and rock.  Therefore,
the resistivity of soils and rocks  is predominantly controlled by
the porosity and the permeability of the system. Electrical sound-
ing is used to reveal the  variations of apparent resistivity with
depth. Horizontal profiling is used to determine lateral variations
in resistivity.

In Situ Measurement Techniques
  The determination of soil properties by in situ measurement
techniques also has advanced greatly. A few of these methods are
described below.

Standard Penetration Test
  The standard  penetration test is the oldest and most widely ac-
cepted method for obtaining shear strength properties. This in-
volves counting the number of blows required of a 140-lb hammer
dropped from a height of 30 in. to move a split spoon sampler 12
in.  through the soil.  The resultant number is termed the  "N"
count and is at present the most widely used index of subsurface
soil conditions. Advantages of the standard penetration test in-
clude economy of use, simplicity of procedure and widespread ac-

Cone Penetration
  Cone penetration devices are useful because of simplicity of
testing, reproducibility of results and ease of correlating results to
other test results. The drawback of the test is that no soil sample is
  The test procedure involves advancing vertically, at a constant
rate, a penetrating cone of about 10 cm2 base area and about 60°
apex angle. The force needed to penetrate  a given distance in a
soil is used in measuring shear strength.

Vane Shear
  In contrast  to methods which determine strength parameters
from empirical correlation,  the  vane  shear  test attempts  to
measure undrained shear strength directly.  The test procedure is
to advance  a vane configuration to a desired  soil  depth and
measure the applied torque as the vane is rotated at a constant
  Shearing resistance  is considered to be mobilized on a cylin-
drical failure surface of rotation, corresponding to the top, bot-
tom and sides of the vane assembly. The vane shear test possesses
the advantages of economy of use and reduced soil disturbance.

  The pressuremeter  is used by lowering the apparatus into a
borehole or, if the machine is self-boring,  by drilling to the re-
quired depth. Increments of hydraulic pressure are applied to the
test cell which acts against the borehole wall. The resulting defor-
mation  of the borehole  wall is determined by  monitoring the
volume  change of the  test cell.
  The relationship between pressure increment, volumetric ex-
pansion  and  time   is  examined  to determine   strength
characteristics. No soil  sample is obtained if  the  self-boring
pressuremeter is used.

Iowa Borehole Shear Test
  The Iowa borehole  shear test is performed by lowering a shear
head consisting of two opposing horizontally grooved shear plates
to the required test position in an uncased borehole. The two
shear plates are expanded until seated in the borehole walls at a
selected pressure. After a time  period allotted to allow for con-
solidation to occur, the  shear  head  is either pulled  upward  or
pushed  downward at  a steady rate of 2 mm/min. The required
forces for shearing are measured, and the shearing stress is plotted
against the normal pressure. By performing a number of tests at
different seating pressures,  the soil strength can be calculated.
  Soil strength may be expressed in a number of ways, and the
particular strength parameters to use in design will depend on the
nature of a specific problem.

  The shear strength behavior of cohesive soils differs consider-
ably from  that of cohesionless soils. This section presents the
details of the shearing resistance of cohesive soils under various
loading and drainage  conditions.  In  order  to obtain meaningful
results from  a laboratory  test program for the shear  strength
parameters of cohesive soils, it is necessary to test good quality
samples representative of the in situ soils or compacted soils pro-
posed for the project. For in  situ soils,  the samples to be used
                                                                                  BARRIERS & WASTE SOLIDIFICATION

should be undisturbed as far as possible and must be tested at the
in situ water content and dry density. For compacted soils, it is
necessary to match the following three factors  between  the
laboratory samples and the soils used in the field:

•  Method of compaction
•  Dry unit weight, 74
•  Water content, u
   It is  necessary to relate  shear  strength to compaction at an
elementary level in this section. A typical compaction curve show-
ing the dry unit weight-water content relation for a cohesive soil is
given in Figure  l(a). The variation of shear  strength with water
content is shown in Figure  l(b). Although a  cohesive soil can be
compacted at the same dry unit weight at points A and B in Figure
l(a), the water content at these points would differ considerably.
Thus,  the shear strength of the compacted  soil at  A would be
higher  than that  at  B.  Therefore, it  is necessary  to  prepare
samples in the laboratory at values of water content  and dry unit
weight comparable to those of the field compacted soils. Further-
more, the structure of a compacted clay is highly dependent upon
the method  of compaction, and the structure has a significant in-
fluence on the shear strength and drainage properties of the clay.
Therefore, it is also important to use  a laboratory compaction
procedure that would yield compacted test samples with a struc-
ture similar to that of the field  compacted soils.
                    Water  Content , oo

                          Figure 1
                      Compaction Curve
Shearing Resistance, Cohesion
  The shearing resistance of cohesive soils differs  from that of
cohesionless soils due to the following differences in soil factors:
 • Particle size
 • Particle shape
 • Permeability
 • Angle of internal friction
 • Plasticity

 The sizes of particles  forming cohesive soils are significantly
 smaller than those of cohesionless soils. Cohesionless  soils are
 made of angular to rounded shapes of particles, whereas cohesive
 soils are made of platey shaped particles. Cohesive  soils drain
 much more slowly than cohensionless soils due to the fact that the
 coefficient of permeability of cohesive soils is lower than that of
 cohesionless  soils. The angle of internal friction of cohesive soils
 generally is slightly lower than that of cohesionless soils.  Further-
 more,  the small  particle sizes of  cohesive soils have attractive
 forces  giving the plastic behavior  for cohesive soils;  this factor
 does not apply to cohesionless soils. Due to the above factors, the
 shear strength behavior of cohesive soils is expressed by the equa-
     t =  c +  crtan 
where c is cohesion and  is the angle of internal friction.
  The Mohr-Coulomb failure envelopes for clays and sands are
shown in Figure 2 for comparison. The value c indicated on the
shear stress, also referred to as the "Cohesion Intercept," is due
to the attractive forces between the platey shaped particles form-
ing cohesive soils such as clays.
                                                                         Envelope  for  Sands-
                                                                                                     Mo hr -Coulomb
                                                                                                     Envelope  for Sands
                                                                              Normal Stress, o
                                                                                           Figure 2
                                                                                      Strength Envelopes
Pore Pressure Parameters
  The low permeability  of  clay causes pore water pressure to
build  up during laboratory  or field tests and  control the shear
strength  properties measured  for  use  in design computations.
Therefore, it becomes necessary to understand the mechanisms of
pore water pressure changes in cohesive soils before making an at-
tempt to understand the shear strength behavior of cohesive soils.
  When  a sample of clay is loaded  in a triaxial test initially by an
equal  all around confining pressure, 0$, followed by a deviator
stress, a, -<73, a total pore pressure change, Au, occurs as shown
in Figure 3, under undrained conditions. The above two types of

                  cr 3 + (a j -a 3 )
                A u3
                    Pore  Pressure
a3 )
Pore Pressure CMange

        = A(  o"!-  0-3 )

     Figure 3
   Pore Pressures
loading independently cause two increments of pore pressures,
Au3 and Aui3, given by:
      Au = Au3 +
        J3 =
      Au13 = A(ff] - CT3)
as shown in Figure 3. In Equations 2 and 4, B and A are referred
to as the "Pore Pressure Parameters." By knowing these pore
pressure parameters, the excess pore water pressures developed in
a triaxial test can be estimated.  Conversely, by  measuring  the
changes in pore pressures, Au3 and Aui3, in a laboratory^ test for
known values of a3 and (al  - a3), the values of B and A can be
  The pore pressure parameter, B, is a function  of the coefficient
of volume compressibility, mv, of the soil sample shown in Figure
4(a) and the degree of saturation, S, indicated in Figure 4(b). The
coefficient of volume compressibility increases as the stiffness of
the soil skeleton decreases and,  as a result,  the value of B in-
creases for the same value of saturation as shown in Figure 4(b).
When the sample is dry due to air being far more compressible
                                           than the soil skeleton, the pore water pressures developed are zero
                                           and, therefore, the value of B is regarded as zero. Since the pore
                                           water will  carry the entire confining pressure  under undrained
                                           loading conditions  for a saturated  sample of clay,  the pore
                                           pressure  parameter,   B,  becomes  1.0.  The pore  pressure
                                           parameter, A,  is a  function  of stress ratio,  oj/cr;, as  shown in
                                           Figure 4(c). For normally consolidated clay,  A is approximately
                                           1.0; for overconsolidated clays, it is between  -0.3 and 0.8.
                                                                                               50          100
                                                                                         Saturation,  S(%)
                                     Stress Ratio,  ai  /o 3

                                        Figure 4
                                 Pore Pressure Parameters

              Unconsolidated-Undrained Test
                The Unconsolidated-Undrained, U-U, Quick or Q Test usually
              is performed to obtain strength parameters, c and , correspond-
              ing to a fast loading of the soils at sites where drainage is poor. In
              this test,  the soil sample is trimmed and placed in a thin mem-
              brane as shown in Figure 5(a). The confining pressure, a3, is ap-
              plied to simulate the confinement available to the sample in the
              field while the drainage value is closed. The sample is sheared
              while the drainage  valve is closed, not allowing any excess pore
              pressures built up in the sample  to dissipate, as shown in Figure
                Since  it is impossible to  saturate the sample without con-
              solidating, pore pressures are never measured in the U-U test. The
              Mohr-Coulomb envelope for a saturated sample of clay in a U-U
              test is horizontal, as shown in Figure 6(a), since the deviator stress
              at failure is independent of the confining pressure applied to the
                                                                                 BARRIERS & WASTE SOLIDIFICATION    193

                                                                                                  O 1
                             Sample of  Soil
                                  o 3
                                   Valve  Closed
                       o i
                                  0 3
                                  Valve  Closed
                                  (Pore  Pressures
                                   are never
                       Figure 5
               Unconsolidated-Undrained Test
sample. For a saturated sample of clay yields the U-U test gives
the following strength parameters:
      c = Su

       = O
where Su is the undrained shear strength of the clay.
  Since the deviator stress  is independent  of  the confining
pressure, it is possible to perform a test with a confining pressure
of zero and obtain the undrained shear strength,  Su> of the soil
from the diameter of the failure Mohr's circle as shown in Figure
6(b). This test is referred to as the "Unconfined Compression
Test," where the  sample is sheared by a compressive load, a^
without the use of a confining cell. The diameter of the Mohr's
circle qu, in Figure 6(b)  is called the "Unconfined Compressive
Strength" of the sample.
  When a partially saturated sample of clay is sheared in a U-U
test, the size of the Mohr's circle increases with increasing confin-
ing pressure as shown in Figure 7. Once the confining pressure
reaches a high enough value  to cause the pore air to dissolve in
pore water, the sample becomes saturated. Beyond this value, the
sample yields a horizontal Mohr-Coulomb envelope as shown in
this figure. Thus,  in order to obtain the strength parameters, c
and , corresponding to a confining pressure, a3 (of interest in the
partially saturated range), it becomes necessary to draw a tangent
at <73  = a'j on the envelope and read values for c* and ' as shown
in Figure 7.
  The strength parameters obtained from a U-U test are  useful
for analyzing geotechnical engineering problems where the soils
did not consolidate appreciably and the drainage conditions were
poor. An  example of such an application is  the slope stability
evaluation for the dike for the end-of-construction loading condi-
                                                                          Unconsolidated-Undrained  Envelope
                                                                                    c = S  .  « =  0
                                                                        r   ZOLA
                                                                             Normal Stress, a
                                Mohr  Circle  for the
                                 Unconfined  Compression
                                 Test  with  O3  =  0
                                                                                     Normal Stress,o
                               Figure 6
                           Strength Envelopes


              Soil  is
              Pa_rt ially
Soil  is Saturated
                                              *  for  03   03 *
        O 3   03
                             A     A
                                 Normal Stress,  o
                               Figure 7
                           Strength Envelopes

Consolidated-Undrained Test
  The Consolidated-Undrained,  C-U or R Test usually is per-
formed to obtain strength parameters, c and , corresponding to a
fast loading of the soils at a site  where drainage is poor and the
site has undergone consolidation.  In this test, the drainage valve is
left open during the application of the initial confining pressure,
'under drained loading conditions. The rate
of loading in this test is slow enough to dissipate the excess pore
Total  Stress
                                                                                                       Effective  Stress
                                                                               Effective Normal Stress,a

                                                                                           Figure 9
                                                                                      Strength Envelopes
water pressures as they build up during the application of the
shear load. The strength parameters obtained from this test pro-
vide the data necessary for long-term  stability calculations of
slopes and embankment dams.

  It is necessary that the dike be stable for the following loading
• End-of-construction or short-term
• Steady seepage or long-term
• Rapid drawdown
• Seismic
  The factor of safety against any form of slope failure usually is
determined using the following:
• Slope geometry
• Shear strength of soil
• Method of analysis

  There are more  than  30  methods of slope stability  analysis.
Among  these,  Bishop's modified method, Ordinary Method of
Slices and Spencer's method are most commonly used.  Bishop's
Method and the Ordinary Method of Slices are used with circular
failure surfaces, and the Spencer's Method is used for noncircular
slope surfaces. These methods can be applied for dike  stability
evaluation using:
                                                                                BARRIERS & WASTE SOLIDIFICATION    195

• Hand computations
• Computer programs
• Stability charts

Hand Computations
  The hand computations are almost never used in practice at the
present time for slope stability calculations. These are tedious and

Computer Programs
  Computer programs for IBM personal computers and  main
frame computers are available from the authors for performing
dike stability evaluation. These computer programs have various
desirable options to handle even the most difficult site conditions.

 Stability Charts
   The stability of slopes  can be analysed efficiently using the
 stability charts shown in the manual by Duncan and Buchignani.'
 Although the charts assume simple slopes and uniform soil condi-
 tions, they can  be used to obtain reasonably accurate answers for
 most complex problems if irregular slopes are approximated by
 simple  slopes and average values of  unit weight, cohesion and
 friction  angle are used.
   When detailed analyses of slope stability are performed using
 the  computer programs based on  either Ordinary Method  of
 Slices,  Bishop's  Modified  Method  or Spencer's  Method, a
 number of slip surfaces must  be examined  to  locate the most
 critical failure surface with the lowest factor of safety. The slope
 stability charts yield minimum factors of safety when properly ap-
 plied. Under most conditions,  the  uncertainties due  to approx-
 imations and assumptions in the  method of analysis  are smaller
 than the uncertainties due to inaccuracies in measuring the shear
 strength. Approximations in the analysis usually amount to 15%
 or less, but the margin for error in evaluating shear strength may
 be considerably greater. The minimum allowable value of factor
 of safety for a slope in a dike depends on  several factors; some
 guidance is given in Table 1 for static and seismic conditions.
                            Table 1
         Recommended Minimum Values of Factor of Safety

                            Uncertainly of Strength Measurements
Costs and Consequences
of Slope Failure                     Small1
Cost of repair comparable to cost
of construction. No danger to human   1.25
life or other property if slope fails.     (1.2)*

Cost of repair much greater than
cost of construction, or danger to      1.5
human life or other valuable          (1.3)
property if slope fails.

or greater
(1.7 or
 1 The uncertainly of ihc Mrcnglh measurement* is smallest when ihc soil conditions arc unilorm
  and high quality .strength lest data provide a conslslcnt, complete and logical picture of the
  strength characteristics.
 ' The uncertainly of (he strength measurements is greatest when the soil conditions are complex
  and when the available strength data do not provide a consistent, complete, or logical piclure
  of the strength characteristics
 * Numbers without parentheses apply for static conditions and Ihosc with parentheses apply lor
  seismic conditions
 (After Duncan and Buchignani')
                  EVALUATION OF BEARING
                    If the foundation of a  proposed dike  is composed of low
                  strength  soils, the factor of safety against any  type of bearing
                  capacity  failure should be evaluated. The bearing capacity consid-
                  eration will control the base width of the dike, especially when the
                  foundation contains a scam of weak material, as shown in Figure
                  10.  In this figure:

                      q'w=7H.                                            0)
                    Using stress influence charts, the influence factor, I, is obtained
                  corresponding to the dimensions, a and b.
                    Using  I, the stress, qw, applied to the seam of weak soils can be
                  calculated as:
                            Iq «
                    The ultimate bearing capacity of the weak material is evaluated
                  using the equation:
    qu  _ cNc
                                      BN- t-  Z"»-NC:
                  where Ne, N,f and Ng are bearing capacity factors obtained from
                  bearing capacity charts corresponding to the angle of friction, ,
                  of the soil.  In this equation, B and y refer to the width of the
                  stress distribution as shown in Figure 10 and the unit weight of the
                  soil, respectively.  The  factor of safety  against bearing capacity
                  failure is calculated using:
                       ' * ~ q»
                  and Fs should be higher than 2.5 to 3.
                              1   1    1   i    1   1    i   1    1

, B ,

                           Figure 10
                    Bearing Capacity Failure
  The RCRA regulations require that the surface impoundments
must  have dikes that are designed, constructed and maintained
with sufficient structural integrity to prevent massive failures of
the dikes. Thus, the engineers involved in certifying dikes need to
use  appropriate  engineering  procedures  to  determine  site
characteristics and  dike integrity. Some guidance is given in this
paper to accomplish these tasks undertaken by engineers involved
in the preparation and review of permit applications for surface
impoundments, waste piles and other land disposal units.

 1.  Bishop, A.W., "The Use of Slip Circle in the Stability Analysis of
    Slopes," Geotechnique, 15, 1955, 7-17.
 2.  Bishop, A.W. and Bjerrum, L., "The Relevance of the Triaxial Test
    to the Solution of Stability Problems," Proc. of the ASCE Research
    Conference on the Shear Strength of Cohesive Soils, Boulder, CO,
 3.  Duncan, J.M. and Buchignani, A.L., "An Engineering Manual for
    Slope Stability Studies," U.S. Berkeley Geotechnical  Engineering
    Report, 1975.
 4.  Holtz, R.D. and Kovacs, W.D.,  An Introduction to Geotechnical
    Engineering, Prentice-Hall, NJ, 1981.
 5.  Hvorslev, M.J., "Subsurface Exploration and Sampling of Soils for
    Civil Engineering Purposes,"  Corps of Engineers, Waterways Ex-
    periment Station, Vicksburg, MS,  1949.
 6.  Janbu,  N., "Slope Stability Computations," Soil Mechanics  and
    Foundation Engineering Report," the Technical University of Nor-
    way, Trondheim, Norway, 1968.
 7.  Jeyapalan, J.K., Theory and Problems of Geotechnical Engineering,
    McGraw-Hill, New York, NY, 1985.
 8.  Lowe, III,  J. and Karafiath,  L.,  "Stability of  Earth Dams Upon
    Drawdown,"  Proc. of the First PanAmerican Conference on  Soil
    Mechanics and Foundation Engineering, Mexico City, Mexico, 2,
    1960, 537-552.
 9.  Mitchell,  J.K., Guzikowski, F. and Villet, W.C.B., "The Measure-
    ment of Soil  Properties In situ,  Present Methods—Their Applica-
    bility  and  Potential," National  Technical  Information  Service,
10.  Morgenstern, N.R., "Stability Charts for Earth Slopes During Rapid
    Draw-down," Geotechnique, 13, 1983, 121-131.
11.  Navfac DM-7,  "Soil Mechanics Design Manual," U.S. Navy  Of-
    fice, 1981.
12.  Seed, H.B., "Considerations in Seismic Design of Earth and Rock-
    fill Drums,"  19th Rankine Lecture, London, 1979.
  The following symbols used in this paper:
Angle of Internal Friction
Undrained Shear Strength
Effective Consolidation Pressure
Unit Weight
Dike Slope Angle
Dike Height
Unit Weight of Shear
Blow Count
Cone Resistance
Factor of Safety
                                                                                        BARRIERS & WASTE SOLIDIFICATION     197

                          Use  of  X-Ray Radiographic  Methods
                                   in the  Study of Clay Liners

                                            Philip G.  Malone, Ph.D.
                                                  James H. May
                                            Geotechnical Laboratory
                            U.S. Army Engineers Waterways Experiment Station
                                              Vicksburg, Mississippi
                                             Kirk W.  Brown, Ph.D.
                                                James C. Thomas
                                      Soil and Crop Sciences Department
                                             Texas  A&M University
                                              College Station, Texas
  X-ray radiography has been widely used in soil investigations
to study the distribution of layers in soil cores and the effects of
changing conditions (loading or impact) on soil structure. X-ray
radiographic techniques also can be useful in  studying clays or
clay soils used in liners.
  Laboratory investigations were undertaken to demonstrate that
X-ray radiographic techniques could be used to detect density and
soil  structure changes that  usually accompany variations in
hydraulic conductivity of clay liners. An example of a real-time
test of a simulated bentonite and sand liner attacked with  acid
lead nitrate and examples of radiographic examination of clay soil
(non-calcareous smectite) samples that have been permeated by
lead acetate or lead nitrate are presented. The changes in density
and structure can be related to changes observed in hydraulic
conductivity during permeation.
  X-ray radiography easily can  be applied  to field samples of
soil or clay liner materials to detect density and structural changes
that occur as the liner and permeating fluid interact. X-ray tech-
niques have applications in both understanding failure mechan-
isms and forecasting liner performance.

  X-ray radiographs of soil cores  have been employed in  geo-
technical investigations  for many years because radiographs  pro-
vide a  unique technique for detecting  changes in the physical
arrangement of soil particles without disturbing a soil core. In
many soil cores, X-ray radiographs will detect  subtle differences
in soil density that might have escaped notice  during the visual
examination  of core sections. The X-ray image displays varia-
tions in the absorption of X-ray photons in different parts of the
object examined.  Differences in absorption can be  related to
changes in sample thickness, density or chemical composition.1
When the sample geometry is controlled by preparing cylinders
or slices of cylinders for examinations, patterns observed in the
radiographs can be reliably related to the differences in sample
density and chemistry.
  This study examined the use of X-ray radiographic techniques
in identifying changes in the density and composition of samples
of simulated  clay  or other soil liners that  have been exposed to
permeants that could be encountered in  groundwater beneath
an industrial or hazardous  waste disposal site. Two approaches
were used: (1) examining the real-time invasion of a bentonite-
sand liner  using a lead nitrate-nitric acid solution; and (2) ex-
amining the  condition  of soil samples after the samples have
been permeated with simulated industrial waste solutions.
           AND  —
s 	 83mm —
|> 	 75 mm 	 ^
	 . 	 _ . '
	 	 . "
/ 	 	 '
/ 	 "

"~ , _ , __ . _ . _ . /

•k ~"-"~ /
% -_ \f
W " ": " \
s - ^!_ ~ - __ - . /

*71 *.



X) mm


> rr

                         Figure I
    Cross-Section of the Test Cell Used in the Real-Time Liner Test
  A  simulated  waste pond liner was prepared  by  mixing one
part of a commercially available bentonite (Aquagel from Baroid
Petroleum Services  of  N.L. Industries, Houston, Texas) with
four  parts by volume of coarse filter sand. The mixture follows
typical procedures for producing a bentonite liner.J The sand and

bentonite were mixed in a dry condition and water was gradually
added with continued mixing to assure that the bentonite was
completely hydrated. The wet mixture was permitted to settle for
24 hours, and excess water was decanted. A test cell was prepared
from a 400-mm length of 75-mm ID schedule 40 PVC pipe (Fig.
  The bottom of the test cell was fitted with a plastic plug and a
hole was drilled through the plug so that a drainage line could be
attached. The bottom of the test cell  was packed with  Dacron
filter floss, and a 100-mm thick layer of coarse filter sand was
poured into the test cell. The bentonite and sand mixture was
poured in on top of the filter sand, and the test cell was gently
shaken to assure that the bentonite and sand layer settled and no
air bubbles formed in the simulated liner. Clean water was placed
on top of the bentonite and sand layer  to verify that the simu-
lated liner would not leak. To initiate the testing, the clean water
was decanted, and a simulated waste formed by saturating a 10%
nitric acid solution with reagent-grade crystalline lead nitrate was
added above the bentonite and  sand layer. The saturated acidic
lead nitrate solution was approximately 140 mm deep over  the
simulated liner.  This depth of  waste  is less than the 300-mm
depth permitted under disposal guidelines.3
                                                                  SUPPORT RING
                                                                     1 cm
                                                                                             Figure 3
                                                                    Print of Radiograph of Simulated Liner One Day after Application of
                                                                                     Acidic Lead Nitrate Solution
                            Figure 2
   Print of Radiograph of Simulated Liner Before Application of Acidic
                       Lead Nitrate Solution
  X-ray radiographs were  made of the test cell at one to  five
day intervals using a Norelco MG 3000 X-ray system equipped
with a tungsten target  tube. Kodak Industrial M film was used
with a 1 to 2 min exposure. The X-ray unit was operated at 290
kV with 11 mA current.
  Figure 2 shows the liner  prior to testing. Breakthrough of the
simulated waste was noted after 16 days. Figures 3 and 4 show the
progressive failure of the liner after a one-day and after a 25-day
exposure to the simulated waste. When the test cell was opened,
the presence of voids in the simulated liner easily could be ob-
served (Fig. 5). The X-ray radiographs had detected shrinkage
cracks forming in the  simulated liner and had accurately indi-
cated the mode of failure in the simulated liner. Lead nitrate, a
radio-opaque solution, was used initially in the tests to assure that
any invasion of the liner could be observed. The formation of the
voids in the clay reduced the usefulness of the lead solution be-
cause the lead salts lowered the contrast between the fluid-filled
voids and  the surrounding soil. Where flocculation and  void
formation are the expected  modes of failure, there is no benefit in
using radio-opaque permeants.
                                                                                    BARRIERS & WASTE SOLIDIFICATION    199

   1 cm
                           Figure 4
    Print of Radiograph of Simulated Liner 25 Days after Application
                     of Acidic Lead Nitrate
  The evaluation of the interaction of soil liners and simulated
or actual industrial wastes generally has been performed by forc-
ing samples of selected permeants (wastes) through packed soil
columns in  a  fixed wall or flexible wall permeameter.' Inter-
actions  are evaluated by determining if the permeability of the
packed  soil is changed by the permeant. The type of interaction
that has occurred between the soil and the permeant cannot be
determined  without examining  the  soil.  For  example,  the
hydraulic conductivity of the liner may decrease, but the reason
for the decrease (for example, swelling of the soil or the forma-
tion of precipitates) cannot be determined without measuring the
change in sample volume (for swelling) or looking for the forma-
tion of  new compounds in the clay (for precipitation). X-ray
radiographs can detect both volume changes and increased liner
density  due to the formation of precipitates. More than 60 soil
samples  from permeameter tests were examined to evaluate the
ability of radiographic data to assist in the  interpretation of
permeability changes.' The major effects observed have included
the breakdown of the clay structure (causing an increase in sample
permeability) and the formation of precipitates (with a decrease
in permeability).
                                                                                          Figure 5
                                                                Photo of Bentonile and Sand Mixture after Permeation with Acidic Lad
                                                                                       Nitrate Solution
  Samples  of permeated soils were prepared  using techniques
proposed for  testing potential  landfill liner materials.' In the
examples discussed here, six samples of the Lufkin soil (a non-
calcareous  smectite) from Brazos County, Texas, were packed
into standard-design fixed wall permeameters (Fig.  6).  Each
sample was compacted in three equal lifts using 25 blows per lift
from a 2.4l-kg hammer falling 300 mm. One set of three samples
was permeated with a 60% (by weight) lead acetate solution; the
other set was permeated with a 50% (by weight) lead nitrate solu-
tion in 0.1% HNOj.
  After a sufficient volume of permeant had passed through the
sample to establish a trend in changes  in hydraulic conductivity,
the permeameters were  opened and checked for swelling. The
samples permeated with lead acetate had increased in  height by
                        PRESSURE INPUT
                                    POROUS STONE
                             TEFLON TUBING
                           Figure 6
     Cross-Section of Permeameter Used in the Study of Soil-Waste

 1  cm
                          Figure 7
   Print of Radiograph of Non-Calcareous Smectite Soil Permeated
                  with Lead Acetate Solution

maximum of 2.7%. Maximum swelling in the samples permeated
with lead nitrate was 1.6%. The samples were extruded and wrap-
ped  in foil. X-ray radiographs were prepared  using a  Norelco
MG 300 Industrial X-ray System. The unit was operated at 290 kV
at 10-12 mA. Exposure times varied from 1 to 3 min using Kodak
Industrial M film.
   Permeameter testing established that the lead acetate caused a
decrease in hydraulic conductivity by a minimum of 50%. X-ray
radiographs showed lead compounds were precipitating in the soil
(Fig. 7).
   The lead nitrate solution caused an increase in hydraulic con-
ductivity by a minimum of over  100%. The X-ray radiographs
showed irregular connecting voids  suggesting the clay had devel-
oped a flocculated or blocky structure similar to that  observed
in the real-time failure study (Fig. 8).

   X-ray radiographic techniques, previously applied to the study
of soils, can be employed usefully in the investigation of the inter-
action of wastes and clay (or clay soil) liners. Using suitable test
cells, real-time effects of waste (or simulated wastes) on sections
of liner can be observed. Soil samples that have been permeated
by simulated wastes can be X-rayed to examine the soils  for den-
sity or structure changes that indicate occurring liner-waste inter-
   Field techniques for sampling soils to obtain undisturbed sam-
ples  are available, and it should be possible to use X-ray radio-
graphic methods along with standard soil description techniques
to examine the condition of liners at landfills or waste storage
areas. X-ray examination of soil cores is completely non-destruc-
tive, and altered sections observed in the X-ray radiographs can
be documented by sampling and examining anomalous soil inter-
  1 cm

                          Figure 8
   Print of Radiograph of Non-Calcareous Smectite Soil Permeated
                  with Lead Nitrate Solution
  The tests described and the resulting data presented herein, un-
less otherwise noted, were  obtained  from research conducted
under the Independent Laboratory Initiated Research Program,
Project No. 4A161102A91D Task Area 02 Work Unit 154, of the
United States Army Corps  of Engineers by the USAE Water-
ways Experiment Station. Permission was granted by the Chief
of Engineers to publish this information.

1. Krinitzsky, E.L., Radiography in the Earth Sciences and Soil Mechan-
  ics, Plenum Press, New York, NY, 1970.
2. Matrecon, Inc., "Lining of Waste Impoundment and Disposal Facil-
  ities," U.S. EPA, SW-870, Washington, DC, 1983.
3. U.S. EPA,  "Hazardous  Waste Management System; Permitting
  Requirements for Land Disposal Facilities," Federal Register, July
  26, 1982.
4. Brown, K.W. and Anderson, D.C., "Effects of Organic Solvents on
  the Permeability of Clay Soils," EPA-600/2-83-016, U.S. EPA, Cin-
  cinnati, OH, 1983.
5. Malone, P.O., May,  J.H., Brown,  K.W. and Thomas,  J.C.,  "Use
  of X-ray Radiographic Techniques in the Evaluation of Soil Liners,"
  Miscellaneous Paper  GL-85-14, USAE Waterways Experiment Sta-
  tion, Vicksburg, MS, 1985.
                                                                                 BARRIERS & WASTE SOLIDIFICATION    201

                 Closure Design and Construction  of Hazardous
                           Wastes Landfills  Using  Clay  Sealants
                                              John F.  O'Brien, P.E.
                                                  Lonnie E. Reese
                                                O.H. Materials Co.
                                                 Rosewell, Georgia
                                                 Ian Kinnear, P.E.
                                                  Dames & Moore
                                                Boca Raton,  Florida
  Clay sealants are employed at an increasing number of sites
securing land disposal of hazardous industrial wastes. Such seal-
ants are employed as topliners to limit infiltration and/or as bot-
tom liners to limit exfiltration and provide for leachate collection.
These sealants are employed both as primary liners and as added
security to synthetic liners. Liner design for hazardous waste land-
fills is well-regulated and quantitative analytical tools are  well-
documented. However,  the practical  implementation of  cost-
effective, buildable and  reliable clay liners requires  design in-
sight beyond the obvious. Key elements in the selection, design
and construction of landfill liners using clay sealants are discussed
herein. The case histories of two landfill closures are presented to
amplify the impact of the designers' assumptions on  implemen-

  The national concern for environmentally secure disposition of
hazardous wastes  has led to increased sophistication in disposal
technology. In land disposal,  landfill design has been expedited
not only by expansive regulatory requirements, but also by ana-
lytical modeling  which  allows cost- and  time-effective  evalua-
tion of design alternatives.
  Ideally, landfills securing hazardous materials are designed
prior to disposal of the materials. Such planning includes design
to limit the development of leachate and preclude exfiltration of
leachate to the groundwater. However, as a result of years of un-
regulated and unenforced disposal, thousands of small to  large
"out back" sites at industrial  facilities exist across the nation. It
is the closure of these existing facilities which is the subject of this
paper.  These existing facilities  include  several  distinguishing
characteristics which impact closure design  and construction:

• The waste material often is poorly defined
• Stabilized or unstabilized, wastes are left in-place and closure
  includes emplacement of a top liner only
• Construction  of the design groundform often includes sub-
  stantial regrading of unknown materials, placing a premium on
  planning equipment, selection and safety.

  Two case histories will amplify this discussion.

• "Site A, Southern Florida": closure of a hazardous waste land-
  fill using an imported bentonitic clay as a topliner
• "Site B, South  Carolina": closure of a solvent contaminated
  landfill using locally available clays

Site A, South Florida
  The landfill at Site A  is located within an industrial facility in
South Florida. It consists of one 4-acre cell that was constructed
to a height of about 35 ft above existing grade. The landfill was in
operation from 1958 until the middle of 1984. The cell was util-
ized primarily for the disposal of paper cartons, boxes, construc-
tion debris,  cafeteria waste,  yard trimmings  and scrap metal.
However, in the early years of filling when operating  practices
and regulations were less stringent than today, unbagged asbestos
and drums containing solvents  and fuels were deposited in the
  Below the cell, the groundwater is contaminated with VOCs;
the resulting contamination plume is slowly migrating eastward,
away  from the landfill. Associated with the landfill closure, a
groundwater recovery well system and  treatment plant will be
constructed around the perimeter of the landfill.
  The landfill is bounded closely on three sides by water. Uncon-
trolled filling over the  years left a very irregular  groundform
with slopes as steep as 1:1.

Site B, South Carolina
  Site B is located within an industrial facility in South Carolina.
Over at least 10 years of operation,  the site was used as a dump
for solvents  and drums. At the time of closure, the drums had
been removed and closure involved capping solvent contaminated
soils on the 5-acre site. The contaminated soils were encountered
largely in depressions in the land surface on this site.

  Closure  design  of industrial  and hazardous waste landfills
usually involves the following generic evaluations:

• Determine the waste  composition and characteristics at the
  closure sites
• Determine  regulatory  and  design  dictated materials  and
  groundform requirements
• Determine liner materials alternatives based on  compatibility
  with the waste, availability, constructability and performance
• Complete analytical evaluations to determine optimum landfill
  design based on cost and performance

  The regulatory  requirements  for design of hazardous  waste
landfills are well-documented by the U.S. EPA under RCRA,
published July 26,  1982. The  Interim Final Rule on "Hazardous
Waste Management System: Permitting Requirements  for Land
Disposal Facilities" became effective Jan. 26,  1983.' These fed-
eral criteria do not dictate either liner parameter, but do include
the criteria for bottom liners for new (not existing) landfills.
  ...The liner must be constructed of materials that prevent
  wastes from passing into the  liner during the active life of
  the facility.1

This requirement virtually requires the use of geomembranes in
lieu of any clay liner as the primary sealant of a new landfill.
  In the closure design of existing facilities, closure  requires a
cap to limit infiltration as well as groundwater monitoring. In the
usual case, closure of an  existing hazardous site allows the de-
signer some latitude in the selection of a landfill cap, either em-
ploying geomembranes or clay liners.
  The regulatory  environment of closure  design requires  con-
sideration. To this end, it is  imperative the owner and engineer
develop communication with the appropriate regulatory agency
early  in the  design process to assure compatibility of the design
with the regulatory expectations.

  The hydrologic processes within landfills were originally mod-
eled using water balance techniques  developed by Thornwaite.3
The U.S. EPA developed the  first widely used landfill  model,
Hydrologic Simulation on Solid Waste Disposal Sites (HSSWDS),
in 1980.  Since that time,  the U.S. EPA has developed Hydro-
logic  Evaluation of Landfill Potential (HELP), the most sophis-
ticated analytical model now available for landfill systems. HELP
was  developed specifically for hazardous waste landfill evalua-
tions  as required by RCRA."
  HELP provides a quasi-two-dimensional hydrologic model of a
landfill section, applying section geometry, climatologic data and
material parameters to account for the movement of water across,
into,  through and out of landfills. Figure 1 presents a generic
landfill section indicating the solutions provided by the model.
The model allows the use of geomembranes in the section,  both
enhancing lateral drainage and limiting percolation.  Despite the
claims of some suppliers, geomembranes (synthetic liners) should
not be considered impervious. Hydrologic modeling of a landfill
performance may be completed using any of the above  tech-
niques:  Thornwaite,  HSSWDS  or  HELP.  Of these  models,
HELP certainly   represents  the  most sophisticated modeling
alternative.  Regardless of the approach used, some quantitative
evaluation of landfill performance should be undertaken.
       ATIVE COVER      » » »             t t T T
               ^,J,x„,„,*,   *  n     L»*-,
               \'/'/'/'/ /'/'/'/'//\/\ V,'/'/'*'*'}•',''/'/'/'
               W'/'/'/.' if.-.'.-.   DISCHARGE
                 'LATERAL DRAINAGE LAYER •-.  '• .  .*;*.'*• ScJSSe'e
                                EXFILTRATION THROUGH
                               BASE OF LANDFILL SECTION
    (after Schroeder, et al)
                                           6" DRAINAGE PIPE
                                           TO DRAINAGE DITCH
                           Figure 1
             Modeling a Landfill Section Using HELP
                           Figure 2
  Typical Landfill Cap Employing Clay, Geomembrane and Geotextile
Sealant Alternatives
  Clay sealants may be used as, or as a part of, a landfill liner.
Figure 2 shows the implementation of a geomembrane and clay
liner as a cap following stabilization of a sludge pond. Whatever
the case, the designer will have two alternatives in clay selection:
• Naturally occurring and locally available clays
• Locallv available soils improved by admixtures
  Clays are chosen as sealants for  their low degree of pervious-
ness. In the design of topliners, a clay barrier with a coefficient of
permeability of 10~7 cm/sec or less normally will be required to
limit infiltration.
  Many  naturally  occurring clays  can readily achieve this low
permeability. However, the designer should not presume that any
available clay may be placed to this  permeability. Design assump-
tions should be correlated with careful laboratory testing. Repre-
sentative samples of the available  clay should be compacted in
laboratory molds and tested to determine permeability.
  In most  topliners  used  in the closure of existing hazardous
waste sites, the permeant will be rainwater. However, if this is not
the case,  a permeant representative of that which will infiltrate the
liner should be used to test the compatibility of the permeant
with the clay minerology.
  Permeability of  clays will vary with relative compaction. The
most commonly used compaction standard is the standard Proc-
tor, ASTM D-698. Laboratory molds compacted to varying  de-
grees of  relative compaction and thereafter tested to determine
permeability generally will show substantial variation. Clays com-
pacted to 95% of the maximum dry density achieved in the stan-
dard Proctor test may  show an order of magnitude decrease in
permeability over a sample of the same soil compacted to 90% of
standard Proctor. Similarly, a clay  compacted from wet material
of the optimum moisture content generally  will have a lower
permeability than the same soil at the same relative compaction,
but compacted from dry  material. Thorough testing to deter-
mine the mechanical properties of clays considered for use as lin-
ers is important not only to determine the suitability of a material,
but also  to determine the  range of performance to be expected
from field variations in installation.
  In many areas, the naturally occurring soils will not provide the
low permeability required for  liner  design. In  this  event, soil
admixes  may be used to lower permeability. Cement,  as well as
industrial wastes such as fly ash and kiln dust, may be added to
soils to lower permeability. However, the brittle behavior makes
these materials undesirable to seal structures as prone to irregular
settlement as  landfills.
  The  most common soil admix is sodium bentonite. Not com-
monly  encountered in nature, sodium bentonite is mined  largely
in the North Central United States. As a soil admixture, sodium
                                                                                 BARRIERS & WASTE SOLIDIFICATION    203

                                TOP SLOPE 4% (TTP)
                                SIDE SLOPE JM. IV
        'SEAL  (4 INCHES)  •
                                    	   OAS CONVCTANCE TRENCH
                                    •f   12' WIDE • ?' OEEPI
                           Figure 3
        Typical Section, Florida Site. Including Gas Conveyance

 bentonite is attractive for its expansive characteristics. The addi-
 tion of 5-15% (by weight) sodium bentonite to soils will lead to a
 coefficient of permeability of 10 7 cm/sec or less. As with natural
 clays, the selection of sodium  bentonite as a soil  admix should
 include careful testing of representative materials.
   Figure 3 presents a section of the top liner used at Site A, South
 Florida. Clayey soils are almost totally absent in South Florida.
 Economic evaluations of geomembranes and  admix improved
 local soils led to the selection of the sodium  bentonite and local
 sand section. Laboratory testing indicated that the  sandy  local
 soils mixed with 13% by weight sodium bentonite and compacted
 to 95% of standard Proctor (ASTM D-698)  would achieve a
 permeability of 10~7 cm/sec or less.
   Figure 4 presents the topliner section used at Site  B, South Car-
 olina. At this site, the Owner's property included an area of thick
 clayey residual soil about 0.5 miles from the landfill. Laboratory
 testing demonstrated a  permeability of 10~7 cm/sec  for these
 clays compacted to 95% of standard Proctor.
                          Figure 4
              Typical Section, South Carolina Site B
Buffer and Cover Protection
   Protection of clay liners from environmental damages is an im-
portant design consideration. Clay liners should, at a minimum,
include a top soil buffer to protect the material from erosion. The
buffer also should be of sufficient thickness to preclude breaking
by roots and burrowing animals. A vegetated soil cover of about
1 ft is sufficient in most cases to adequately protect the clay.

Specifications and Quality Control
   Certainly the best closure planning and design is impotent with-
out carefully developed  specifications and good field quality con-
  The designer should be responsible for  the technical specifi-
cations. In the writers' experience, many landfill closure specifi-
cations include exacting criteria for material type, degree of com-
paction and permeability. Such tightly constructed specifications
open the owner to substantial costs in change orders.
   If   the  material  to be  used  has been  pre-determined  by
 laboratory testing, specifications for liner construction should be
 limited to degree of compaction and material type. Additionally,
 to assure good  field control,  a minimum compactive effort and
 soil moisture may be  specified (for example, "three passes of a
 kneading-type compactor weighing a minimum of 8 tons, com-
 pacting soils within 3%  of their optimum  moisture content").
   If  the specific  clay  material has  not been  predetermined,
 specifications should  be limited  to general material  type  and
 permeability requirements.
   Installation  of  clay  liners should  be  under  the  full-time
 surveillance  of  a representative  of the designer's office. At a
 minimum, this surveillance should be supplemented by testing for
 soil parameters  detailed  in the drawings and specifications to in-
 clude material type, thickness, degree of  compaction,  moisture
 and permeability.
   State regulated closure criteria for Site A in South Florida (Fig.
 3) dictated permeability requirements (1Q-7 cm/sec),  but  left
 materials selection to  the contractor. Quality control testing for
 the liner  installation included  thickness, moisture, relative com-
 paction and laboratory permeability of field samples.  Additional-
 ly, the contractor was required to  submit  his sodium bentonite-
 local  sand mix  design, which was  monitored in  the field. The
 average  permeability determined  from  field samples was 10~8
   At Site B, South Carolina, liner construction specifications de-
 tailed the borrow source and compaction requirements only. Pre-
 testing of  these  soils  had  well-established  the density  and
 permeability relationships of  these soils.  A full-time  represen-
 tative from the designer's office monitored compaction and liner
 thickness. Additionally, it was required that this 2-ft thick liner be
 placed and compacted in increments no greater than  6 in.

   Closure construction of hazardous waste landfills requires con-
 siderable skill, equipment and experience. Such work should be
 undertaken only by a contractor experienced in the field.
   Construction  planning  for closure  involves  the following
 generic process:

 • Equipment scoping, based  on materials,  earthwork  and pro-
   ductivity requirements
 • Project safety planning
 • Decontamination planning
 • Contingency planning

   The contractor views the closure of a hazardous waste landfill
 differently than does the engineer. The good design engineer best
 serves his client, the owner, if designs and specifications consider
 the constructability questions  the contractor faces.

 Equipment Scoping
   Both sites A and  B present good examples of the constructor's
 response in equipment  and  approach to particular site  re-
 quirements.  Site A, in South  Florida, was characterized by con-
 strained work  space and relatively steep side slopes. Site B, in
 South Carolina, required effective  hauling of a large quantity of
 clay cover to the site.
   The constrained (bounded closely on three sides by water) site
 conditions at Site A, South Florida, required the landfill ground-
 form  to  be constructed with a side slope as steep as practical. The

designer selected 3 horizontal to 1 vertical (3:1). Construction of
the topliner section required the importation of about 16,000 yd3
of soil. Permit requirements required construction to be complete
in an 8-week period. The physical constraints not only dictated
the landfill groundform, but also limited the  area available for
materials staging.
  The steep  side  slopes and  moderate  materials requirements
complicated earthwork at Site A.  All  soils were placed using
12-yd3 scrapers pushed by a dozer up the sides  of the landfill and
depositing  soils as they glided down-slope. Materials  deliveries
were carefully coordinated with placement productivity because
of the limited staging area.
  The sodium bentonite-local sand liner was mixed on-site using a
pugmill to  effect thorough mixing.  About half-way through the
clay liner construction, mechanical problems forced abandon-
ment of this mixing technique.  Thereafter, the clay and sand were
mixed using bulldozers to place the materials together  in a con-
trolled area. The scrappers then picked up and placed the mix.
  The 4-in. thick clay liner was difficult to place on the steep side
slopes. Measured thickness of  the liner varied  from 3.5 to 7 in.,
averaging about 5 in.
  A self-propelled, rubber tired vibratory compactor was used to
compact the sand-clay topliner at Site A. Again, because of the
steep side slopes, the compactor was pushed up-slope, then glided
and compacted moving down-slope. The thin liner section limited
equipment  alternatives. The contractor, concerned that a dozer-
towed vibratory compactor would damage the 4-in. liner with its
tread, elected to use the rubber tired compactor. Adequate com-
paction was  achieved in  a single  pass, which  alleviated the
awkwardness of compaction,
  Experience at Site A indicated that 3:1  slopes are  about the
steepest which may be practically constructed, limiting equipment
and earthwork  control. Indeed, flatter slopes are more  desirable
from a construction standpoint.
  At Site B,  South Carolina,  emplacement of the clay liner re-
quired moving about 30,000 yd3 of clay from the borrow source 1
mile (by road) from the site. To optimize placement, a  high pro-
ductivity earthwork program was established,  moving 3,000 yd3
per  day. Self-loading 12-yd3  scrapers  were used to  pick up,
transfer and place  the clay. Kneading compactors ("sheepsfoot"
compactors) towed behind bulldozers were used to compact the
clay to specifications.

Project Safety Planning
  Safety should be given the  highest priority during any work
with hazardous wastes. Personnel involved in cleanups at  both
Sites  A and  B were participants  in  comprehensive medical
monitoring programs—both that on-going by the contractor and
project  specific medical monitoring.
  No closure of a  hazardous waste  landfill should be completed
without a Site  Safety Plan (SSP) developed by a Certified In-
dustrial Hygienist (CIH). This  document not only advises person-
nel of hazards  and potential hazards caused by exposure to the
landfill  materials,  but also establishes site safety protocol.  At a
minimum, this protocol includes definitions of levels of protec-
tion required for different areas of the site as well as site monitor-
ing procedures and emergency equipment and procedures.
  Both Sites A and B required personnel to wear full face air-
purifying respirators and skin protection afforded by disposal
suits and gloves. Working in this wear (particularly in summer) is
quite  fatiguing and should be associated with increased breaks
and liquid intake. Project planning should consider the associated
loss in productivity.
  Both Sites A and B were attended by a full-time health and
safety officer. A portable  decontamination trailer,  including
showers, and contingency decontamination equipment  was util-
ized on each site. Daily "tailgate" safety sessions were conducted
throughout the projects to address safety concerns for each day's
planned work.

Equipment and Personnel Decontamination
  Equipment and personnel decontamination is required follow-
ing any exposure to hazardous materials. The best decontamina-
tion program is, of course, one which minimizes the contact of
equipment and personnel with the hazardous wastes. However, as
this exposure is not always avoidable, equipment and personnel
decontamination should be provided for.
  Equipment decontamination procedures should be such that no
decontamination is trafficked off-site. Equipment decontamina-
tion normally is undertaken on a specially constructed pad de-
signed to collect and store any wash water.
  A convenient personnel decontamination area should be pro-
vided. At this station, workers should have easy access to all safe-
ty equipment required for work on the site. It is here that workers
remove street clothes and put on protective clothing before enter-
ing the site.

  Landfill design has progressed to the point where a number of
liner alternatives may be evaluated quickly  and cost effectively.
While quantitative methods to model landfill performance are
available,  practical  implementation   of  cost-effective   and
buildable liners requires  design insight beyond the obvious.
Closure of existing hazardous waste facilities should be completed
by contractors experienced in the field. Experience not only in the
work  but also in the associated personnel protection and decon-
tamination concerns, should be closure imperatives.
1. U.S. EPA, Federal Register, July 26, 1982, 32274-32388.
2. U.S. EPA, 40 CFR 264.301(a)(l).
3. Thornwaite, C.W. and Mather, J. R., "The Water Balance," Clima-
  tology, 8, No. 1, 1955.
4. Schroeder, et al., "The Hydrologic Evaluation of Landfill Perfor-
  mance (HELP) Model," Municipal Environmental Research Labora-
  tory,  Office of Research and Development, U.S. EPA, Cincinnati,
  OH, 1984.
                                                                                  BARRIERS & WASTE SOLIDIFICATION    205

             Soil Liners  for  Hazardous Waste  Disposal Facilities
                                                  D.C. Anderson
                                       K.W.  Brown  and Associates, Inc.
                                              College Station, Texas
  The Hazardous and Solid  Waste Amendments of 1984 con-
tained a requirement for double liners containing both a flexible
membrane liner and compacted soil liner. The soil component can
be used either as the sole secondary  liner or as the lower com-
ponent in a composite secondary liner.  In either case, soil liners
must be at least 3 ft thick with a demonstrated field permeability
of less than or equal to 1 x 10- ' cm/sec.  Factors affecting field
permeability of soil liners  include soil  characteristics  and liner
construction methodology. Some soil characteristics which can
yield low permeability liners (e.g., high plasticity) also can give
rise to the potential for shrinkage and permeability increases.
  One  of the most important considerations in the selection of
methods for soil liner construction is the destruction of clod struc-
ture during compaction. After the selection of a specific construc-
tion methodology, the best way to achieve design specifications
is  by implementating a sound construction  quality  assurance

  Soil liners are a required component of double liner systems
for hazardous waste disposal  facilities (Fig. 1). This requirement
was established in the Hazardous and Solid  Waste Amendments
of 1984 and further defined in the Minimum Technology Guid-
ance (MTG) on Double Liner Systems.'  In this paper, the author
discusses ways to incorporate soil liners into these liner systems
and factors most greatly affecting the field permeability of these
  All double liner systems  are required  to have primary flexible
membrane liners (FML). The MTG provides two options for the
design of secondary liner systems, both of which include soil lin-
ers. Following is a discussion of the soil liner components to these
liner systems.

Soil Liner Component to
Secondary Composite Liners
  One of the designs suggested in the MTG incorporates a secon-
dary composite liner composed of an upper FML and a lower
compacted soil liner (Fig. 1A). In the configuration suggested in
the MTG, the FML and soil liner are sandwiched together to min-
imize the potential for lateral flow between the liners. It has been
suggested that this configuration would limit leakage through de-
fects in the FML to the rate at which leachate could flow into the
area of soil liner directly under the FML defect. If this theory is
correct, composite liners should be able to reduce leakage from
disposal facilities to a much greater extent  than either an FML or
soil liner alone.
                                            PRIMARY FML
                                           -SECONDARY FML
                   UMSATUSATED ZONE
                                       -PRIMARY FVL
                   UNSATURATEC 20NE

                         Figure 1
Double Liners for Hazardous Waste Landfills Must Have a Primary FML
     and Either an (A) Composite, or (B) Thick Compacted Soil
                      Secondary Liner
  Requirements for the soil component to secondary composite
liners are given in Table 1. The traditional requirement for a
permeability of less than or equal to 1 x 10-' cm/sec must be
verified in the field.' Laboratory permeability tests are no longer
considered adequate to verify actual field performance.

Secondary Soil Liners
  The other secondary liner design given in the MTG is that of a
thick compacted