United States
             Environmental Protection

             Agency
xvEPA
             Office of Research and

             Development

             Washington, DC 20460
EPA/540/R-94/515

August 1994
Symposium on

Intrinsic

Bioremediation of

Ground Water

Hyatt Regency Denver
Denver, CO
August 30 - September 1,1994

-------

-------
                                                        EPA/540/R-94/515
                                                        August 1 994
Symposium  on  Intrinsic Bioremediation of Ground Water
                       Hyatt Regency Denver
                           DenveV, CO
                  August 30 to, September 1, 1 994
                 Office of Research and Development
                 U.S. Environmental Protection Agency
                         Washington, DC
                                                    Printed on Recycled Paper

-------
                                     Disclaimer
The projects described in this document have been reviewed in accordance with the peer and
administrative  review  policies of the U.S. Environmental  Protection  Agency and the U.S.
Geological Survey, and have been approved for presentation and publication. Mention of trade
names or commercial  products does not constitute endorsement or recommendation for use.

-------
                                 Acknow edgments
The papers abstracted in this book were presen ed atthe Symposium on Intrinsic Bioremediation
of Ground Water, held August 30  to  Septsmber 1, 1994,  in  Denver,  Colorado.  The
Symposium was a joint effort of the U.S. Environmental Protection Agency's (EPA's) Biosystems
Technology Development Program and the LJ.S. Geological Survey (USGS), with additional
sponsorship from the U.S. Air Force.  Fran V. ^remer (EPA, Cincinnati, Ohio), John T.  Wilson
(EPA, Ada, Oklahoma), and Gail E. Mallard (USGS, Reston, Virginia) served as co-organizers
of the Symposium, along with support from Lt. Col. Ross N. Miller, Headquarters, U.S. Air Force
Center for Environmental Excellence,  Brooks Air Force Base, Texas.

-------

-------
                                      Contents
Site Characterization: What Should We Measure, Where (When?), and How?
   Michael J.  Barcelona	
Processes Controlling the Distribution of Oil, Air, and Water
  John L Wilson 	
New Tools To Locate and Characterize Oil Sp
Us in Aquifers
   Bruce J.  Nielsen
Microbiological and Geochemical Degradation Processes
   £ Michael Godsy	
Field and Laboratory Results: Getting the Whole Picture
   Mary Jo BaedecJcer	
In Situ Bioremediation at the Seventh Avenue pite in Denver: Remediation of Soils
and Ground Waters
   Christopher Nelson	
The Role of Intrinsic Bioremediation in Closure of Sites After Cleanup Through
In Situ Bioremediation: The Regulator's Perspective
   Mar/c E. Walker and Lisa C. Weers  ....
The Importance of Knowledge About Intrinsic Bioremediation for Cost-Effective Site
Closure: The Client's Perspective
   Harry E. Moseley  	
The Role of Intrinsic Bioremediation in Closurs of Sites After Cleanup Through
In Situ Bioremediation: The Role of Mathematical Models
   T/ssa H. Illangasekare, David C. Szlag, anc

Intrinsic Bioremediation of JP-4 Jet Fuel
   John T. Wilson, Frederick M. Pfeffer, James
   Todd H. Wiedemeier, Jerry E. Hansen, and
10


25


35


41



47



49



51
John T. Wilson   	  53
W. Weaver, Don H. Kampbell,
Ross N. Miller 	  60
A Natural Gradient Tracer Experiment in a Hoterogeneous Aquifer With Measured
In Situ Biodegradation Rates: A Case for Natiiral Attenuation
   Thomas B. Stauffer, Christopher P. Anfworf/j, J. Mark Boggs,
   and William G. Maclntyre  	,
Traverse City: Distribution of the Avgas Spill
   David W. Ostendorf	
                                       73
                                       85

-------
 Traverse City: Geochemistry and Intrinsic Bio-remediation of BTX Compounds
   Barbara H. Wilson, John T. Wilson, Don H. Kampbell, Bert E. Bledsoe, and
   John M. Armstrong  	  94

 Mathematical Modeling of Intrinsic Bioremediation at Field Sites
   Hanadi S. Rifai	  103

 Biogeochemical Processes in an Aquifer Contaminated by Crude Oil: An Overview
 of Studies at the Bemidji, Minnesota,  Research Site
   Robert P. Eganhouse, Mary Jo Baedecker, and Isabelle M. Cozzarelli	  Ill

 Simulation of Flow and Transport Processes at the Bemidji, Minnesota, Crude-Oil
 Spill Site
   Hedeff I.  Essaid	  121

 An Overview of Anaerobic Transformation of Chlorinated Solvents
   Perry L McCarty	  135

 Contamination of Ground Water With Trichloroethylene at the Building 24 Site at
 Picatinny Arsenal, New Jersey
   Mary Martin and Thomas E. Imbrigiotta  	  143

 Intrinsic Bioremediation of TCE in Ground Water at an NPL  Site in St. Joseph,
 Michigan
   John T. Wilson, James W. Weaver, and Don H. Kampbell	  154
Poster Session

Technical Protocol for Implementing the Intrinsic Remediation With Long-Term
Monitoring Option for Natural Attenuation of Fuel-Hydrocarbon Contamination
in Ground Water
   Todd H. Wiedemeier	  1 63

Wisconsin's  Guidance on Naturally Occurring Biodegradation as  a Remedial
Action Option
   Michael J. Barden	  I/O

Assessing the Efficiency of Intrinsic Bioremediation
   Francis H. Chapelle  	  171

A Practical Approach to Evaluating Natural Attenuation of Contaminants in
Ground Water
   Paul M. McAllister and Chen Y. Chiang	  1 72

The Use of Low Level Activities To Assist Intrinsic Bioremediation
   Robert D. Norris, Jeffrey C. Dey, and Daniel P. Shine  .	 .  1 73

-------
Natural Attenuation of Jet Fuel in Ground Wafer
  Greg Doyle, Dwayne Graves, and Kandi Brown  	  1 75
Evaluation of Intrinsic Bioremediation at an Urderground Storage Tank Site in
Northern Utah
  R. Ryan Dupont, Darwin L Sorensen, and /v <.
Case Studies of Field Sites To Demonstrate Nctural Attenuation of BTEX Compounds
in Ground Water
  Chen Y. Chiang and Paul M. McAllister
arion Kemblowski	  1 76
                                     178
Demonstrating Intrinsic Bioremediation of BTEX at a Natural Gas Plant
   Keith Piontek, Tom Sale, Steve de Albuquerc
Demonstrating the Feasibility of Intrinsic Bioremediation at a Former Manufactured
Gas Plant
   /an D. MacFarlane, Edward J. Bouwer, and Patricia J.S. Colberg	  181

Natural and Enhanced Bioremediation of Aronatic Hydrocarbons at Seal Beach,
California:  Laboratory and Field Investigations
ue, and John Cruze	  1 79
   Harold A. Ball, Gary D. Hopkins, Eva Orwir
The Complete Dechlorination of Trichloroetheie to Ethene Under Natural
Conditions in a Shallow Bedrock Aquifer in New York State
   David Major, Evan Cox, Elizabeth Edwards,
, and Martin Reinhard	  1 83
and Paul W. Hare  	  187

-------

-------
Site Characterization: What Should We Measure, Where (When?), and Why?
Michael J. Barcelona
Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, Ml
Abstract
Site characterization represents the initial phase of the active monitoring process that occurs as
part of intrinsic organic contaminant bioremediation efforts. Initial characterization work sets the
stage for evaluating the progress of the nature  transformation of contaminants. The following
have frequently been observed: parent compound disappearance, active microbial populations
with biotransformation capabilities,  and the appearance or disappearance of organic and
inorganic constituents that provide  evidenc
Quantitative evidence is lacking, however, for
mixtures solely by biological processes. This is
      3  of bioremediation  at  contaminated  sites.
      net removal of toxic compounds from complex
      due largely to the reliance on monitoring well
samples for evidence of biological activity, rather than on identifying the mass of contaminants
(and total reactive organic carbon) and estimating the net removal/transformation of reactive
compounds over time.

A dynamic approach to quantitative site characterization is needed that recognizes intrinsic
bioremediation  as an active  cleanup  approach. Careful attention must be  paid  to the
identification  of the  three-dimensional   dis
correspondence between contaminant distribu
microbial conditions in the subsurface overtim
      tribution  of  contaminant  mass.  Then  the
      ion and favorable physical, geochemical, and
      3 provides a basis for net contaminant-removal
estimates. Mere adaptations of detective ground-water monitoring networks are insufficient for
quantitative evaluation of intrinsic bioremediafon technologies.
Introduction

The practice of site characterization for
evolved  slowly  in  the  past  decade.  Early
contamination detection monitoring (i.e., moni
been applied to many sites of potential concern
phases.
This  minimal  approach  has  been  applied
remediation of subsurface organic contaminants has
      guidelines  (1 -3)  for minimal  ground-water
      oring wells upgradient and downgradient) have
      rom detection through remedial action selection
       widely,  regardless  of the  physicochemical
characteristics of contaminant mixtures orthe complexity of hydrogeologic settings. With solubl
inorganic constituents, this approach may be adequate for detection purposes, but assessment
efforts require substantially  more  comprehensive  approaches.  For organic  contaminant
assessment efforts (i.e., determinations of the rature and extent of contamination), wells alone
have been found to be inadequate monitoring 1ools. Recognition of the value of subsurface soil
vapor surveys for volatile organic components cf fuel and solvent mixtures has generated a flurry
of modified site characterization approaches bused on monitoring wells (4). These approaches
to site characterization and monitoring networljc design suffer also from a failure to identify the
total mass of contaminant in the subsurface.
Symposium on Intrinsic Bioremediation of Ground Water

-------
This failure occurs for three main reasons. First, although volatile organic compounds (VOCs)
are mobile in ground water and are frequently early indicators of plume movement (5), their
detection in vapor or well samples and their apparent aqueous concentration distribution do not
identify the total  mass distribution of organic contaminant (6). Second,  efforts to correlate
observed soil vapor or ground-water VOC concentrations with those in subsurface solid cores
have often been unsuccessful, because current bulk jar collection/refrigeration at4°C guidelines
for solid core samples for VOC analyses lead  to gross negative errors (7). Third, "snapshots"
(i.e., one-time surveys) of background and disturbed ground-water chemistry conditions  have
been interpreted as "constant," ignoring temporal variability in subsurface  geochemistry.

The unhappy result of the slow improvement in site characterization and monitoring practices
has often been the very low probability of detecting the source of mobile organic contaminants.
This outcome may be followed by the misapplication of risk assessment or remediation models
and fiscal resources. Nonetheless, good reasons exist for a more optimistic view of the future
reliability of site-characterization and monitoring  efforts.

The  shortcomings  of  previous contaminant detection and assessment  efforts  have  been
recognized. New guidelines and recommendations for network design and operations will lead
to more comprehensive, cost-effective site characterization  (7, 8)  in general. Also, excellent
reviews of characterization and long-term monitoring needs and approaches in support of in situ
remediation efforts should guide us in this regard (9, 10). Site characterization efforts provide
a  basis  for long-term monitoring design and actually continue throughout the life of a
remediation project.
Advanced  Site Characterization and Monitoring

How do we estimate the potential for subsurface intrinsic bioremediation success and track its
performance into  the future?  Clearly, we should seek  to design technically defensible
characterization and monitoring networks that will  provide reasonable estimates of in-place
contaminant distributions overtime. Therefore, a dynamic, ongoing site-characterization effort
includes the following objectives:

       •      Identify the spatial  distribution of  contaminants, particularly their relative
              fractionation in subsurface solids, water, and vapor, along potential exposure
              pathways, recognizing that the mass of contaminants frequently resides in the
              solids.

       •      Determine the corresponding spatial  distribution of total reactive organic matter
              (e.g., degradable normal, aliphatic, and aromatic hydrocarbon compounds),
              because overall microbial activity and disruptions in subsurface geochemical
              conditions (and bioremediation indicators) are due to the total mass of reactive
              organic carbon.

       •      Estimate the temporal stability of hydrogeologic and geochemical conditions that
              may favor microbial transformations  in background, source, and downgradient
              zones during the first year of characterization  and monitoring.

-------
               Derive initial estimates of net
               organic matter overtime that i
               network design.
                                         rr icrobial transformations of contaminant-related
                                         tmy be built into an efficient long-term monitoring
The first three objectives establish the environment of major contamination and the conditions
under which bioremediation may occur. The later two objectives are vitally important, because
evaluating  the  progress of intrinsic  bioremediation  processes  depends on  distinguishing
compound  "losses"  due  to dilution, sorption, and  chemical  reactions  from  microbial
transformations. This approach has been suggested emphatically by Wilson (.9) and was recently
developed  into a  draft technical U.S. Air Force (USAF) protocol by Wiedemeier et al. (10).
The latter reference focuses directly on the implementation of intrinsic remediation for dissolved
fuel contamination in ground water. The general approach is shown in Figure 1, which has been
modified from the original work. The draft USAF protocol (10) has as its goals the collection of
data necessary to support:

        •       Documented loss of contaminants at the field scale
        •       The use of chemical analytica

        •       Laboratory microcosm studies

These data, if collected in three dimensions
implement intrinsic  remediation  successfully
characterization effort (Figure 1) support the d
                                           data in mass balance calculations

                                           using aquifer samples collected from the site

                                           :or an extended period, should be sufficient to
                                            (11).  The data  collected  in the  initial  site
                                           evelopment of a site-specific conceptual model.
This model is a three-dimensional representatijon of the ground-water flow and transport fields
based on geologic, hydrologic, climatologic, and geochemical data for a site. The conceptual
model,  in  turn,  can  be tested, refined, and used  to determine  the suitability of intrinsic
remediation as a risk-management strategy. The validity of the conceptual model as a decision
tool depends on the complexity of the actual hydrogeologic setting and contaminant distributions
relative to the completeness of the characterization database. The draft USAF protocol is quite
comprehensive in identifying important parameters, inputs, and procedures for data collection
and analysis.  The major categories of  necessary data are  listed in Table 1 from the draft-
protocol (10). Ongoing work on the protocol has revised some of the  detailed guidance it
provides on sampling and analytical protocol! for these critical parameters; thus, recent drafts
of the protocol should be even more useful to practitioners.
 Typical detective monitoring data sets availab
 likely to contain contaminant-related informcjiti
 property, hydrogeologic, or geochemical da
 recognition of  the  variability inherent  in tl
 characterization efforts.
                                           3 prior to in-depth site characterization are more
                                           ion rather than the three-dimensional aquifer
                                           ra  needed to formulate a conceptual model. A
                                           lese parameter distributions is  critical to site

-------

	
Review AvalaWe
SHe Data
4f
Develop Preliminary
Conceptual Modal
Jr
Make PreEminary
Assessment of Potential
For Intrinsic Remediaflon
Based on Existing Sits
Characterization Data
• Ccrtmww* Typ»
KidDcMajten
- Hydmosofagy
• LooCon oJ Rece$(oa

V f
Perform Sfta Characterization 1
In Support of Intrinsic Remediation f
4
Refine Conceptual Model and
Complete Pre-Mode5ng
Cateuiatfons
A
Document Occurrence of
. Intrinsic Remediation and
J Model Intrinsic Remediation
/ Using Nunerieal Models
4
Use Results of Modeling and
Sfta-Spetific Information in
an Exnosura Assessment
.-^\^*^-. ^s

v Evaluate Use of


> Conjunction With
Intrinsic Remediation
^l^\\
^acov8*^ / 111 XL^^ISI^J
rrcfei \
Tmt 1 1 1 Sparging
IBI^^}— J|l — I'Ssfj
iErtuncad
BJorefnediioi

^^Unacceptable Risk ra"*^.
^^s^Potential Recsptorsjx"^
^s^^X^ NO
wv
jv Examine Temporal/
•r Spatial Distributions
Sits Point-O(-Complianca
v Monitorina Weds and
' MonHorinqPlan

d> r__™^™T 	
Prepare Refined
Monitoring Plan

MBsera Midngs
andLong-Tenn
	 \ Monitoring Plan To
7 RegUaioiy Agencies
and Reach Agreement
en Monitoring Strategy

I
Assays Potential For
.,.._„,..._.>> IrtJinsir SnrrKvflaffon
^ W* Remedaflon
System Instaled
V
Refine Conceptual Model and
Complete Pre-Modefng
Calculations
V
Model Intrinsic Rsmedatfon
Combined with Remaotal
Option Selected Above
Using Numerical Models
4
Usa Results of Modeling and
Site-Specific Information in
an Exposure Assessment
YES ^X^s^^
^'iJnaccaptabla Risk To^*^.
^•s^Potential ReceptofsTX'^
NO 	 ^^^^>^
Figure 1. Intrinsic remediation flow chart.

-------
Table 1. Site-Specific Parameters To Be Determined During Site Characterization (10)
  Fractionation and
  Spatial Extent of
  Contamination
Extent and type of soil and ground-water contamination

       Location and extent of contaminant source area(s) (i.e.,
       areas containing free- or residual-phase product)

       Potential for a continuing source due to leaking tanks or
       pipelines
  Hydrogeologic
  and Geochemical
  Framework
Ground-water geochemical parameter distributions (Table 2)

       Regional hydrogeology, including:

              Drinking water aquifers
              Regional confining units

       Local and site-specific hydrogeology, including:

              Local drinking water aquifers
              Location of industrial, agricultural, and domestic
              water wells
              Patterns of aquifer use
              Lithology
              Site stratigraphy, including identification of
              transmissive and nontransmissive units
                                    Grain-s
                                    Aquifer
                      ze distribution (sand versus silt versus clay)
                      hydraulic conductivity determination and
                                    estimates from grain-size distributions
                                    Ground-water hydraulic information
                                    Preferential flow paths
                                    Location and type of surface water bodies
                                    Areas of local ground-water recharge and  discharge

                             Definition of potential exposure pathways and receptors
 Sampling in Space

 The initial site characterization phase should I
 of critical data over volumes corresponding to
 flow paths.  If the flow path intersects a
 should be scaled accordingly. For example, if
 volume would be  1  yr of travel time. The '
 hydrogeologic and geochemical parameters
 derived from data sets that are large enough
 mean, median, correlation distance, and
                 variance
     ie designed to provide spatially dense coverage
     1 0-yrto 1 00-yrtravel times along ground-water
discharge zone in less than 1 00 yr, then the volume
     the flow path discharges  after 10 yr, the critical
     /olume-averaged"  values  of the contaminants,
     within zones along the flow path(s),  should  be
    to permit estimation of statistical properties (e.g.,
         . In general, this means that the data sets for

-------
derived mass loadings of contaminants, aquifer properties, and geochemical constituents (Table
2) derived from spatial averages of data points must include approximately 30 or more data
points (12-14).  Indeed, this minimum data-set size strictly applies to points in a plane.
Table 2. Target Constituents for Site Characterization in Support of Intrinsic Bioremediation
Contamination
Area
Source
Downgradient
Upgradient/
Far-field
downgradient
Apparent/
Geochemical
Redox Zone
Reducing
Transitional/
Suboxic
Oxic
Contaminant
Mixture
Fuels
Chlorinated
solvents
Fuels
Chlorinated
solvents
Fuels
Chlorinated
solvents
Inorganic
Constituents
O2, CO2, H2S;
pH, Fe2+,
HS-/S-, NCy,
NH3, alkalinity
02, C02, H2S;
pH, Fe2+,
alkalinity, NO2",
NCy, NH3,
HS-/S-
02, C02, H2S;
alkalinity,
Fe2+, NO3-,
NCy, NH3
Intrinsic Constituents
Organic carbons,
CH4, organic acids,
phenols
As above and:
chlorinated
metabolites,
ethylene, ethane
Organic carbon,
CH4, organic acids,
phenols
As above and:
chlorinated
metabolites,
ethylene, ethane
Organic carbon,
CH4, organic acids,
phenols
As above and:
chlorinated
metabolites,
ethylene, ethane
Two major decisions  must be made with regard  to  how spatially  averaged  masses  of
contaminants, electron donors  (e.g.,  organic carbon, Fe2+, S=, and NH3), and electron
acceptors (e.g., O2/ NO3", NO2",Fe and Mn oxides, and SO4=) are to be estimated.

The first question deals with identification of the media in which the bulk of the constituent's
mass  resides. For aquifer  properties (e.g., grain size and  laboratory estimates of hydraulic
conductivity), the answer is simple. In this case, the solids are clearly the media of interest. For
constituents, particularly VOCs, which are sparingly water soluble, the bulk of the mass may in
fact reside in the solids, though both solids and water samples must be collected carefully.

-------
The second question pertains to the depth intsrval over which "planar" data points might be
averaged. With fuel-related aromatic contaminants, the depth interval above and below the
capillary fringe/water table interface typically exhibits order-of-magnitude differences in solid-
associated concentrations. In this situation, averaging data points over depths greater than 0.5
m could easily lead to order-of-magnitude errors in estimated masses for a site. Continuous
coring of subsurface solids and close interval  (i.e., <1 m) sampling  of water should be
considered in many VOC investigations. To approach this level of depth detail in sampling,
"push" technologies  and/or  multilevel sampl
characterization. Push technologies rely on hycraulic or hammer-driven, narrow diameter (i.e.,
<2 in.) probes for solid or water sampling. These technologies have the potential to provide
greater spatial coverage of the subsurface at
                                           ess cost than drilling techniques.
The approach to site characterization forchloririati
Very few models of site characterization for
in specific media. Many of the previously
detection, assessment, and quantitation,
sampling than intuition based  on experience.
                                       "these
                                    'referenced
                                           ng  devices  present very useful tools  for site
        ed hydrocarbons is significantly more difficult.
         contaminants have estimated mass loadings
          methods may work satisfactorily. Free-phase
however, may be more a matter of luck and exhaustive
Sampling Over Time

VOC compounds (e.g., aromatic hydrocarbons and chlorinated solvents) are among the target
contaminants  that are considered constituents of concern in remedial investigations. Their
aqueous solubility and demonstrated associatiDn with aquifer solids require sampling of these
media during the site characterization phase. Tr
of complex organic mixtures  (e.g.,  ethylene,
                                           is suggestion also applies to organic metabolites
                                           vinyl  chloride, aromatic acids, and phenols).
Aqueous plumes that develop subsequent to the release of these organic mixtures and byproduct
compounds have  received the most attention in the past. The fact that the mass  of these
contaminants frequently resides in the solids strongly suggests that the solids should receive the
most attention in the  initial  site characterizati*
physical, geochemical, and  microbial determi
                                               ffort. This should also  be the case for th
                                           lotions.
Initially, conventional nested monitoring wells with screened  lengths of 1  m or more will  be
useful for estimating the spatial  extent of fie dissolved  plume, for delineating apparent
geochemical zones, and for providing data on
                                           water level and aquifer property (e.g., slug- and
 particularly multilevels appropriately designed
 course of the long-term monitoring program. S
pump-test derived hydraulic conductivity estimates). Semiannual or annual sampling of wells
                                           and completed, should be quite useful over the
                                           impling should track the downgradient progress
of risk-associated target compounds and permit testing predictions of intrinsic bioremediation
effects on risk reduction.

Proof of the effects of the net removal of specific solid-associated contaminants due to intrinsic
bioremediation, however, will depend on solid sampling and  analysis  at annual or greater
intervals,  because solid-associated concentrations may be expected to change slowly. Unless
 biotransformation can be shown to be a major
 over extended periods, it will remain an area

 Because very few contamination situations have
 several years, it is difficult to define specific sanr
                                           oss mechanism for contaminants mainly in solids
                                           of research rather than practice.

                                           been monitored intensively for periods exceeding
                                           pling frequencies forthe range of hydrogeologic

-------
and contaminant combinations that may be encountered.  The adoption and future refinement
of  recently developed, technically  defensible  protocols  will improve intrinsic remediation
approaches to risk management in subsurface contamination situations.
Acknowledgements

The author would like to express his gratitude to the following individuals who aided  in the
preparation of the manuscript: Dr. Gary Robbins, Mr. Todd H. Wiedemeier, Dr. John T. Wilson,
Dr. Fran Kramer, and Ms. Rebecca Mullin.
References

1.     Sea If, M.R., J.F. McNabb, W.J. Dunlop, R.L Cosby, and J.S. Fryberger. 1 981. Manual
       of ground-water sampling procedures. National Water Well Association.

2.     Barcelona, J.J., J.P. Gibb, J.A, Helfrich,  and E.E. Garske. 1985. Practical guide for
       ground-water sampling. Illinois State Water Survey, SWS Contract Report 374 U.S.
       Environmental Protection Agency, Ada, OK.

3.     U.S.  EPA.  1 986.  RCRA technical enforcement guidance document, OSWER-9950.1.
       Washington, DC.

4.     Eklund, B.  1 985. Detection of hydrocarbons in ground water by analysis of shallow soil
       gas/vapor. API Publication No. 4394. Washington, DC.

5.     Plumb, R.H. 1 987. A comparison of ground-water monitoring data  from CERCLA and
       RCRA sites. Ground Water Monitor. Rev.  7:94-100.

6.     Robbins, G.A. 1989.  Influence of using purged  and partially penetrating  wells on
       contaminant detection, mapping, and modeling. Ground Water 2:1 55-1 62.

7.     U.S.  EPA. 1992. RCRA ground-water monitoring: Draft technical guidance document.
       EPA/530/R-93/001. Washington, DC.

8.     U.S.  EPA.  1 994.  Proceedings of the Ground Water Sampling Workshop, Dallas, TX,
       December 8-10,  1993. U.S. Environmental Protection Agency, Ada, OK.

9.     Wilson, J.T. 1993. Testing bioremediation in the field  In: National  Research Council.
       In situ bioremediation—when does it work? Washington, DC: National Academy Press.
       pp. 160-184.

10.    Wiedemeier T.H., D.C. Downey, J.T. Wilson, D.H. Kampbell, R.N. Miller,  and J.E.
       Hansen. 1994. Draft technical protocol for  implementing the intrinsic remediation
       (natural  attenuation)  with  long-term  monitoring option  for dissolved-phase fuel
       contamination in ground water. Air Force  Center for Environmental Excellence, Brooks
       AFB,  San Antonio, TX (March).

-------
1 1.     National  Research  Council.  1 993.
       Washington, DC:  National Academy
n situ  bioremediation—when does it work?
 ress.
12.    Journel, A.G. 1 986. Geostatistics: Models and tools forthe earth sciences. Math. Geol.
       18:119-140.

1 3.    Hoeksema, R.J., and P.K. Kitanidis. 1 985. Analysis of the spatial structure of properties
       of selected  aquifers. Water Resour. Res. 21:563-572.

14.    Gilbert, R.O.,  and J.C. Simpson. 19B5. Kriging for estimating spatial  patterns of
       contaminants:  Potential and problems.! Environ. Monitor. Assess. 5:113-135.

-------
Processes Controlling the Distribution  of Oil, Air, and Water

John L Wilson
Department of Geoscience, New Mexico Institute of Mining and Technology,
Socorro, NM
Abstract
Oils  and  other nonaqueous-phase liquids  (NAPLs) are  a  major source  of  dissolved
contamination in aquifer systems. Three major forces control the movement and distribution of
NAPLs, as  well as air and water, in both porous and fractured media. These are  capillary,
viscous, and buoyancy forces. These forces interact with the complex pattern of geological
features, making fluid behavior difficult to predict and fluid distribution difficult to characterize.
Intrinsic bioremediation within a  NAPL-contaminated zone is  probably of limited effectiveness
because of mass transfer limitations, toxicity concerns, and the limited  availability of other
nutrients. Thus,  intrinsic bioremediation is aimed at the contamination in the downgradient
aqueous phase of the  dissolved plume.  Evaluating the effectiveness of bioremediation  is
impossible within  the  plume without knowing  the distribution of the  NAPL—the plume's
source—or determining whether that source is  still moving.
Introduction

Oil or another NAPL, such as gasoline or trichloroethylene (TCE), may be released at or near
the ground surface.  These  liquids  are primary sources for dissolved  contaminant plumes in
ground-water systems. Even if the NAPL has ceased to move, trapped by capillary forces ds we
describe below, it remains a long-term source of dissolved contamination. The limited aqueous
solubility of the chemicals composing these oily liquids implies that even in small volumes they
can lead to ground-water plumes of enormous dimensions. Intrinsic bioremediation is aimed at
the contamination in the aqueous-phase  plume, for reasons that will  become clear. The plume
location and concentrations depend on understanding the spatial pattern of its source, the NAPL.
In the vadose zone, the problem is complicated by the presence of a third phase, air, and the
propensity of the  NAPL to spread at the air-water interface and to volatilize.

In this review, we first describe the three major forces controlling the movement and distribution
of fluids in the subsurface, using natural processes such as infiltration to illustrate. We then add
NAPLs, to relate the discussion to the contamination issue, and aquifer heterogeneity, to relate
the discussion to natural hydrogeological systems. Finally, we discuss the implications for intrinsic
bioremediation. Many of these issues are illustrated with photomicrographs taken of appropriate
processes ;n situ.
Three Major Forces That Control  Processes

Ground-water systems are composed of porous and/or fractured aquifer material containing
water in the spaces between the solids  (see Figure  la). Above the water table, in the vadose
10
Symposium on Intrinsic Bioremediation of Ground Water

-------
zone (see Figure 2), air is also present within thi
control both the movement and distribution of
forces, and gravity or buoyancy forces (1).
between the solid phase and each of the fluids.
inversely proportional to the pore size. In the va
                                           ; pore space (see Figure 1 c). Three major forces
                                             h of the fluid phases: capillary forces, viscous
eac
Capillarity is the result of the cohesive forces within each fluid phase and the adhesive forces
                                           \capillaryforce is proportional to the interfacial
tension at the fluid-fluid  interface and the strength of fluid wetting to the solid  surface, and
                                           dose zone, with air and water present, interfacial
tension is the same as air-water surface tension. In the saturated zone, beneath the water table
(see Figure 2), the only fluid is water, and capillary forces are absent (see Figure la).  The
exception occurs beneath a fluctuating water table, where gas bubbles may become entrapped
by  capillary forces  (see  Figure 1 b and  Figu'e 3;  Figure 3  is a  photomicrograph  from a
visualization experiment (see also the appendix to this paper). For most aquifer materials, water
is the wetting fluid; that is, the solid aquifer ma
and the air or gas is almost always the nonwetting phase (see Figures 1 b and 1 c). Water occurs
as an interconnected film or layer of wetting
vadose zone. The nonwetting gas phase occupies the larger pores.
Viscous or flowing forces within a fluid phase
                                           erial has a greater affinity for water than for air,
                                           quid covering and connecting the solids in the
                                           require an expenditure of potential energy. For
in
example, in the saturated zone the ground-voter flow rate  is proportional to the hydraulic
gradient. The flow rate is also a function of tre aquifer material and the structure of its pore
space, as represented by the effective permeability. If more than one fluid phase is present, as
   Figures  Ib  and  1 c, the  interconnected paths  in  each  phase  are more tortuous.  The
                                           relative permeability closer to zero than to one,
                                           occupies the pore space.

                                           ho the density difference between two fluids. The
permeability of each phase is reduced, with a
the value that applies when  only a single fluid

Buoyancy is a gravitational force proportional
gas  phase has a  much  lower density than water, so that gravity (buoyancy) forces play a
significant role in the vadose zone. Water infiltrates downward, toward the watertable, primarily
under the influence of gravity. As it moves downward, water easily displaces the less dense and
less  viscous air phase. In the saturated zone, j/ith only water present, gravity usually plays no
direct role (although, of course, it ultimately drives the hydraulic gradient). In the saturated zone,
if the chemical concentration varies enough td influence water density, gravity can again  play
a direct role, causing the water and its chemicals to move.
 Oil and Other NAPLs

 A third fluid  is often present at many hazardous waste sites and most leaking  underground
 storage tanb. NAPLs often share the pore space with both gas and water, in the  vadose zone,
 or water alone, in the saturated zone. The NAP^_ may be more dense than water (a dense NAPL,
 or DNAPL), such as TCE, or it may be lighter than water, such as gasoline.  In either case, a
 NAPL is heavierthan air, so it easily moves downward through the vadose zone. Once it reaches
 the vicinity of the watertable, it may continue to move deeper if it is denser than water (DNAPL),
 or it may remain in the vicinity of the water table if it is lighter, as illustrated in Figure 2 (2, 3);
 in the  latter case, the term "water table" begins to lose its meaning. In either case, the NAPL
 leaves behind a trail of residual as it moves.
                                                                                     11

-------
                                 water   air    solid
                                                  •js^sN
                                                S!s&
           (a) Water Saturated  (b) Residual Gas  (c) Continuous Gas
Figure 1.  Diagram of fluid saturation in a porous media (4).
                       hazardous waste site
                                                              ground surface
          vapor -^   *jw-A
            phase \  ''*?:** <;'
            organic
                                                                VADOSE
                                                                 ZONE
                                                            floating NAPL
    phase liquid4$
    saturation
                                                          SATURATED
                                                            ZONE
   capillary fringe
       •M
water table
     residual
     non-aqueous*'
Figure 2. Diagram of the saturated and vadose zones, showing the migration pattern for a
         NAPL more dense than water (left), and less dense than water (right) (3).
12

-------
Most NAPLs appear to be of intermediate wetta
jility in typical aquifer materials; that is, they are
nonwetting relative to water but wetting  relative to gas. Thus, depending on  whether it
encounters gas or water, a NAPL can then display either wetting or nonwetting behavior or both.
Many NAPLs, such as gasoline, also have low hternal cohesion and will spread at a gas-water
interface,  presumably forming a film between the water phase, which because of capillarity
preferentially occupies the smallest pores, and t le gas phase, which preferentially fills the largest
pores. In the  vadose zone, this film interconne :ts the pockets of NAPL, which  even  at residual
saturation should be largely continuous, as shown  in Figure 4. Some NAPLs, such as TCE, have
more internal cohesion and do  not spread, h this case, the sum  of the interfacial tensions
between the NAPL and the water and air, <7OW + aao, exceed the surface tension  between the gas
and  water, aaw  (see Figure 5, 0>0). These NAPLs will  not spread as films.  On a flat water
surface (the gas-water interface), nonspreadind liquids tend to coalesce into lenses that float on
the surface (much as depicted in Figure 5), even though  many of these liquids are denser than
water. In porous media, this leads to complications, as shown in Figure 6.

In the saturated  zone, the NAPL fills the larger pores, while water occupies the smaller pores and
lies as a film between the NAPL and the solid surfaces. After the  NAPL moves on, it leaves
behind a residual saturation trapped by capilla -y forces, which occupies the larger pores. NAPL
blobs occupying one ortwo pore bodies are shown in Figure 7. A similar process occurs in the
vadose zone, when a rising water table entraDS air bubbles, as shown in Figures  2b and 3.
Because of lower interfacial tension, NAPL blob
forming interconnected groups of larger pore
> are often much more complex than air bubbles,
bodies and pore throats, as shown in Figure 8.
So far we have looked at water-wet porous media. Wetting, however, may change for a variety
of reasons, including the adsorption at hazardous waste sites of (polar) organic compounds onto
the solid. Figure 9 shows the saturated zone residual in a pore space with altered wetting and
significant  wetting  hysteresis.  The residual  NAPi.  is not trapped in the  pore  bodies, the
characteristic location fora nonwetting fluid. Some of it is in the pore throats, some in the pore
 bodies, some has an end in both. Because of
 wet conditions  prevailed during  the invasion
he large wetting hysteresis of this surface, water
of the NAPL, while intermediate wet conditions
 prevailed when the water reentered to trap the NAPL. The NAPL-water interfaces in the figure
 clearly show evidence of both histories. In some locations, clearly water wet, we can infer that
 the interface was not disturbed as the water  'centered. In other locations,  the contact angle
 shows the recent displacement of the NAPL by water.
 Heterogeneity

 All of this is complicated by the geology of the ground-water system. The material composing
 an aquifer is  always  heterogeneous. For example, fluvial-deposited materials contain sand,
 gravel, and clay in a complex geometric pattern of geologic facies. Many formations, even those
 containing clay, are fractured and sometimes faulted. Heterogeneities provide preferential paths
 for fluid and chemical migration in an interp ay of the geology with the forces of capillarity,
 viscosity, and gravity. For example, a DNAPL reaching the saturated zone will tend to seek out
 the  heterogeneities with larger pore spaces  ;uch  as  the coarser sands  or, more insidiously,
 fractures where  its movement will be hard to trace. Heavier than water, it tends to move farther
 downward into the aquifer, fingering out the  Bottom of a sand lens or moving snakelike down
 the  inside of a fracture.
                                                                                    13

-------
 Figure 3.  Photograph of a micromodel with entrapped air bubbles.  The bubbles are trapped
           in the pore bodies, which are roughly 200 jum in diameter, while the pore throats are
           50/im in diameter (4).
Figure 4.  A micromodel after a spreading NAPL (Soltrol 130) is drained with air. Air fills the
          pore bodies and is interconnected through some of the pore throats. Water fills the
          other pore throats and some of the pore wedges and is a wetting film everywhere
          else. The NAPL forms a thick film filling some pore wedges and surrounding the air
          everywhere.  This  ubiquitous and interconnected film is particularly thick near the
          water-filled pore throats (3).
14

-------
Figure 5.  Cross-sectional diagram of spread! ig potential for a drop of NAPL floating on the
          air (gas)-water interface, after Adamson (1 0) and others. The water is wetting, the gas
          is nonwetting, and the NAPL is intermediate wetting.
                                                    Water

                                                   N&BLpoefc
Figure 6.  A micromodel after a nonspreading DNAPL (mineral oil) is drained with air. The
          DNAPL is not connected. As it leave's the photo at the bottom, the air is surrounded
          by a film of oil, similar to that seen with the spreading oil. As the air leaves to the
          left, the film no longer exists; it trunc ates somewhere in between. In this area, the air-
          water interface  is dimpled, with sir all  lenses of mineral  oil.  These  lenses are not
          interconnected. The various pockets! of oil, such as on the right side of the figure, are
          also generally discontinuous and not connected (1 1).
                                                                                    15

-------
Figure 7.  A micromodel with a singlet and doublet NAPL blob (dark gray) in a water-filled pore
          space (lighter gray) of an irregular pore network. The pore bodies are up to 1,000
          fim across, and the model was 1 0 cm x 1 5 cm in area (8).
Figure 8.  A micromodel with a  large,  branched, nonaqueous, nonwetting phase blob (dark
          gray) in an irregular pore network (3).
16

-------
Figure 1 0 shows a pore system with embedded
zones of larger pores surrounded by a network
of smaller pores. This simulates lenses of coarse porous materials (e.g., coarse sand) embedded
in a matrix of finer material (e.g., fine sand), the NAPL invaded from the side, mimicking the
lateral migration of a NAPL plume along the bottom of an aquifer. Afterwards, the NAPL was
displaced by water, leaving behind  a residual saturation. The photograph in Figure  10 was
taken at this time. The residual is dominated by the coarse lenses. In the fine material, we see
the typical single- and multiple-pore body blobs, but the coarse lenses have been completely
bypassed. Capillary forces are too strong, and the viscous forces in the moving watertoo weak,
to displace  the NAPL from these isolated coarser zones. Note that the zones with higher
saturated permeability trap the NAPL.
This experiment was repeated at a much highe
was occurring is shown in Figure 1 1. A signifi
- water flow rate. A photograph taken while this
:ant amount of NAPL has been swept from the
coarse lenses, even though the initial condition was essentially the same. In the swept area, on
the right side, there is still some bypassed NAPL, but it is now located on the downstream end
of the lenses. Sufficient viscous forces were generated to overcome the capillary forces that held
NAPL in the coarse lenses. The displacement is incomplete because, as the wetting front reaches
the end of a lens, it closes together and surrounds the lens. At that point, the nonwetting NAPL
in the lens is no longer connected with the downstream  NAPL. These two photographs suggest
that in heterogeneous  systems the  residual nonwetting-phase saturation  is a function of the
structure of the heterogeneities and the fluid h
dissolved plume, we would like to be able to prc
of the geology and the fluid flow history. We use mathematical models to assist with this kind
of prediction. There is a problem, however. Th
3 current generation of multiphase flow models
assumes that there  is a single value of residual and, worse, that we know what it is. These
models offer little assistance with the important
of residual.

These photographs were taken from experiment
to stratigraphic bedding planes—for instance,
story. Because the residual is the source of the
diet its location and magnitude from knowledge
issue of predicting the amount and distribution
; that simulated the movement of DNAPL parallel
as it migrates  along the bottom of an aquifer.
What happens as it moves perpendicular to tie bedding, for example, during its movement
downward from the water table toward the bo torn of the aquifer? Figure 1 2 shows the same
kind of heterogeneous lenses, except for this ch Jnge in geometry. The flow is downward, across
the lenses, with both gravity and viscous components. As the DNAPL encounters its first coarse
lens, it tends to spread horizontally because of c apillary forces. As soon as it fills a lens, it begins
to spill downward, but because of the gravity instabilities—DNAPL is heavier than water—this
movement occurs as fingers on the scale of a few pores. When one of these fingers encounters
another, deeper coarse lens, the downward migration of the DNAPL is again arrested, until the
lens is filled. In essence, the lenses promote horizontal spreading, here confined by the lateral
boundaries  of the  model, and  vertical finge
displacement, with lots of water bypassed by the
it is difficult to directly measure DNAPLs in  the
remediation measures.
ring. The process leads  to  a very inefficient
 DNAPL, and illustrates some of the reasons why
field, predict their behavior, or design effective
                                                                                    17

-------
Figure 9.  Residual nonaqueous-phase saturation in a micromodel treated with an alkoxysilane
          of proprietary composition, GlassClad 1 8 (Huls Americal Petrarch, Bristol, PA). This
          treated surface has significant contact angle hysteresis. The water is light gray, and
          the NAPL is darker gray in this image (9).
Figure 10.  Water has swept through a heterogenous micromodel containing a NAPL as a
            nonwetting fluid and water as a residual wetting fluid. The nonwetting NAPL in the
            isolated coarse-grained heterogeneities has been bypassed (3).
18

-------
Figure  1 1.  Water is displacing the NAPL at a higher flow rate through the same heterogeneous
           micromodel. The  flow  is from  right to left.  With larger  viscous forces,  the
           displacement is more efficient, and less NAPL is bypassed. This photograph was
           taken while the displacement was still underway. The bypassing is complete on the
           right side of the model and has not yet occurred at the left side (3).
Figure 1 2.  DNAPL is moving downward through a water-saturated heterogeneous micromodel
            (9).
                                                                                   19

-------
 Influence  on Intrinsic Bio remediation

 The distribution of fluid and solid phases controls the mass transfer and transport of chemicals
 and their  availability for biotransformation by the microbial  population. For all practical
 purposes, biotransformation takes place in the aqueous phase or in contact with the aqueous
 phase. Bacteria colonize surfaces. In water-saturated media, the only surfaces available are the
 solids, but in the vadose zone and in NAPL-contaminated areas there are also fluid-fluid (e.g.,
 air-water, oil-water) interfaces that some bacteria will colonize. These colonies develop in part
 due to  the same  interfacial forces that attract abiotic colloids to solid-liquid or fluid-fluid
 interfaces.  Figure  13 shows fluorescing colloids that are repulsed from the solid surface by
 electrostatic forces but that are attracted to an air-water interface by hydrophobic and capillary
 forces.  Similar observations have been made for microorganisms  (4, 5). Other important
 processes also lead to bacterial attachment and colonization at interfaces that  are unique to
 living bacteria,  with  their complex surface chemistry  and ability to produce extracellular
 polymers. Where NAPLs are concerned, there may also be toxicity effects repelling the bacteria,
 or, if the NAPL is  an important nutrient source,  it may be an attractant. Figure 14 shows a
 bacteria colony in the vicinity of a blob of iso-octane. The bacteria have colonized the solid
 surface and the surface of the blob, and a few were even  observed to be living in the iso-
 octane. In  a similar experiment with toluene and a toluene-degrading bacteria, the bacteria
 found high concentrations of toluene toxic and  set up housekeeping at a distance that allowed
 ambient flow and mixing to dilute the concentrations many times.

 The geometry of the fluid distributions have other indirect influences on intrinsic bioremediation.
 Consider mass transfer of chemical components between a multicomponent residual NAPL and
 the water phase in the saturated zone. Many spilled NAPLs—from gasoline and other fossil fuels
 to transformer oils containing  polychlorinated  biphenyls (PCBs), as well as the  mixed  bag of
 organics that are sometimes found  in industrial waste pits—are  mixtures  having  many
 (sometimes hundreds of) organic chemical components. The more soluble components of these
 mixtures can dissolve comparatively quickly, leaving behind  less soluble components to leach
 out more slowly.

 There are  also cosolvency effects to consider. Interaction  between components can either
 enhance the solubility of a given component (cosolvency) or reduce the solubility  of that
 component (by a kind of salting-out process). Capillary-trapped residual blob size and shape
 influence the partitioning of NAPL components to the aqueous phase. Mass transfer coefficients
 used in the mathematical models of this partitioning often employ the analogy of an equivalent
 spherical blob. Certainly singlet blobs can be represented by a spherical model, but it is less
 clear that this model works for the tortuous  multiple pore-body  elongated or branched blobs
 shown in Figure 8, or in the presence of heterogeneity shown in Figure 1 0. Large complex blobs
 hold the majority of the NAPL  volume,  and  mass transfer from them to the  aqueous phase is
 clearly rate limited  (6-8); it can take decades under natural conditions. Intrinsic bioremediation
 can do  little to speed up this mass-transfer process, except to possibly increase  concentration
 gradients  between the NAPL and the aqueous phase,  or perhaps increase the solubility of
 chemical components through the production of biosurfactants.

 Even  in  the absence of toxicity effects and mass-transfer limitations, there may  be little intrinsic
 bioremediation within the NAPL zone.  The aqueous-phase flow field  is complicated by the
 presence of the NAPL. Among other  things,  the  effective permeability to water  is lower.
 Consequently, it is difficult to supply the nutrients necessary for growth (3, 8).
20

-------
Figure 13.
Negatively charged hydrophilic pclysty
bubbles entrapped in a micromcd
surface at this ionic strength (4).
   rene particles attached to the surface of gas
el. There is almost no sorption on the glass
Figure 14.  A colony of bacteria (Arfhrobacfersp
            after 48 hr of growth. The blob h
            gray (5).
                                 . ZALOOl) in the vicinity of an iso-octane blob
                                trapped in a pore body. The bacteria are light
                                                                                     21

-------
For these reasons, intrinsic bioremediation of the NAPL itself may be of limited value. If instead
we consider the downgradient plume of dissolved contamination, the basic concern with the
NAPL is as a source of that contamination.

To understand the plume, we must know something about the spacial distribution of the source
and determine whether or not it is still moving. This  both locates the source and controls the
effective  mass-transfer rate between the source and the ground water.  If the source is a NAPL,
it is obvious from some of the earlier photographs that even a small amount of liquid may be
spread throughout a relatively large volume of aquifer material, and in a most complex pattern
(see Figures 11  and 12). Characterization of this distribution and movement is difficult and may
be beyond the current state of the art.
Summary

Three mafor forces control the movement and distribution of fluids in the vadose and saturated
zones: capillary, viscous, and buoyancy forces. These forces interact with the complex pattern
of geological features, making fluid behavior difficult to predict and fluid distribution difficult to
characterize. Wetting fluids have an affinity forthe solid surfaces of aquifer materials. In general,
NAPLs are wetting relative to air, but nonwetting relative to water. Consequently, they tend to
fill the larger pores of the saturated zone, leading to entrapment in what are usually regarded
as high-permeability materials: the coarse sand  but not the fine silt. This tendency, together with
the propensity for gravity fingering through the finer  materials, causes a complex pattern of
migration and  distribution in the saturated zone. In the  vadose zone, behavior is further
complicated by the presence of air and the question of whether or not the NAPL tends to spread
at the air-water interface. Wettability can be altered by sorption processes, changing the relative
roles of water and the NAPLs and leading to new pathways and fluid distributions.

Intrinsic bioremediation  within a NAPL-contaminated zone is probably of limited  effectiveness.
Mass-transfer limitations constrain the rate at which dissolved components reach the aqueous
phase. There are toxicity concerns with the high concentrations that are found nearer the NAPL
blobs orfilms. Some other nutrients are only limited in availability because of the more complex
flow patterns that water  must take in this region.

Consequently, intrinsic bioremediation should probably be aimed at the contamination  in the
downgradient aqueous  phase  of the dissolved  plume.  It is  impossible to  evaluate  the
effectiveness of bioremediation within the plume without knowing the answer to two questions:
Where is the source? and, How much contamination is entering the flowing ground water from
the source? The answer to  both questions requires that we know  the distribution of  the
NAPL—the plume's source—and determine whether that source is still moving.  It is a  direct
answerto the "where" question. The "how much" question depends on the effective mass-transfer
rate between the  source and the  passing ground  water. The  trouble is that, even in a
geologically simple aquifer, a small amount of nonaqueous contamination may be spread
throughout a relatively  large volume  of aquifer material, and in the  most complex pattern.
Characterization of this  distribution and movement is  difficult and may be beyond the current
state of the art.
22

-------
Appendix: Visualization of Processes in Micromodels

Micromodels are transparent physical models of a pore space network, created by etching a
pattern onto two glass plates which are then fuLed together (3, 8, 9). The resulting pores have
complex three-dimensional  structures, although the  network is only two dimensional. The
micromodel in Figure 3 shows pore bodies connected together by narrower pore throats. The
fused glass located in between these connected channels represents the solid material in this
model of a porous media.  Glass is also at the top of the channels, toward the viewer, and
below the channels.  When  the pore  body contains a nonwetting fluid, a  thin water film is
attached to the top and bottom of the pore body. The film is thin enough to partially  exclude
the bacteria from colonizing the top surface of the iso-octane  blob  in Figure 14. All of the
photographs in this paper can be found collected together and described  in more detail in
Wilson (9). Greater detail is given in the original references.
Acknowledgements

The photographs presented in this paper were taken from work sponsored by the Subsurface
Science Program of the Department of Energy', or from work sponsored by the Robert S. Kerr
Environmental Research Laboratory of the U.S.
New Mexico Institute of Mining and Technology
References
 i.

 2.



 3.



 4.



 5.



 6.
Bear, J. 1972. Fluid flow in porojs media. New York: Elsevier.
Mercer, J.W.,  and  R.M.  Cohen
subsurface: Properties, models,
Hydrol. 6:107-163.
                               Environmental Protection Agency. M. Flinsch of
                               assisted with the preparation of this manuscript.
 1990. A review of immiscible  fluids  in the
:haracterization and  remediation. J. Contam.
Wilson, J.L., S.H. Conrad, E. Hagan, W.R. Mason, W. Peplinski, and E. Hagan.
1 990. Laboratory investigation of lesidual liquid organics. ReportCR-81 3571. U.S.
Environmental Protection Agency,

Wan, J., and J.L Wilson. 1 994. V
on the fate and transport of collo
23.
 Ada, OK.

sualization of the role of the gas-water interface
ds in porous media. Water Resour. Res. 30:1 1 -
Wan, J., J.L. Wilson, and T.L. Kiett. 1 994. The gas-water interface as an influence
on  the transport of microorganisms through unsaturated  porous media. Appl.
Environ. Microbiol. 60:509-516
Miller, C.T., M.M. Poirier-McNeil
nonaqueous  phase liquids:  Ma
26:2783-2796.
, and A.S. Mayer. 1 990. Dissolution of trapped
s transfer characteristics. Water Resour.  Res.
                                                                                   23

-------
 7.
 8.
 9.
 10.
 11.
Powers, S.E., C.O.  Louriero, L.M. Abriola, and WJ. Weber.  1991. Theoretical
study of the significance of nonequilibrium dissolution of NAPLs  in subsurface
systems. Water Resour. Res. 27:463-478.

Conrad, S.H., J.L Wilson, W.R. Mason, and W. Peplinsld. 1 992. Visualization of
residual organic liquid trapped in aquifers. Water Resour. Res. 28:467-478.

Wilson, J.L. 1 994. Visualization of flow and transport at the pore level. In: Dracos,
T.H., and F. Stauffer, eds. Transport and reactive processes in aquifers. Rotterdam:
Balkema. pp. 1 9-36.

Adamson, A.W. 1 982. Physical chemistry of surfaces, 4th ed. New York: Wiley and
Sons.

Wilson, J.L. 1 992. Pore scale behavior of spreading and nonspreading organic
liquids in  the vadose  zone. In:  Weye, U., ed.  Subsurface  contamination  by
immiscible fluids. Rotterdam: Balkema. pp. 107-1  14.
24

-------
New Tools To Locate and Characterize Oil  Spills in Aquifers
Bruce J. Nielsen
Environics Directorate, Armstrong Laboratory,
                                           Tyndall AFB, FL
Abstract

The vision of the Tri-Services (Air Force, Army, and Navy) scientists is about to become reality,
as a partnership between the Department of Defense (DOD), academia, and private industry
evolves into a combined technology that can :>ave millions of dollars in long-term hazardous
waste site cleanup costs.
                                         man
DOD has about 20,000 contaminated sites,
also may require monitoring for more than 3C
The cost of site characterization and monitoring
remediation costs.
                                        eth
-------
 This paper briefly describes the Tri-Services program for developing this cone penetrometer
 technology and recent results of the Air Force program.
 Methods

 Cone Penetrometry

 The Site Characterization and Analysis Penetrometer System (SCAPS), developed jointly by the
 Tri-Services, has proven to be an effective technology for characterizing contaminated  DOD
 sites. The Army has provided leadership on developing SCAPS, including the concept of using
 a sapphire window in the cone rod and of using fiberoptics and spectroscopy to analyze the soil
 for  contamination.  The Army  has  patented  a  "Device  for  Measuring  Reflectance  and
 Fluorescence of In Situ Soil" and is  now licensing it.

 The typical  cone penetrometer is mounted on a 20-ton truck and driven to the site requiring
 characterization,  where a  conical rod  is  hydraulically  pushed  into the  ground to be
 characterized. The rod is equipped with a variety of sensors or soil and ground-water sampling
 tools. The cone penetrometer can characterize several aspects of the subsurface, depending on
 the types of sensors integrated into the penetrometer. Strain gauges measure the forces against
 the tip and sleeve of the cone tool, allowing determination of soil type (e.g., sand, silt, clay) and
 stratification. Other sensors provide electrical resistivity, pore pressure, spectral characteristics,
 and other properties of the soil and contamination. The sensors provide information on
 hydrogeology and contamination; the samplers verify it. The real-time ability to receive and
 assess monitoring data on site without laboratory analysis is critical, facilitating decision-making
 during  site  investigation projects while ensuring  accurate and  efficient completion of site
 investigations and optimization of remedial activities.

 Laser Spectrometer System

 Each of the services  has significant programs  for developing  and demonstrating  laser
 spectroscopy and other sensor systems. One of the key components of the cone penetrometer
 is a  neodymium:yttrium aluminum garnet (Nd:YAG) laser, which pumps a dye laser system to
 induce fluorescence of fuel products as the cone penetrometer probe is advanced into soils. LIF
 has  been shown to be useful in identifying petroleum, oil, and lubricant (POL) contamination,
 such as gasoline and JP-4 jet fuel. The Armstrong Laboratory's Environics Directorate, working
 with  North  Dakota State  University  (NDSU),  has developed  a  tunable  laser/fiberoptic
 spectrometersystem, which uses laser-generated ultraviolet light, optical fibers, and spectroscopy
 for hazardous waste site monitoring. The basic detection approach takes advantage of the fact
that certain  substances fluoresce when a particular wavelength of  light shines on them. The
 spectral emission, including fluorescent lifetime, is somewhat like a fingerprint and therefore is
 useful  in identifying the  contaminant.  The fluorescent intensity  indicates concentration of the
 contaminant. The transportable laser system is unique because  its output may be tuned to the
 optimum frequency for detecting the pollutants of interest.

 Optical fibers are used to transmit ultraviolet light to subsurface monitoring points and return
 resulting lightforthe spectroscopic analysis. The system can identify aromatic hydrocarbons such
as benzene, toluene, and xylene (BTX) and naphthalene by their fluorescent spectra.  Jet fuel,
which contains naphthalene and BTX, is the most common contaminant at Air Force sites.
26

-------
This  technology can provide semiqualitative
  and  semiquantitative information, on site,  in
minutes.  The LIF response  can be  correlated  to the  total petroleum hydrocarbon (TPH)
concentration within the soil. The system has been tested  in the field with detection limits as low
as parts-per-million levels on soil when used with a  cone  penetrometerand in the laboratory at
parts-per-billion levels for BTX in water using f beroptic probes. Laser spectroscopy technology
could also be used to monitorthe progress of site remediation and to provide baseline data for
intrinsic bioremediation modeling studies.

Technology Transition

Armstrong  Laboratory and  Unisys Corp. sigred a Cooperative Research  and Development
Agreement (CRADA) to commercialize the Air
 :orce-developed laser spectrometer system. The
laserspectrometerwas initially developed for making ground-water measurements in monitoring
wells or in implanted monitoring points. A consortia consisting of the CRADA partners, Dakota
Technologies Inc., and NDSU submitted a proDosal to the Advanced Research Projects Agency
(ARPA), Technology Reinvestment Project (TRP)
$ 1,600,000 grant with two 1 -year options for
partners provide in-kind contributions and me
of the  laser spectrometer developed for the
  ARPA selected the proposal to receive a 2-year,
 follow-on technology development. The industry
 tching funds.
The Rapid Optical Screening Tool (ROST) is the proposed product from the commercialization
 Air Force  by NDSU.  ROST will  build upon the
previous Environics Directorate research by automating the collection and mapping of data,
making equipment components smaller and more rugged, and developing a more user-friendly
interface to allow use by environmental technicians involved in site characterization and cleanup.
ROST also has potential for process monitoring and for medical diagnostics. Initial commercial
use will be with cone penetrometers for soil characterization.
Use of the proposed ROST technology shoulc
  result in substantial savings in costs associated
with characterization, monitoring, and remediation of hazardous waste sites. The participants
are committed to commercializing the resulting instrumentation for worldwide sales by U.S. firms
or companies.  In short, DOD will benefit from the  technology and knowledge  gained; the
private sector will receive a highly transferable and profitable technology; the U.S. economy will
be helped; and all will benefit from  a cleaner
Combined Technologies

The combined cone penetrometerand transpo
Force installations having fuel-contaminated
  environment.
 table laserspectrometerwas demonstrated at Air
 areas.  To date, the laser spectroscopy system's
primary function with the cone penetrometer h 3s been to define the oily-phase plume. At Tinker
AFB, the tunable laser system was configured to optimize the system for jet fuel and heating oil,
the known petroleum contaminants. Laboratory fluorescence spectra from these fuels suggest
that naphthalene  produces the  maximum
wavelength appropriate for the known fuels vas utilized during the field program.
The system is designed to collect data in two
mode,  laser excitation  frequency  is fixed,
penetrometer probe is advanced. Operation
allows collection of LIF multidimensional data
intensity, and time of decay matrices (WTM).
various fuels.
 luorescence; consequently,  a laser excitation
 different modes:  "push" or "static." In the push
 and the  LIF signal  is monitored  as  the cone
 in the static mode, or with the probe stopped,
 sets, typically fluorescence emission wavelength,
V/TMs have proven to be very useful in  identifying
                                                                                    27

-------
 Results

 Studies to  demonstrate site amenability towards intrinsic demonstrations were conducted at
 Plattsburgh, Patrick, and Dover AFBs. The sites selected for the demonstrations included closed
 fire-fightertraining areas, gasoline service stations, and aircraft hydrant fueling systems. The old
 fire-fightertraining areas were composed of unlined pits used to ignite fuels such as JP-4, waste
 oils, and otherflammable substances. The service stations or hydrant fueling systems had leaking
 underground storage tanks or piping. Substances from these operations have percolated through
 the vadose zone into the unconfined aquifer. The penetrometer determined the areal extent and
 volume of oily-phase contamination, and obtained ground-water and soil  core samples, to
 determine site amenability, acquired data were then fed into BIOPLUME®!!, a computer model
 for in situ contaminant biodegradation. The cone penetrometer system rapidly located and
 defined the leading edge of the oily-phase petroleum plume. The technology proved that it can
 be used to provide timely and accurate data for intrinsic bioremediation modeling.

 The Tri-Services conducted a series of laboratory tests, and some of the preliminary results are
 calibration curves with different fuels on various soil matrices. The calibration curve obtained in
 the laboratory for diesel fuel  marine (DFM) on  a sand matrix  indicates a detection limit that is
 lower than  30  mg/kg  (ppm) (Figure 1). The collection of LIF multidimensional data sets
 (fluorescence  emission wavelength,  intensity,  and lifetime)  or  WTMs for diesel  #2, JP-4,
 unleaded gasoline, and diesel fuel marine show how each one has a characteristic pattern.
 These  patterns make possible reliable fuel-type  identification without the need for bringing
 samples to the surface (Figure 2). In  the field, the LIF count measurements can be correlated
 with collected samples and analytical results. To assist in the  correlation, several WTMs were
 conducted at various depths. Color-coded WTMs from the North Tank Area (NTA) and Fuel
 Purge Area (FPA) at Tinker AFB indicate different fuel types. The shapes of these spectra identify
 the contaminants as fuel oil at the NTA and JP-4 at the FPA (Figure 3). The fluorescence versus
 depth  profile from  push  location  84-L at  Plattsburgh  AFB  indicates  narrow bands  of
 contamination are in the "oily phase," which rests just above the water table. Note that discrete
 sampling at 5-tt intervals (25, 30, 35 ft,  etc.) could easily skip over the contamination (Figure
 4). A series of fluorescence versus depth  profiles taken across a north-south  transect at the
 Plattsburgh AFB fire-fightertraining area show  the extent of contamination (Figures 5 and 6).
 Location 84A is upgradient, 84D is in the center of the burn pit, and the  remainder are
 downgradient. The contamination traveled directly down from the burn pit and  then along the
 water table.
Conclusions

Currently, these technologies are being further developed and demonstrated within numerous
DOD, Department of Energy, and EPA programs. Ongoing research will develop techniques to
monitor contaminants such as chlorinated solvents, metals, and explosives that do not naturally
fluoresce. Refining this technology is at the heart of site remediation because it provides cost-
effective characterization before, during, and after remediation. We can use it to determine if
remediation is needed, what remediation technology we should apply, whetherthe remediation
is working, and whether the cleanup effort has  been successful—all with a minimum of risk,
time, labor, and cost.
28

-------
  100,000
   10,000
1,000
 c
 U)

55
  100
       10
                  i i I I I I	1	1	L
                    10
                                o-
                                      •o-
                                     -o-
                           100
1,000       10,000
                Concentration DFM (ppm)
Figure 1.  Calibration curve for diesel fuel mcirine on fisher sand.
                                                         29

-------
       Diesel #2
Diesel Fuel Marine
on
70
^ 60
8.50
o 40
.1 30
H 20
10
0
3(

A
)0 350 400 450 5C
IZU
100
80
60
40
20
0
)0 3(

•*•».,


)0 350 400 450 500
Wavelength (nm) Wavelength (nm)
80
70
^ 60
c 50
o 40
.1 30
H 20
10
0
3C
JP-4



'-
)0 350 400 450 5C
I
80
70
60
50
40
30
20
10
0
10 3(
Unleaded Gasoline



ite*




)0 350 400 450 500
     Wavelength (nm)
    Wavelength (nm)
Figure 2. Wavelength time matrices of various fuels.
30

-------
  120

  100


1  80

E  60

    40

    20
      0
       300
       300
                            NTA
                 350       I  400         450
                      Wavelength (nm)
                 350
       400
Wavelength (nm)
450
Figure 3. Wavelength time matrices at two di
       Area.
                               500
500
                            ferent sites: North Tank Area versus Fuel Purge

-------
                 0
           
-------
Figure 5.  Base map of the fire training area showing location of CPT soundings and wells.
                     Fuel Pit
                   84D-LIP
                   •p •
                   1
                                       SECTION A-A-
Leading Edge of Oily Phase

      J=rLIF  BAF.LIP
                     03 GrSand
                       S«nd
                     CD Sand Mix


                    x Sampl*: Water L»v«l (PE)
          O       1OO     2OO

        VERTICAL EXAGGERATION: 6
Figure 6.  Cross section showing  contaminated zone and water table  along section A-A' in
          Figure 5.
                                                                                       33

-------
References

1.        Bratton, W.L., J.D. Shinn, S.M. Timian, G. Gillispie, and R. St. Germain. 1 993. The
          Air Force Site Characterization and Analysis Penetrometer System (AFSCAPS), Vols.
          I-V. AL/EQ-1993-0009.

2.        Gildea, M.L., W.L. Bratton, J.D. Shinn,  G. Gillispie,  and R. St. Germain.  1 994.
          Demonstration of the Air Force Site  Characterization and Analysis Penetrometer
          System (AFSCAPS)  in support of the intrinsic bioremediation (natural attenuation)
          option (interim report).

3.        Schroeder, J.D., and S.R. Booth.  1991.  Cost effectiveness  analysis of the Site
          Characterization and Analysis Penetrometer System (SCAPS). Los Alamos National
          Laboratory, Los Alamos, NM.
34

-------
Microbiological and Geochemical Degradation  Processes
E. Michael Godsy
U.S. Geological Survey, Menlo Park, CA
Introduction

Ground-water contamination is perhaps the
impact of toxic chemical spills and leaks. The
foremost importance. In terms of treatment pro
                                          nost dangerous and intractable environmental
                                          prevention of ground-water contamination is of
                                          :esses, the present technology of excavation and
relocation of contaminated soils to "secure landfills" (which are seldom secure) and "pump and
treatment" of contaminated ground water has proven to be totally inadequate; these processes
just transfer contaminants from one environmental phase to another. Bioremediation, however,
achieves contaminant decomposition or immobilization by exploiting the existing metabolic
potential of microorganisms with novel catabolic functions derived through selection. In response
to the introduction of a toxic contaminant,  an indigenous  bacterial population arises that is
unique from the standpoint of physiological capabilities and species diversity.

Conditions that restrict  life or inactivate microbial enzymes are  incompatible with intrinsic
bioremediation efforts. Although the physical and chemical characteristics of the contaminants
and the metabolic potential of microorganisms determine the feasibility of biotransformation
reactions, actually achieving  biotransformation also depends on the prevailing geochemical
conditions. The ecological  constraints  to  bioremediation  can be classified as  microbial,
                                           of these constraints, as required, assesses the
                                           article focuses   on  the  microbiological and
chemical, or environmental, and recognition
feasibility of intrinsic  bioremediation.  This
decomposition of reduced organic matter (heh
geochemical constraints influencing intrinsic bioremediation processes.
Bacterial Metabolic  Diversity

Bacteria comprise a large and diverse group :>f microorganisms that obtain their energy from
a  variety of  sources,  including  1)  light (j/iotosynfhef/c bacteria), 2)  the  oxidation  or
                                           -otrophic bacteria), or 3) the oxidation of reduced
inorganic compounds (autotrophic bacteria). Some bacteria derive energy from more than one
source, such as combinations of light and reduced inorganic or organic compounds; however,
heterotrophic  bacteria are the major group
compounds.
                                          responsible for the biodegradation of organic
All living organisms must generate reducing
systems and maintaining the oxidation-
oxidized compounds by the addition of elect;
energy production. The electron acceptor can
For many bacteria, most fungi, and all higher
process termed aerobic respiration. The final
and  the final oxidized compound respired
                                       from
     power for the purpose of replenishing enzymatic
reducfon power cycle. This involves the reduction of
      ons released from compounds oxidized during
        either an organic or an inorganic compound.
     organisms, the final electron acceptor is O2 in the
       duced substrate in aerobic respiration is H2O,
        energy production is CO2.
Symposium on Intrinsic Bioremediation of Ground Water
                                                                                    35

-------
 In the absence of O2, certain  bacterial populations respire other less  oxidized inorganic
 compounds (inorganicrespiration) or use only organic compounds (fermenfaf/onj.Denitrification
 is a process promoted by bacteria that can thrive under either aerobic or anaerobic conditions
 (facultative anaerobes) by using O2 as the terminal electron acceptor, when available, or, in the
 absence of O2, by using NO3" as the terminal electron acceptor  (inorganic respiration). When
 oxidation-reduction potentials within soils are even lower, other inorganic compounds are used
 by specific groups of bacteria as terminal  electron acceptors. Because O2 is toxic to these
 bacteria, they are called obligate anaerobes. Several common alternative electron acceptors and
 associated bacterial groups include ferric iron (iron reducers), SO42" (sulfate reducers), and CO2
 (methanogens).
Incorporating  Contaminants Into Bacterial Ecosystems

The ecosystem as a whole can be thought of as a series of integrated oxidation-reduction (redox)
reactions driven ultimately by the radiant energy of the sun. Microorganisms catalyze many of
these reactions and play an essential  role in maintaining the electron balance of complex
ecosystems. For a contaminant to be incorporated into these redox reactions, it must be able
to serve as either an electron donor or an electron acceptor. Moreover, its tendency to  either
donate or accept electrons—and thus be oxidized or reduced—depends on the chemistry of the
compound. For example, many halogenated organic compounds are highly oxidized relative to
their nonhalogenated counterparts, and thus tend to accept electrons and to  be reductively
dehalogenated. The halogenated compounds must compete with other physiological electron
acceptors in order to be incorporated into microbial energy cycles. Thgs, the effectiveness :of
reductive dehalogenations is often influenced by the presence of other electron acceptors, such
as NO3" or SO42". This effect would explain why reductive dehalogenations are more frequently
observed under methanogenic conditions, where generally a paucity of these onions exist.


Bacteria-Contaminant Interactions in Aquifer Material

A minimum of three conditions must  be met before a  contaminant can be  degraded or
transformed by bacteria: 1) the bacteria must be in the immediate vicinity of the contaminant;
2) the contaminant must be available to the bacteria; and 3) the bacteria must have the capacity
to participate in some part  of the degradation or transformation process. Specific bacterial
populations preferor require particularenvironmental conditions. If these conditions do not exist,
these populations tend to become quiescent until more ideal conditions return or develop, or
in some cases they may even die off. The nature of limiting environmental factors often dictates
which  bacterial populations exist.  While  the  subsurface  environment can  be  modified,
modification is often accomplished with great difficulty and always at great expense.

Bacterial Nutrition

Bacteria are composed of combinations of elements that are the components of their genetic
material, structural molecules, enzymes, and intracellular plasma. Because of the great diversity
among bacteria, the proportion of nutrient elements required for growth varies widely; however,
the major required elements  that make up bacteria are carbon, hydrogen, sulfur, nitrogen, and
phosphorus. In soils and aquifer materials contaminated with most organic compounds, carbon
and hydrogen are not typically limiting because they  are the major components of organic
36

-------
compounds, and sulfur is generally available ir
limiting elements for growth are nitrogen and
ratio (C/N/P) usually considered ideal is 300
vary depending the nature of the contaminant
  sufficient quantities for growth. Thus, the major
  phosphorus. The carbon/nitrogen/phosphorus
 fo 1 00:10:1 to 0.05 (1); however, this ratio can
 (s).
pH and Redox Potential
Near-neutral aquifer pH  values are usually
organic material. The hydrogen ion concentrat
of compounds produced by bacterial activity, a
equilibrium rates (2). Because hydrogen ion transfer is
pH  and redox potential are interdependent (3
O3timum for the biodegradation of contaminant
  on of the ground water is governed by the types
  id is controlled especially by CO32":HCO3":CO2
        commonly involved in electron transport,
The redox potential, termed Eh, is extremely important in the biotransformation schemes of both
organic and inorganic contaminants. Usually] a heavily contaminated site is anoxic because
ongoing bacterial respiration has depleted all available O2. The resulting anoxic conditions tend
to favor different electron acceptors, with the most oxidized compounds (higher Eh) being used
first. The resultant scheme is NGy (denitrification) utilization after O2  depletion, Fe3+ (iron
reduction)  utilization  after NGy depletion, SO42~ (sulfate reduction)  utilization  after Fe3+
depletion, and finally CO2 reduction to CH4 after depletion of the available SO42" . As a  result,
bacterial populations having different degrade ive potentials can be operative at different times
at the same contamination site as the redox potential varies.

Although Eh measurements can provide valuaDle clues about the functioning of geochemical
systems in  aquifer material, they alone will not give definite  information about the chemical
species present. These measurements indicate
such as those involving bacterial respiration.
  hat a potential exists for certain redox reactions,
  Various organic and inorganic redox reactions
activity and rates of organic matter decompos
temperatures, however,  bacteria are  generc
temperatures (4). Perhaps more importantly,
cannot be predicted because of their complexit/ and different interdependent reaction rates (2).

Temperature

Ground-water temperature is  often one of the most important factors  controlling microbial
  tion. Generally, rates of enzymatic degradation
and bacterial metabolism double for every 10°C increase in temperature until close to inhibitory
temperatures, which are usually around 40°C to 50°C for most bacteria. Except for subfreezing
  lly capable  of  degradation at most ambient
  •emperature can influence biodegradation of a
compound or contaminant mixture by changing its physical properties, bioavailability, ortoxicity
to bacteria. For example, an increase in temperature usually increases the equilibrium vapor
concentration,  resulting in an increased vola
sorption to aquifer particles (5).

Physical Deterrents to Biodegradation
  ilization rate, but it can at times also increase
 Physical or physicochemical factors can also effect the biodegradation of contaminants. Some
 molecules are recalcitrant to degradation because they are too large to enter bacterial cells,
 which is  usually required  for complete  degradation by membrane-bound  enzymes. Some
 substances are difficult to biodegrade because the number, length, or  location of functional
 groups impede enzyme attack. Strong sorption
 of bacteria to attach to, absorb, or enzymatically attack the molecule (6).
  on aquifer material can greatly hinderthe ability
                                                                                     37

-------
Sorption and solubility of organic contaminants are complex interdependent phenomena that
vary with the  composition  of the aquifer material and complex  contaminant mixtures.  For
example, aromatic hydrocarbon concentrations in water extracts of 31 gasoline samples varied
over an order of magnitude (7). Although gasoline variability could account forthis, the solubility
of each component of mixtures can vary from ideal conditions, with each compound acting as
a cosolventto increase  hydrophobic hydrocarbon solubilities (8).  Organic solvents can also
affect the sorption of organics on soils in general (9).
Contaminant Metabolism: Aerobic Versus Anaerobic

A common misconception is that all organic contaminants are biodegraded most rapidly and
thoroughly under aerobic conditions. Although this is commonly the case, anaerobic conditions
promote some very important degradative processes. For example, compounds that are highly
oxidized, such as polychlorinated biphenyls or chlorinated solvents, are more susceptible to
reductive processes than to oxidative processes  (10) during the initial stages of mineralization.

Most organic compounds found in crude oil, refined oils, and fuels are known to degrade under
aerobic conditions (4); however, current research efforts have shown that the biodegradation of
many monoaromatic compounds common to most fuels also occurred in the laboratory under
anaerobic conditions. This biodegradation readily occurs not only with NO3" serving as the
terminal electron acceptor (1 1) but also under SO42' reducing conditions (12), Fe3+  reducing
conditions (13),  and  methanogenic conditions (14,15).

Benzene has been especially recalcitrantto anaerobic biodegradation in laboratory studies under
denitrifying and sulfate-reducing conditions (1 0,12). Yet some laboratory (1 6) and field studies
(1 7) have shown the  depletion of all common monoaromatic hydrocarbons found in gasoline
under denitrifying conditions. In 1 986, Vogel and Grbic-Galic (1 8) demonstrated that benzene
and toluene were degradable to  CH4 and  CO2 in laboratory microcosms; however,  the
confirmation of the degradation of benzene in field situations under methanogenic conditions
has  been rather elusive. One of the few well-documented  instances of the methanogenic
degradation of benzene in the field  is a crude oil spill in Bemidji, Minnesota (1 9), where many
of the water-soluble  monoaromatic hydrocarbons present in crude  are undergoing intrinsic
bioremediation.
Conclusion

Although bioremediation in general has gained considerable attention, obstacles remain before
bacteria  can be  used  effectively  for detoxifying  wastes  affecting ground water. A lack  of
knowledge or misunderstanding concerning what can and cannot be done with bioremediation
has resulted  in unrealistic expectations, leading in turn to disappointments and ultimate failures.
Continued research into and application of sound bioremediation schemes will undoubtedly
prove the viability of intrinsic bioremediation in the overall  remediation efforts of contaminated
ground water.
38

-------
References

1.      Torpy, M.F., H.F. Stroo, and G. Bruefc
2.


3.


4.


5.



6.


7.



8.


9.
 11.
 12.
 13.
       wastes. Poll. Eng. 21:80-86.
                                  aker. 1 989. Biological treatment of hazardous
Stumm, W., and J.J. Morgan. 1981. Aquatic chemistry. New York, NY: John Wiley &
Sons.

Grundi, T. 1 994. A review of the currsnt understanding of redox capacity in natural
disequilibrium systems. Chemosphere ~~
Leahy, J.G., and R.R. Colwell.  1990.
environment. Microbiol. Rev. 54:305-
 28:613-626.
 Microbial  degradation of hydrocarbons in the
 315.
Lyman, W.J., W.F. Reehl, and D.H. Rosenblatt. 1982. Handbook of chemical property
estimation methods: Environmental behavior of organic chemicals. New York, NY:
McGraw-Hill.

Cheng, H.H., K. Haider, and S.S. Harper. 1 983. Catechol and chlorocatechols in soil:
Degradation and extractability. Soil Bipl. Biochem. 15:31 1-317.

Cline, P.V., J.F. Delfino, and P.S.C. Rod. 1 991. Partitioning of aromatic constituents into
waterfrom gasoline and other complex solvent mixtures. Environ. Sci. Technol. 25:914-
920.

Groves, F.R.  1988. Effect of  cosolve its on the solubility of hydrocarbons  in water.
Environ. Sci. Technol. 22:282-286.

Fu, J.-K., and R.G. Luthy. 1 986.  Effect of organic solvent on sorption of  aromatic
solutes onto soils. J. Environ. Eng. 1 12:346-366.
 10.    Kuhn,  E.P., and J.M. Sulflita. 1989
                                    Dehalogenation of pesticides by anaerobic
       microorganisms in soils and ground water: A review. In: Sawhney, B.L, and K. Brown,
       eds. Reactions and movement of organic chemicals in soils. Special Publication 22.
       Madison, Wl: Soil Science Society of America, pp. Ill -1 80.
Evans, P.J., D.T. Mang, and L.Y. Yourjg
and transformation of o-xylene by
Microb. 57:450-454.
 ^. 1991. Degradation of toluene and m-xylene
cenitrifying enrichment cultures. Appl.  Environ.
 Edwards, E.A.,  LE.  Wills, D.  Grbic-Galic, and  M.  Reinhard.  1991. Anaerobic
 degradation  of  toluene and  xylene:  Evidence  for sulfate as the terminal electron
 acceptor. In: Hinchee, R.E.,  and R.F. Olfenbuttel, eds. In  situ  bioreclamotion:
 Applications  and investigations for hydrocarbon and contaminated site remediation.
        Boston, MA: Butterworth-Heinemann.
                                  pp. 463-471.
 Lovely, D.R., and DJ. Lonergan. 1 990. Anaerobic oxidation of toluene, phenol, and p-
 cresol by the  dissimilatory  iron-reducing organism, GS-15. Appl.  Environ. Microb.
 56:1,858-1,864.
                                                                                  39

-------
 14.    Grbic-Galic, D., and T.M. Vogel. 1987. Transformation of toluene and benzene by
        mixed methanogenic cultures. Appl. Environ. Microb. 53:254-260.

 15.    Wilson,  B.H.,  G.B. Smith, and  J.F. Rees. 1986.  Biotransformations  of  selected
        alkylbenzenes  and  halogenated  aliphatic  hydrocarbons in  methanogenic  aquifer
        material: A microcosms study. Environ. Sci. Technol. 20:997-1,002.

 16.    Major, D.W., C.I. Mayfield, and J.F. Barker. 1988. Biotransformation of benzene by
        denitrification in aquifer sand. Ground Water 26:8-14.

 1 7.    Berry-Spark, K.L., J.F.  Barker, D. Major, and C.I. Mayfield.  1 986. Remediation of
        gasoline-contaminated ground-waters: A controlled  experiment. In: Proceedings of
        petroleum hydrocarbons and organic chemicals in ground water: Prevention, detection,
        and restoration, NWWA/API. Dublin, OH: Water Well Journal Publishing.

 1 8.    Vogel, T.M., and D. Grbic-Galic. 1 986. Incorporation  of oxygen  from  water  into
        toluene  and benzene during anaerobic  fermentative transformation. Appl.  Environ
        Microb.  52:200-202.

 19.    Cozzarelli, I.M.,  R.P. Eaganhouse,  and  MJ. Baedecker. 1990. Transformation of
        monoaromatic  hydrocarbons to organic  acids in anoxic  ground-water environment.
        Environ.  Geol.  Water Sci.  16:135-141.
40

-------
Field and  Laboratory  Results: Getting the Whole  Picture
Mary Jo Baedecker
U.S. Geological Survey, Reston, VA
Abstract
Concern over contamination of ground water in  the last decade has led to an increased
awareness of the need to understand the transport and fate of organic contaminants and the
geochemical  processes that result from their bresence in the subsurface.  A large number of
contaminated sites contain  petroleum-derived hydrocarbons. Many of the hydrocarbons are
biodegraded in ground-water environments, and the extent of their removal by natural processes
has  been evaluated  in  field  and  laboratory investigations.  Both types of  investigations
demonstrate that  natural biodegradation can  be  an important component in remediation
strategies for some contaminated sites.
Results and  Discussion
The processes that control the attenuation of organic compounds in the subsurface .a re complex,
and many investigations have been undertaken in the field and in laboratories to understand
better the factors that control degradation reactions. Parts of contaminant plumes often become
anoxic, and the fate of contaminants at anoxic field sites has been reported in several studies
(1-5). One of the most widespread types of contaminants is petroleum-derived  hydrocarbons
from pipeline breaks, leaking storage tanks, spills, and disposal of wastes. Many of these sites,
such  as those  contaminated by  leaking  small underground storage tanks, are  easier to
remediate than sites with contaminants such as chlorinated compounds. The  number of such
sites is large, however, and much effort has been spent trying to understand processes and to
develop effective remediation strategies to deal with petroleum-derived hydrocarbons. Several
research efforts have been undertaken in the field and laboratory to determine processes that
affect the fate and transport of individual hydrocarbons. In field investigations, degradation of
soluble aromatic hydrocarbons has been shown to occur downgradient from source areas (6-
10). Hydrocarbon-degrading bacteria were found and quantified in soil and  ground water at
a fuel-oil contaminated site (11). Degradation of petroleum-derived hydrocarbons is generally
considered to occur more rapidly in aerobic or suboxic environments, where oxygen or nitrate
is available as an electron acceptor (7,12-1 3)j but anaerobic biodegradation also may remove
significant amounts of hydrocarbons from ground water (14-17).
An investigation of the effects of a crude-oil s
   on an aquifer was undertaken near Bemidji,
Minnesota, as part of the Toxic Substances Hydrology Program of the U.S. Geological Survey.
An underground pipeline carrying crude oil ru Dtured and sprayed oil over land surface. Part of
the oil was removed  during remediation, but part of it infiltrated through the unsaturated zone
and accumulated.as an oil body floating on tne ground water. A more detailed description of
the processes that occurred at the site is avai
Eganhouseetal. (9), Baedecker and Cozzarell
etal. (21).
able in Baedecker et al. (5), Bennett et al. (1 8),
 et al. (19), Cozzarelli et al. (20), and Eganhouse
 Symposium on Intrinsic Bioremediation of Ground Water
                                          41

-------
 The concentrations of benzene, toluene, ethyibenzene and o-, m-, and p-xylene (BTEX) in the
 upper 1.5-m thickness of the aquifer downgradient from the oil body are shown in Figure 1. The
 concentrations of BTEX decreased with  increasing  distance from the oil body and were
 attenuated under anoxic and oxic conditions. Where oxygen was encountered, concentrations
 decreased by several orders of magnitude (56 m to 1 37 m). The mass of BTEX lost near the oil
 body (26 m to 56 m), however, was as high as that lost farther downgradient. The decrease in
 concentrations of BTEX near the oil body (26 m to 56 m) was in the anoxic part of the plume,
 where no oxygen was detected in the ground water over an 8-yr period. Another indicator that
 this part of the aquifer was anoxic is that ferrous iron concentrations were 1.8 mg/L at the water
 table and 71 mg/L at only 0.3 m deeper in the aquifer (22). Ferrous iron precipitates where it
 encounters trace quantities of oxygen. Thus, the high concentrations of ferrous iron near the oil
 body indicate that dissolved oxygen was not transported to the water table. Degradation of
 hydrocarbon vapors in the unsaturated  zone  most likely consumed the  oxygen by aerobic
 respiration (23). The downgradient movement of the BTEX plume averaged about 8 m/yrfrom
 1987 to 1992,  but movement has not been at a steady rate. Near the oil body, microbial
 degradation is the primary process of hydrocarbon attenuation. As the  hydrocarbons are
 transported fartherfrom the oil body, additional processes such as dispersion, mixing (24), and
 sorption become more important.
      ea
      a
      a.
      O)
      .g
      £
      c
      0>
      o
      o
      O
             100
              10
              1.0
             0.1
0.01
           0.001
          0.0001
                                               BTEX
                             50            100          150

                             meters downgradient from oil body
                                                          200
Figure 1.  Concentrations of benzene, toluene, ethyibenzene, and o-, m-, and p-xylene (BTEX)
          in the upper 1.5-m thickness of an aquifer contaminated with crude oil (1 9).
42

-------
Laboratory experiments demonstrated that these hydrocarbons can degrade under controlled
laboratory conditions. To verify that the hydrocarbons were degrading under anoxic conditions,
microcosm experiments were undertaken with sediment and water from the anoxic part of the
plume (5, 20). In two separate experiments, benzene and a mixture of toluene and naphthalene
were added to microcosms under anaerobic conditions. These compounds were also added to
microcosms that were poisoned and sterilized for controls. In the microbially active microcosms,
benzene decreased  in concentration by 98 p srcent in 1 25 d, and toluene decreased by 99
percent in 45 d (5).  No loss of naphthalene
controls, no  loss of  benzene, toluene, or
experiments and field  results provide strong e
  was observed during the same period. For the
  naphthalene  was observed. These  laboratory
  /idence that hydrocarbons are degrading in an
anoxic environment. By comparing the results from the microbially active microcosms with results
from  controls, sorption and  abiological chemical oxidation were eliminated  as  possible
explanations for the loss of benzene and toluene.

Field  and  laboratory studies of anaerobic bictransformation or aromatic hydrocarbons were
reviewed by Barker  and Wilson (25)  who  found evidence  at  five methanogenic  sites  for
biodegradation of benzene,toluene, ethylbenzene,and thexylenes. The estimated half-lives were
0.5 yr and 3.8 yr. The longest half-lives were for benzene. Laboratory experiments also have
indicated biodegradation by several pathways ,(26); however, for benzene the results have been
contradictory. Large concentrations of benzene are biodegraded  in the subsurface (27), yet in
some ground-water  environments  benzene  is the most persistent hydrocarbon among the
monoaromatics.

Natural processes can remove  significant cc ncentrations  of hydrocarbons and prevent the
spreading of  a  plume.  At  sites where the rates of  solubilization, volatilization, and
biodegradation of hydrocarbons are such thai the  plume is either contained or spreading at a
slow  rate, these  natural  processes can  be  considered in the design of a site  remediation
program.  Even at sites  where plumes are spreading at a fast rate, knowledge of the natural
processes that a re attenuating contaminants provides information that can be used to accelerate
biodegradation processes.
Note

The data and interpretation in this report were previously published in Baedeckerand Cozzarelli
(1 9) and Baedecker et al. (5).
References
 1.     Baedecker,  M.J., and W.  Back.  1
       reactions at a landfill. Ground Water

 2.     Nicholson, R.V., J.A. Cherry,  and E.J
       ground water at a landfill: A case study
979
  .  Hydrogeological  processes  and chemical
17:429-437.
   Reardon.  1983. Migration of contaminants in
   6. Hydrogeochemistry. J. Hydrol 63:1 31 -1 76.
        Lesage, S., R.E. Jackson, M.W. Priddle, and P.G. Riemann. 1 990. Occurrence and fate
        of organic solvent residues in anoxic ground water at the Gloucester landfill, Canada.
        Environ. Sci. Technol. 24:559-566.
                                                                                   43

-------
4.     Godsy, E.M., D.F. Goerlitz, and D. Grbic-Galic. 1 992. Methanogenic biodegradation
       of creosote contaminants in natural and simulated ground-water ecosystems. Ground
       Water 30:232-242.

5.     Baedecker, M.J., I.M. Cozzarelli, R.P. Eganhouse, D.I. Siegel, and P.C. Bennett. 1993.
       Crude oil in a shallow sand and gravel aquifer, III. Biogeochemical reactions and mass
       balance modeling in anoxic ground water. App. Geochem. 8:569-586.

6.     Barker, J.F., J.S.Tessman, P.E. Plotz,andM. Reinhard. 1 986. The organic geochemistry
       of a sanitary landfill leachate plume. Contam. Hydrol. 1:171-1 89.

7.     Major, D.W., C.I. Mayfield, and J.F. Barker. 1 988. Biotransformation of benzene by
       denitrification in aquifer sand. Ground Water 26:8-14.

8.     Cozzarelli,  I.M.,  R.P.  Eganhouse,  and  M.J. Baedecker.  1990. Transformation of
       monoaromatic hydrocarbons to organic acids  in anoxic ground-water environment.
       Environ. Geol. Water Sci. 16:135-141.

9.     Eganhouse, R.P., M.J. Baedecker, I.M. Cozzarelli, G.R. Aiken,  K.A. Thorn, and T.F.
       Dorsey. 1 993. Crude oil inashallowsand and gravel aquifer, II. Organic geochemistry.
       Appl.  Geochem.  8:551-567.

10.    Davis, J.W., N.J. Klier, and  C.L Carpenter.  1994. Natural biological attenuation of
       benzene  in  ground  water  beneath  a   manufacturing  facility.   Ground   Water
       32(2):215-226.

11.    Kampfer, P., M. Steiof, and W. Dott. 1 991. Microbiological characterization of a fuel
       oil-contaminated site including numerical identification of heterotrophic water and soil
       bacteria. Microb. Ecol. 21:227-251.

12.    Kuhn, E.P., J. Zeyer, P. Eicher, and R.P. Schwarzenback. 1 988. Anaerobic degradation
       of alkylated benzene in denitrifying laboratory aquifer columns. Appl. Environ. Microb.
       54:490-496.

13.    Hutchins, S.R., G.W. Sewell, D.A. Kovacs, and  G.A. Smith.  1 991. Biodegradation of
       aromatic hydrocarbons by aquifer microorganisms underdenitrifying conditions. Environ.
       Sci. Technol. 25:68-76.

14.    Wilson,  B.H.,  G.B.  Smith,  and  J.F. Rees. 1986.  Biotransformations  of selected
       alkylbenzenes and halogenated  aliphatic  hydrocarbons  in  methanogenic  aquifer
       material: A microcosm study. Environ. Sci. Technol. 20:997-1,002.

15.    Grbic-Galic, D.,  and T.M. Vogel. 1987. Transformation of toluene and  benzene by
       mixed methanogenic cultures. Appl. Environ. Microbiol. 53:254-260.

16.    Lovley, D.R., M.J. Baedecker,  DJ. Lonergan, I.M. Cozzarelli, E.P.  Phillips, and  D.I.
       Siegel. 1 989. Oxidation of aromatic contaminants coupled to microbial iron reduction.
       Nature 339:297-299.

-------
17.    Haag,  P.M., M. Reinhard, and P.L
18.
20.
21
22.
23.
25.
26.
                                  McCarty.  1991. Degradation of toluene and
       p-xylene in anaerobic microcosms: Evidence for sulfate as a terminal electron acceptor.
       Environ. Toxicol. Chem. 10:1,379-1,390.
Bennett, P.C., D.I. Siegel, M.J. Baedec
sand and gravel aquifer, I. Hydrology
8:529-549.
                                  cer, and M.F. Hult. 1 993. Crude oil in a shallow
                                   and inorganic geochemistry. Appl. Geochem.
19.    Baedecker, M.J., and I.M. Cozzarelli.
       of aqueous constituents  in ground
       substances and the hydrologic scieno
                                   994. Biogeochemical processes and migration
                                  voter contaminated with crude oil. In:  Toxic
                                  s. American Institute of Hydrology, pp. 69-79.
Cozzarelli,  I.M., M.J.  Baedecker, R.P. Eganhouse, and  D.F. Goerlitz. 1994. The
geochemical  evolution of  low-molecular-weight organic acids derived  from the
       degradation of petroleum contaminan
       58(2):863-877.
                                         3 O
Eganhouse, R.P., M.J. Baedecker, and
in an aquifer contaminated by crud
Minnesota,  research  site.  Presentee
Bioremediation of Ground Water, Denver,
                               . Aiken
Cozzarelli, I.M., M.J. Baedecker, G.
heterogeneities  in  a  crude-oil-con
Morganwalp, D.W., and D.A. Aronsot
the U.S. Geological Survey Toxic Substances
(September 20-24, 1 993). Water Res
Hult, M.F., and R.R. Grabbe. 1988.
the unsaturated zone. In: Ragone, S.E.
water contamination study. Proceeding
Survey,  Cape Cod,  MA (October
C21-C26.
                                        21
24.    Essaid, H.I., M.J.  Baedecker, and  I .A/
                                  s in ground water. Geochim. Cosmochim. Acta
                                   M. Cozzarelli. 1 994. Biogeochemical processes
                                        An  overview of studies at the Bemidji,
                                    at  the  U.S.  EPA  Symposium on Intrinsic
                                       CO (August 30 to September 1).
  , and C. Phinney. 1 994. Small-scale chemical
aminated  aquifer,  Bemidji,  Minnesota.  In:
i, eds. Proceedings of the Technical Meeting of
    Hydrology Program, Colorado Springs, CO
 Invest. Rep. 94-4014. In press.
                                 pistribution of gases and hydrocarbon vapors in
                                   ed. U.S. Geological Survey toxic waste/ground
                                  ; of the Technical Meeting of the U.S. Geological
                                   -25, 1985).  Open-File  Report  86-481. pp.
                                  . Cozzarelli. 1994. Use of simulation to study
       field-scale solute transport and biodegiadation at the Bemidji, Minnesota, crude-oil spill
       site.  In: Morganwalp, D.W., and D.A. Aronson, eds. U.S. Geological  Survey toxic
       substances hydrology program. Proceedings of the Technical  Meeting  of the  U.S.
       Geological Survey Water Resources, Colorado Springs, CO (September 20-24,1 993).
       Water Res. Invest. Rep. 94-4014. In press.
Barker, J.F., J.T.Wilson. 1992. Natural
under anaerobic conditions. Proceed!
Dallas, TX. pp. 57-58.
Grbic-Gdlic, D. 1 990. Anaerobic
and alicyclic compounds in soil, subsu
and G. Stotzky, eds. Soil Biochemistry
117-189.
                                  biological attenuation of aromatic hydrocarbons
                                  ngs of the Subsurface Restoration Conference,
                              micrbbial transformation of nonoxygenated aromatic
                                  face, and freshwater sediments. In: Bollag, J.M.,
                                   Vol. 6. New York, NY: Marcel Dekker, Inc. pp.
                                                                                  45

-------
27.    Hadley, P.W., and R. Armstrong. 1991. Where's the benzene? Examining California
       ground-water quality surveys: Ground Water 29(1):35-40.

-------
In Situ Bioremediation at the Seventh Avenue Site  in  Denver-.
Remediation of Soils and Ground  Water
Christopher Nelson
Groundwater Technology, Inc., Englewood,
                                         00
                                      sthod
In situ bioremediation is a cost-effective method for the remediation of soils and ground waters
contaminated with petroleum hydrocarbons. T iis paper presents a case study of the successful
application of in situ bioremediation at a pub
                                          tie
The site  was used as  a truck maintenance
nonvolatile petroleum hydrocarbons from used motoroil
fluids  were released to a used oil sump at
claystone and sandstone bedrock, covered with
gravels,  as well as silty sands and  clays.
interbedded alluvial sands and gravels.
                                           ic utility site in Denver, Colorado.
                                         The
acility for almost 30 years. During that time,
       il, diesel, gasoline, and other automotive
   site. The  local geology of the site  includes
recentdeposits of interbedded alluvial sands and
   principal  aquifer at the  site lies within the
Soil and water sampling confirmed the presence of petroleum hydrocarbons in both the vadose
and  saturated zones.  Laboratory studies  showed  that the chemical, microbiological, and
hydrogeological characteristics of the site werj conducive to bioremediation.
Site Assessment/Treatability Study

An  extensive site assessment and treatability study was conducted at the site to  determine
physical, chemical, microbiological,  and hydrogeological characteristics controlling the bio-
degradation of contaminants and the mass tra
indicated that the primary contaminant at the
                                           isport of nutrients and oxygen. Nine monitoring
wells were installed at depths of about 7.6 m, using a 3-m screen interval. The site assessment
                                           site was waste oil  located in the saturated and
unsaturated sediments beneath the former used oil  sump. Samples showed high levels of
benzene, toluene, ethylbenzene, and the xylenes (BTEX); total petroleum hydrocarbons (TPHs);
and  total  organic carbon  (TOC)  in ground water, with localized but detectable levels of
chlorinated  organics.  A  relatively large  population  of bacteria was  found  within  the
contamination zone; however, its growth appeared to be restricted by nutrient and oxygen
conditions.
Feasibility  studies  were  performed  on  soi
                                             and  sediment  samples  to  determine  the
biodegradability of contaminants under variou; nutrient loads and aerobic conditions. Aerobic
testing was performed to simulate optimal conditions for the bioremediation of hydrocarbons.
Column studies also were conducted to determine how nutrient and hydrogen peroxide loading
would affect the hydraulic conductivity of sediments in the subsurface above the ground-water
table.

The results showed that the loading of nutrients and hydrogen peroxide would be critical to the
success of in situ bioremediation at this site ana that the loading should be minimized in the silty
sand  zone because of high reactivity. A nutrient adsorption test indicated that ammonia and
phosphate loading was feasible in the soil at
                                           he site.
Symposium on Intrinsic Bioremediation of Ground Water
                                                                                    47

-------
 System  Design and Installation

 After reviewing several remedial operations, in situ bioremediation was selected. The conceptual
 design included stimulating indigenous bacterial populations through the introduction of oxygen
 and inorganic nutrients. The primary functions of the remediation system included ground-water
 recovery, treatment, and  reinjection; vapor  extraction and discharge;  stimulation of in  situ
 bioremediation by subsurface inorganic nutrient and oxygen additions;  and phase-separated
 hydrocarbon recovery.

 Laboratory tests showed that hydrogen peroxide and nutrient loading worked best in sediments
 from the coarse sand interval. Nutrient injections and the addition of atmospheric oxygen from
 the vapor extraction system stimulated bioremediation in the unsaturated zone and enhanced
 the desorption of adsorbed hydrocarbons for recovery in the monitoring wells.
 Operation, Monitoring, and Results

 Once the bioremediation system was installed, it was inspected weekly to adjust and maintain
 the water-table depression pump, the hydrogen peroxide and nutrient injection equipment, and
 the soil vapor extraction system. Crews conducted field analyte tests, sampled ground water, and
 gauged  monitoring wells. The bioremediation  system began operating in July 1989 and
 continued through March 1992.

 Approximately 36,000 Ib of hydrocarbons have been removed. Nearly 94 percent of  the
 contaminant mass was  degraded  biologically,  as  evidenced  by the low concentrations of
 dissolved oxygen and the  relatively high  concentrations of  background heterotrophs and
 hydrocarbon-utilizing  bacteria. Approximately 9  million gal of ground water were recovered,
 amended with nutrients,  and reinfiltrated.  The site  is currently undergoing  closure with  the
 Colorado Department of Health.
Reference

For more information about the Denver site and supporting data, refer to:

Nelson, C.H., R.J. Hicks, and S.D. Andrews. 1 994. An integrated system approach for in situ
bioremediation of petroleum hydrocarbon contaminated soil and ground water. In: Flathman,
P.E., D.E. Jerger, and J.H. Exner, eds. Bioremediation: Field experience.
48

-------
The Role of intrinsic Bioremediotion in Closure of Sites After Cleanup Through
In Situ Bioremediation: The Regulator's Perspective
Mark E. Walker and Lisa C. Weers
Colorado Department of Health, Denver, CO
Because staff are assigned such a large numbe r of sites, reviewing all of the hard data for each
and every site is not conceptually possible. Therefore, we do not want to receive all of the
technical information and correspondence generated at every site;  instead,  we prefer the
technical interpretation. Ourapproach to site evaluation is to reviewthis interpretation, reserving
the right to request any and all hard data generated during the course of the investigation.

Although the technical staff at the Colorado Department of Health (CDH) review an abridged
version of the total information generated at a particular site, our version of the file on the Public
Service Company of Colorado (PSCo)  occup
abridged version of this site presented a daunting task.
I needed to get up to speed quickly on the file
to verify if an approved corrective action plan
had been placed  on this approval, if any. In a
been documented, we check to see if the poin'
   ed 6 feet of shelf space.  Even a review of the
   o consider this site for closure. My first task was
   was in place and to determine what conditions
   iy case where ground-water contamination has
   of compliance (POC) wells have been affected
in excess of state maximum contaminant levels (MCL). The second task was to conduct an
evaluation of the existing monitoring program to determine if the program was  adequate for
detecting contamination emanating from the
   •elease. Based on the calculated ground-water
velocity, was the monitoring period long enoug h to detect contamination from the source? With
respect to monitoring well construction and location, were these wells placed downgradient?
(Has direction changed with the seasons?)
Afterthe monitoring plan was completed and th
the next topic for examination prior to closure
and its potential for adversely affecting the environment,
identified receptors in the immediate vicinity an|d
by the remaining contamination.
   3 closure borings had been drilled and analyzed,
   was the extent of the remaining contamination
           This information is considered with the
    the potential for these receptors to be affected
After closure is  complete,  one might ask, "Well, where  does this site stand now, from a
regulatory point of view?" The following items should be considered: 1) after closure is complete,
CDH does not release the owner/operator
remaining  in the subsurface;  2)  the  closure
submitted;  3) similar to the owner's liability
   rom any  liability regarding the  contamination
    decision is based solely on the information
   issue, the CDH wants to be  informed  of any
developments that could increase the potent al for this site to affect human health and the
environment adversely, at which point we assess the need for additional work/remediation.
In conclusion, CDH  is willing to consider
proposing such technologies should be prepared
implemented in a manner that does not adve 'sely
Proposal of an innovative technology should b
innlovative remedial technologies. Those who are
     to demonstrate that these technologies can be
       affect human health and the environment.
   3 approached from the viewpoint of a regulator:
Symposium on Intrinsic Bioremediation of Ground Water
                                            49

-------
consider receptors and aspects of the technology that might be detrimental to the environment,
and address these concerns in a responsible, forthright manner.
50

-------
The Importance of Knowledge About Intr
Closure:  The Client's Perspective
   nsic  Bioremediation  for  Cost-Effective Site
Harry E. Moseley
Public Service Company of Colorado, Denvei
  , CO
Public  Service Company of Colorado (PSCol is Colorado's  largest electric and gas utility.
During an  underground storage tank removal and replacement project in 1 987 at PSCo's
Seventh Avenue Service Center, a facility used for automotive maintenance, an oil sump was
discovered. Because the sump did not have a concrete bottom or other type of platform, the
soil beneath the sump was saturated with used oil.

After an extensive study to determine the nature and extent of the contamination beneath the
facility, PSCo officials decided that/n situ bibremediation of the site would be more cost effective
than removing the building that housed the center. Construction of a new facility would cost
at least $1  million, excluding demolition of the existing facility and removal or remediation of
the contaminated soil and ground water. After discussions with the Colorado  Department of
Health and 1 8 months of ground-water monitoring, the site received approval for final closure
in March 1 994, 7 years after discovery of the contamination.
PSCo, a business without extensive expertise in
with a fairly direct goal: to remediate the si
Intrinsic bioremediation was found to be a key
was known about intrinsic bioremediation, anc
additional knowledge of these fields might
For example, in investigating whether biorem
enhancing the  activities of the bacteria that
   nvironmental restoration, undertook the project
  e as efficiently and cost-effectively as possible.
   element of the project's success.  In 1 988, little
   some opportunities may have been lost; indeed,
have reduced costs.
  sdiation was feasible, emphasis was  placed on
  vould perform the remediation, rather than on
determining how long the enhancement COL Id sustain  bacterial growth, which  might have
revealed a more effective strategy for delivering nutrients.  Once the nutrients penetrated across
the site, batch feeding of the nutrients may halve been more effective for the sustained growth
of the bacteria than the continuous recirculation of the enriched ground water. If this fact had
been considered before  the start of the project, operation and maintenance costs incurred
during the project's life cycle could  have been reduced.

The risk assessment and ground-water data suggested that the chance of any contaminants
appearing in any ground water being consumed or used by humans was minimal.  The natural
ground-water flow across the site was very slow, and the site might have remediated itself with
the help of a small amount of oxygen and other nutrients.
To  evaluate the  progress of the project and to adhere to Underground Injection Control
regulations, many tests were required that monijtored the concentration of the injected chemicals,
or "nutrients"; these tests necessitated costly site visits and report submissions.  The chemicals
were not injected  at harmful levels, however, nor was the risk of human exposure significant due
to the location of the injections and the velocity of the ground water.  Three years into the
project, the reporting requirements were reduc
bioremediation process would have reduced
  ed to quarterly, but a better understanding of the
  hese requirements earlier.
Symposium on Intrinsic Bioremediation of Ground Water
                                            51

-------
The bioremediation process came in on budget and successfully removed all threat to human
health and the environment (see Table 1).
Table 1. Resources Used at the Seventh Avenue Bioremediation Project
1990
1991
1992
1993
Labor hours
Activities
Testing cost
Materials cost
Labor hours
Activities
Testing cost
Materials cost
Labor hours
Activities
Testing cost
Labor hours
Activities
Testing (quarterly)
2,024 hours total
1 68 hours per month average
O&M costs
Quarterly and monthly reporting
$12,778
$16,473
1 ,746 hours total
1 45 hours per month average
Onsite inspections (once or twice a week)
Risk assessment formulation
O&M costs
Quarterly and monthly reporting
$13,451
$12,000
551 hours total
46 hours per month average
Quarterly reports
Monthly site visits
Closure report submittal
$11,539
94.25 hours total
7.85 hours per month average
Closure monitoring
$8,420
Information from past projects, sound scientific judgment, and risk analyses will provide an
understanding of intrinsic bioremediation that could increase the number of cleanups, speed up
the closure process, and  reduce testing and  reporting requirements, thereby substantially
reducing costs to industry.
52

-------
The Role of Intrinsic Bioremediation  in  C
In Situ Bioremediation: The  Role of  Mat
                                          osure of Sites After Cleanup Through
                                          lematical Models
Tissa H. Illangasekare and David C. Szlag
Department of Civil, Environmental, and Arch
University of Colorado, Boulder, CO

John T. Wilson
Robert S. Kerr Environmental Research Laborctory,
U.S. Environmental Protection Agency, Ada, OK
                                           ectural Engineering,
Abstract

This pa per discusses the important processes ir
tools  for  the  design and  evaluation  of
contaminated  with  nonaqueous phase orgc
application  of a  mathematical model that
methodology for retrospective evaluation of b
Introduction
                                           developing mathematical models to be used as
                                        grsund-water remediation schemes  in aquifers
                                          nic chemicals.  The paper also presents  the
                                           considers these  processes  in  developing a
                                          oremediated sites.
Mathematical  models of water flow and chemical transport have been extensively used in
ground-water quantity and quality managem
involve bioremediation. The  basic  processes
                                          snt. Some of the applications of these models
                                           of flow of water  and transport of dissolved
substances are fairly well understood, and numerical models for the solution of the governing
equations have been developed. Field applications of models as prediction arid design tools,
however, have not been very successful for many reasons, including the complexities associated
with natural heterogeneities and the inadequacy of available field techniques for physical and
chemical  characterization; this is especially
involving  chemicals  in the form of separate
                                          true in ground-water contamination situations
                                           phase organics. The  models that have been
 developed to simulate the transport and entrapment behavior of nonaqueous phase chemicals
 and waste products have not been adequatelj validated due to the scarcity of laboratory and
 field data. These models sometimes fail to simulate flow and entrapment behavior under
 heterogeneous soil conditions that are commonly encountered in the field. Accurate calibration
 and prediction become difficult due to the limitations of field and laboratory techniques that are
 used to obtain model parameters. Some of the scaling issues related to multiphase flow model
 parameters  are not very well understood. The assumptions that are  made in modeling mass
 transfer from  entrapped chemicals to the aqueous phase become questionable under some
 conditions of  ganglia formation and macroscale entrapment.

 The movement of nonaqueous phase liquids  (NAPLs) in the subsurface is a complicated process
 leading to large amounts of NAPLs becoming trapped in the soil. Most of the common organic
 wastes in the  form  of NAPLs are only  sparingly soluble  in water and thus act as long-term
 sources of ground-water contamination. In a case study presented in this paper, we will show
 that after bioremediation,  pockets of NAPLs  remained in the soil. These  entrapped fluids have
 Symposium on Intrinsic Bioremediation of Ground Water
                                                                                   53

-------
 the  potential  to contribute to ground-water  pollution  after active  remediation  has been
 discontinued.

 In our research, we have identified two modes of entrapment:  microscale and  macroscale.
 Microscale entrapment, which occurs at residual levels, is primarily governed by fluid properties
 and pore characteristics  (1). Macroscale entrapment is defined  as entrapment at saturations
 higher than residual (or irreducible) due to heterogeneities in the soil  (2).

 Many models employed  in remediation design make use of the  local equilibrium assumption
 (LEA); in other words, any water exiting a zone of entrapped NAPL will be completely saturated
 with the contaminant, regardless of the system parameters. The LEA is conservative in predicting
 the maximum  concentration observed in ground water but may lead to gross underprediction
 of the contaminant source lifetime.  It is our hypothesis that entrapment itself will change the
 system parameters controlling mass transfer into the flowing water, and the LEA assumption is
 not a physically realistic way to quantify mass transfer.

 Laboratory experiments were conducted to obtain a fundamental understanding of the processes
 that govern the transport and distribution of organic chemicals in soils and to generate data for
 validation  of models that will  be used  as tools for the design of remediation schemes and
 monitoring systems. A detailed investigation of these processes under controlled conditions was
 done in small soil cells, columns, and large flumes.

 In our ongoing research, new models and modeling approaches have been developed. To
 improve these models, we have focused on issues related to entrapment, mobilization, and mass
 transfer associated with organic waste chemicals in heterogeneous aquifers. The effectiveness
 of models as tools to design and evaluate treatment and remediation technologies depends on
 their ability to accurately  represent the above processes. A case study conducted in Colorado
 identifies some of these basic processes that are of importance in remediating and monitoring
 sites contaminated with organic fluids, and demonstrates the use of a  mathematical model.
Study Objectives

Conventional methods for determining the extent of cleanup at a bioremediation site can often
be misleading. Monitoring wells may show very low or zero levels of contaminants after active
bioremediation, but levels may increase overtime. In most cases, regulatory authorities require
a direct measure of the residual NAPLs after bioremediation in addition to monitoring well data.
Often  the  relative composition  of the  oily phase is assumed  to remain  constant during
bioremediation. This is a conservative assumption and generally leads to target levels of total
petroleum hydrocarbon (TPH) concentrations on the order of 10 to 1 00 mg/kg  aquifer material.
Many  bioremediation  schemes, however,  may  preferentially degrade the compounds of
regulatory concern, leaving relatively high TPH levels in the soil that pose a minimal risk. This
modeling study focuses on developing a methodology  to evaluate the possible risk, if any,
associated with benzene, toluene, ethylbenzene, and xylene (BTEX) sources left in soils after the
implementation of a bioremediation scheme. This developed methodology will assist in providing
answers to the following questions: 1) Will BTEX reappear in ground water? 2) How long will
it take the plume to reappear? 3)  What concentration level may be expected? The results of the
case study will also assist in providing a technical basis for implementing monitoring schedules,
locating compliance wells, and constructing rational criteria for site closure.
54

-------
Problem Description

A temporary holding tank under a garage in an industrial area in Denver, Colorado, leaked
used crank case oil, diesel fuel, gasoline, and other material into a shallow water table aquifer.

Remediation involved removal of separate oily phases, in situ  bioremediation with hydrogen
peroxide and mineral nutrients, and bioventing. An estimated 2,147 Ib of hydrocarbons have
been removed by pumping hydrocarbon emulsion from monitoring wells or by volatization
through soil aeration. In the treatment system,
recovery well was treated and amended with
 he ground water pumped from a downgradient
 organic nutrients  and hydrogen peroxide. The
solution was then injected into the subsurface system upgradient of the contaminated site. The
system operated from October  1989 to March  1992. Table 1 compares the  reduction in
concentration  of benzene and  total  BTEX compounds  in  ground water achieved by this
remediation scheme (3, 4).
Table 1.  Reduction in Concentration (ug/L) of Hydrocarbon Contaminants in Ground Water
         Achieved by In Situ Bioremediation
Well
MW-1
MW-8
MW-2A
MW-3
RW-1
Benzene
Before
220
180
—
11
<1
During
<1
130
n
5
2

AfteV
<1
16


0.4
2
<1


Total BTEX
Before
2,030
1,800
—
1,200
<1
During
164
331
1,200
820
2
After
<6
34
13
46
<1
Water from the monitoring wells and the rec
contaminants by March 1 992. Active remediati
of postremediation monitoring. In June 1 992
determine the  extent  of hydrocarbon  rema
contamination could return once active remedi
downgradient of the release. The cores we
hydrocarbons and for the concentration of in
boreholes with the highest concentrations of h
sampling, the elevation of the water table wa
 rculation well contained low concentrations of
Jon was terminated, and the site entered a period
), core samples were taken from the aquifer to
 ning  and to determine whether a plume of
 ition ceased. The site was cored along a transect
 B extracted  and analyzed for total petroleum
 ividual BTEX compounds. Data from one of the
 drocarbons are given  in Table 2. At the time of
  5,280.5 ft above mean sea level (AMSL).
                                                                                  55

-------
 Table 2.  Vertical Extent of Total BTEX and TPH (mg/kg) at a Borehole
Elevation (Feet AMSL)
281.14-5,280.31
5,280.31-5,279.97
5,279.97-5,279.56
5,279.56-5,279.14
5,279.14-5,278.97
5,278.97-5,278.64
5,278.64-5,278.22 .
5,278.22-5,277.14
TPH
<44
227
860
1,176
294
273
<34
<24
BTEX
<1
5.1
101
206
27
7.4
<1
<1
Benzene
<0.02
<0.2
<0.2
4.3
0.68
0.26
<0.2
<0.2
Color and Texture
Brown sand
Brown sand
Brown sand
Brown sand
Brown sand
Brown sand
Brown sand
Brown/yellow sand
To hydradically characterize the aquifer, selected soil cores that had previously been used to
measure hydrocarbon saturations were reconstructed in a load cell, and hydraulic conductivity
was determined with a constant flux apparatus. The hydraulic conductivity varied two orders of
magnitude across the site, with some highly permeable channels evident.

The results of this field investigation suggest that hydrocarbon in the form of nonaqueous fluids
moved into preferential flow channels created by the local heterogeneities. After direct recovery,
the fluids  remained entrapped at saturations  that may  be  higher than residual. During
remediation, the treating agents did not reach some of the locations where the chemicals were
entrapped.

A modeling study  was conducted at the site  to  make a retrospective  evaluation  of  the
effectiveness of the remediation scheme. The results of this study are reported in Szlag et al. (3).


Model Selection

The following observations  at the site and  in the laboratory indicated the need  for three-
dimensional simulation: 1) visual inspection of aquifer material indicated the presence of coarse
gravel lenses, clayey sands, and sands of varying gradation; 2) light nonaqueous phase liquid
(LNAPL) plumes are inherently three-dimensional, forming thin, pancake-like plumes in  the
capillary fringe and just beneath the water table; 3) LNAPL can become entrapped in coarse
lenses that act as preferential flow channels well  beneath the water table; 4) solute plumes  are
not vertically homogeneous, and biological activity will not be uniformly distributed vertically.
Bypassing due to the lowered hydraulic conductivity of the central part of the LNAPL plume by
nutrients and electron acceptors resulted in high TPH and BTEX levels in some cores. Coupled
with the clear need  for a three-dimensional model are other criteria such as availability, ease
of use, reliability, and cost. We have selected MODFLOW, a three-dimensional ground-water
flow model developed  by the U.S. Geological Survey (USGS), to simulate ground-water flow.
Solute transport is simulated with a three-dimensional random walk called RAND3D.
56

-------
Modeling Approach

The problem domain was modeled as a rectangular area 300 ft long and 200 ft wide. Two
wells outside of the modeled domain were used as reference head locations for general head
boundaries. The NAPL contaminant zone covers approximately 1,600 ft2 of area and a  soil
depth of approximately 1.7 ft. To accurately assess the mass flux from the LNAPL contaminant
source, three layers were chosen in the  model. The upper two layers are 1 ft thick, and the
bottom layer is 1 8 ft thick. The LNAPL organic contaminant is confined to the uppertwo layers.
Data for the period June 8, 1 989, to April  1,
standard deviation of the mean residual. The b<
 990, were used for calibration. The goodness
of fit between the model and measured well cata was characterized by a mean residual and
st fit model was obtained by assigning the pump
test average hydraulic conductivity to the bottom layer, which carried the majority of water.

The main focus of this modeling effort is to cetermine how much contaminant mass will  be
transported from the remaining residual and
whether it will generate a plume of regulatory
concern. A monitoring well screened in onl^ the upper two layers would see the highest
contaminant concentrations. A pumping well screened over the entire aquifer thickness is also
being considered; in this case, however, dilution will play a major role in reducing the maximum
concentrations. Two significant assumptions a B used in the solute transport modeling:  1) the
concentration  of BTEX in the  source  zone remains constant, and 2) water flowing from the
contaminated  cells is in equilibrium with the
MODFLOW, ground-water velocity through ea
BTEX concentration was then used to calculate t
residual NAPL. Using the heads generated by
ch source cell was calculated. The known NAPL
ie equilibrium concentration and, consequently,
the mass flux. Estimated benzene mass fluxes wsre converted into particle inputs for each layer.
The particle tracking  model was  used to simulate solute transport  using the velocity field
generated from MODFLOW.
Results and Discussion

Any simulation of solute transport requires spec
of a NAPL spill, the source function will consi
quickly reached in spill scenarios if ground-wa
ification of the contaminant source. For the case
>t of a continuous mass flux of solute from the
residual NAPL phase to the aqueous phase. A/any researchers have shown that equilibrium is
er velocity is low and the "residence time" of the
water in contact with NAPL is "long enough." From a regulatory standpoint, the assumption of
equilibrium is conservative, as greater mass flu
tes cannot be achieved. The water flux can vary
significantly in the source zone, which often gives misleading indications that the contaminant
transport is rate-limited. This primary problem has been the focus of our work. Preferential flow
paths often develop within the source zone in areas with low BTEX and TPH concentrations,
allowing water flow to bypass the more highly contaminated areas. Laboratory determination of
the hydraulic  conductivity  in the samples  containing high amounts of TPH confirms  this
observation.
A key result of the modeling study is that the so
homogeneous. In general, the solute plume wil
ute plume emanating from a NAPL source is not
 consist of subplumes at different depth intervals
and widely different concentrations, and moving at different velocities. A regulatory question
                                          I  I
posed earlier in this paper is, "How long  should the compliance wells be monitored?" The
answer is when all the subplumes have reached steady-state. The plumes in the middle and
                                                                                   57

-------
 bottom layer have reached or are close to equilibrium by 330 days. The plume in the top layer,
 which has the highest benzene concentration, has not reached equilibrium at 420 days.

 The design of the  compliance wells will have tremendous impact on the actual  sampled
 concentration. If the wells are bailed or pumped so that the well volume is completely mixed,
 significant dilution will occur. The existing  monitoring wells at the site are screened overthe top
 5 ft of the aquifer. The maximum concentration achieved in the well screened overa 5-ft interval
 reaches a steady-state concentration of 26 ppb. If that well is screened overthe entire saturated
 thickness, a concentration of 15 ppb is achieved. Even greater dilution will occur if the well is
 pumped.

 Several operational considerations for risk assessment and compliance well monitoring can be
 made from the modeling study: 1) a  benzene plume will reestablish itself at the site, but it will
 be  three orders  of magnitude  lower than  the federal   maximum contaminant level—new
 standards may be set, however, and the  risk from this plume may be deemed significant; 2)
 local hydraulic conductivity plays a significant role in determining the contaminant mass flux and
 in creating subplumes of different concentration and velocity; 3) compliance well monitoring will
 have to be continued past August 1993 so that solute plumes in all levels will reach steady-
 state; 4) retardation coefficients and effective porosity data would significantly improve the time
 of arrival estimate of the solute plume; 5) compliance well design should be carefully considered
 when  sampling a  three-dimensional plume because well design  can  lead to significant
 contaminant dilution.
Conclusions

A modeling  methodology  for the  retrospective  evaluation  of bioremediated  aquifers
contaminated with organic chemicals was developed. The primary hypothesis on which the
methodology  was based is  that  during  a spill,  NAPL contaminant  becomes entrapped
preferentially in coarse formations in the saturated zone and fine formations in the unsaturated
zone. This hypothesis is supported by laboratory experimental (2) and field data. Flow channels
created by naturally occurring aquifer soil heterogeneities as well as macroscale entrapment of
the NAPL will also produce preferential paths forthe treating agents. The proposed methodology
requires that these local heterogeneities  in the contaminant zone of the spill be  captured.
Standard pump tests, which provide regional values for transmissivity, will not have adequate
resolution to capture these spill-site-scale heterogeneities. Even though hydraulic conductivity
values determined in the laboratory on  disturbed soil samples were used in this study, a more
appropriate characterization method would be well-designed bail tests (orslug tests) that capture
the local layered heterogeneities more accurately. These local hydraulic conductivity values allow
us to obtain the  velocity field in the contaminant zone and to subsequently determine the
contaminant mass flux. Solute breakthrough curves determined by this method can then be used
to conduct risk analysis  and to provide a rational basis for postremediation well monitoring.
References

1.     Szlag, D.,  and T.H. Illangasekare.  1994. Quantification of residual entrapment of
       nonaqueous phase organic fluids in soils. Ground Water. In review.
58

-------
2.
3.
Illangasekare, T.H., D. Szlag, J. Camp sell, J. Ramsey, M. Al-Sherida, and D.D. Reible.
1 991. Effect of heterogeneities and preferential flow on distribution and recovery of oily
wastes  in  aquifers. Proceedings of the Conference  on Hazardous  Waste Research,
Manhattan, KS. Manhattan, KS: Kanscs State University.

Szlag, D.C., T.H. Illangasekare, and  I.T. Wilson.  1993. Use of a three-dimensional
ground-water  model  for  retrospectve   evaluation of  a  bioremediated   aquifer
contaminated with  organic chemicals. Proceedings  of  the Ground-Water Modeling
Conference, Golden, CO (June 10).
4.
Wilson, J. 1 993. Retrospective perf
characterization. In: U.S. EPA.
Research, development, and  field
Washington, DC (May).
   orrnance evaluation on in situ bioremediation: Site
Symposium on bioremediation of hazardous wastes:
     'evaluations  (abstracts). EPA/600/R-93/054.
Acknowledgments
The support of the U.S. Environmental Protection
Research Center at Kansas State University (agreement
We would also like to thank Ms. Lisa Weers
assistance.
                                       Agency through the Hazardous Substance
                                          R-81 5709) is gratefully acknowledged.
                                   of the Colorado Department of Health for her
                                                                                   59

-------
 Intrinsic Bioremediation  of JP-4 Jet Fuel
John T. Wilson, Fredrick M. Pfeffer, James W. Weaver, and Don H. Kampbell
U.S. Environmental Protection Agency, Robert S. Kerr Environmental Research Laboratory,
Ada, OK

Todd H. Wiedemeier
Engineering Science, Denver, CO

Jerry E. Hansen and Ross N. Miller
Air Force Center for Environmental Excellence, Brooks AFB, TX
Introduction

Intrinsic bioremediation is a risk management option that relies on natural biological processes
to contain the spread of contamination from spills. The option is most appropriate when the
concentration of contaminants is reduced to regulatory limits before ground water discharges
to surface water or is collected by a pumped well.

In the past, remedial action plans have proposed the intrinsic remediation option based solely
on the apparent attenuation of contamination in water from monitoring wells that are distant
from the spill. These  plans were  often criticized because it was impossible to  distinguish
attenuation due to contaminant destruction from attenuation due to simple dilution in the aquifer
or in the monitoring well. Convincing regulators that the wells  with low concentrations of
contaminants actually sample the plume of contaminated ground water has been difficult. This
lack of credibility  has led to the "one-more-well" syndrome, with excessive investment in a
monitoring approach that focuses on the compounds of regulatory concern but fails to earn the
confidence of the regulatory community.

During characterization of JP-4 jet fuel spills at Eglin Air Force Base (AFB), Florida, and Hill AFB,
Utah, three approaches were  used to distinguish contaminant attenuation due to  destruction
from attenuation due to dilution or sorption.

To distinguish attenuation  due to biological destruction of the contaminants from attenuation
due to dilution, the attenuation of the compounds of regulatory concern—benzene, toluene,
ethylbenzene, and the xylenes (BTEX)—was compared with the attenuation of other components
of the fuel that were relatively  recalcitrant. Tracers have been used successfully to correct for
dilution and  sorption of hydrocarbons in ground water. Cozzarelli et al. (1)  used 1,2,3,4-
tetramethylbenzeneto normalize the concentrations of other alkylbenzenes  and their anaerobic
degradation products in ground water that had  been contaminated by a spill of light crude oil.
Wilson etal. (2) used 2,3-dimethylpentaneto normalize the concentrations  of BTEX compounds
in ground water contaminated  with gasoline from an underground storage tank. In both cases,
the tracer was a component of the spilled fuel.

To distinguish attenuation due  to biological destruction from attenuation due to sorption, core
samples were analyzed for the total  quantity  of individual BTEX compounds and for total
petroleum hydrocarbons. Partitioning theory was used to predict the concentration of individual
60
Symposium on Intrinsic Bioremediation of Ground Water

-------
hydrocarbons in ground water in contact with
were compared with concentrations in water
water in the plume was in sorptive equilibriurr
To prove that the attenuation was due to bio
 fhe core material. The predicted concentrations
from monitoring wells to determine if the ground
  with the spilled fuel.
 ogical activity, the geochemistry of the ground
 3olism  of   petroleum   hydrocarbons   has
water   was   examined.   Microbial   meta
predictablegeochemical consequences. The hydrocarbons can  be respired,  resulting in the
consumption  of oxygen, nitrate, sulfate, or irsn II  minerals in the aquifer matrix and in the
production of water, dinitrogen, sulfide, or iron
microbial  respiration as electron acceptors. A
production of methane. Simple stoichiometry
  I. Microbiologist often referto the substrates for
  kylbenzenes can be fermented, resulting in the
 can be used to predict the quantity of electron
acceptors consumed or the quantity of methane produced during biotransformation of a given
concentration of petroleum-derived hydrocarbons.
Intrinsic Bioremediation of Ground Watei
POL is located over sands and silty peats chan
of the study, a plume of contaminated grounc
discharged to a small creek approximately 300
water table is approximately 8.4 ft above  sea
table in the creek is 1.4 ft above mean sea leve
  at Eglin AFB, Florida
Leaking distribution pipes from an underground storage tank released JP-4 jet fuel to the water
table aquifer under the petroleum, oil, and lubricants storage depot (POL) at Eglin AFB. The
  cteristic of a barrier island complex. At the time
  water moved away from the residual JP-4 and
  ftdowngradient (Figure 1). The elevation of the
  level in the area with residual JP-4. The water
  . Hydraulic conductivity determined by pumping
tests in monitoring wells varied from 48 ft to 1 02 ft per day. Based on these data, and assuming
an effective porosity of 30 percent, the residerce time along the flow path from the spill to the
creek is on the order of 1 0 wk. Water samples
screen. Seasonally, the temperature of ground
the pH varies from 5.6 to 6.7. Samples prod
1 993, when ground-water temperatures varied from 24°C to 28°C.
 were acquired with a geoprobe, using an 1 8-in.
 water at the site varies from 1 9°C to 28°C, and
 jcing the data in Table 1  were taken in August
Correcting for Dilution  at Eglin  AFB

Table  1  presents the  changes  in  conceri
trimethylbenzenes (TMB) along the flow  path
 rations  of BTEX compounds  and  the three
  from the spill to the creek (Figures 1  and 2).
Sample 80H-3 is from a location just outside the JP-4 spill and appears to be in chemical
equilibrium with the weathered residual fuel. Samples 83H-1 and 83Z-2 are from locations
approximately 1 50 ft and 300 ft downgradienlj from the spill. Samples 83U-2 and 83U-3 were
taken 0.5 ft and 4.1  ft below the sediments of the creek receiving discharge from the plume.
Sample 83U-1 is water from the creek at the sediment boundary, taken when the tide was going
out and the plume was actively discharging to the creek.
When trimethylbenzene concentrations  in samples 80H-3 and 83U-3 were compared, the
reduction in the .concentration  of these compounds was found to be remarkably uniform.
Concentrations under the creek were 36, 27, and 46 percent of the concentrations near the
spill, while the concentrations of toluene, ethylbenzene, p-xylene, m-xylene, and o-xylene were
0.02, 0.08, 0.44, 0.20, and less than 0.02 percent of the initial concentrations.
                                                                                   61

-------
                  ®PL-2
  .
 /


A - A'
                         SOIL BORING

                         2" MONITORING WELL

                         LINE OF EQUAL
                         WATER TABLE ELEVATION
                         (FEET ABOVE MEAN SEA LEVEL)

                         UNE  °F HYDROGEOLOGIC SECTION
SCALE
         rm
       , T
          FIGURE 1

  Water Table Elevation Map
Showing Sampling Locatlona and
  Flow Path of Ground  Water
   from JP-4 Spill to Point
  of Surface Water Discharge.

     EGLIN AIR FORCE  BASE
Figure 1.  Flow path of ground water from a JP-4 spill at Eglin AFB to the point of discharge to
            surface water: plan view.
                            A
                          North
                         -20
                            LEGEND

                      X  Sample Location
                     -sr_ Water Table
                                                            A'
                                                           South
                                                                              r '5
                                                                               . 0    E
                                                                               -20
                                                                             FIGURE 2
                                                          Hydrogeologlc
                                                          Section A-A'
                                                                        EGLIN AIR FORCE BASE
Figure 2.  Flow path of ground water from a JP-4 spill at Eglin AFB to the point of discharge to
            surface water: cross section.
62

-------
Table 1.  Bioattenuation in Methanogenic
         Along a Flow Path From a JP-4 Spill
Ground Water: Changes in Ground-Water Chemistry
     to the Point of Discharge to Surface Water

Compound

Benzene
Toluene
Ethyl benzene
p-Xylene
m-Xylene
o-Xylene
1 ,3,5-TMB
1 ,2,4-TMB
1 ,2,3-TMB

BTEX and TMB
Oxygen
Nitrate +
Nitrite - N
Sulfate
Methane
Iron II
Location Ale
80H-3
83Z-2

TOO
5,150
1,700
3,120
6,750
5,480
327
1,090
406
153
18.3
227
594
1,270
<1
114
420
182

24.1
0.4
0.14
1.6
3.7
2.3
2.97
0.2
0.12
0.62
16.8
7.8
>ng Flow Path to Surface Water

83U-3
83U-2
83U-1
fcg/L)









198
1.1
1.4
13.8
13.5
<1
70.9
172
115
6.9
<1
1.4
25.1
39.8
<1
119
299
187
<1
<1
<1
<1
<1
<1
2.3
<1
<1
(mg/L)






0.59
0.3
<0.05
5.6
12.5
3.3
0.68
0.6
<0.05
1.76
14.2
2.8
0.002
3.8
<0.05
<0.5
0.7
<0.5
Apparently, concentrations were reduced from one-half to one-fourth of the initial concentration
due to dilution, with further reductions due to biological activity. Benzene was not degraded in
the anaerobic portion of the flow path.

As the plume moved  up into the sediments of the creek, the concentration of benzene was
reduced more than 20-fold (compare 83U-2 with 83U-3 in Table 1).  If we assume that the
trimethylbenzenes are recalcitrant, the dilution bf 1,2,4-trimethylbenzene can be used to correct
forthe dilution of benzene and determine the true removal due to biodegradaWon. The corrected
concentration of a biologically transformed compound in a  downgradient well would be the
measured concentration in the downgradient well, multiplied by the measured concentration of
 1,2,4-trimethylbenzene in the upgradient well
 1,2,4-trimethylbenzene in the downgradient v
      and divided by the measured concentration of
                                                                                   63

-------
The concentration of benzene measured in sample 83U-2 was 6.9/ig/L Compared with sample
83Z-2, the concentration corrected for dilution would be 9.7/ig/L. Compared with 1 53 /tg/L
in 83Z-2, there was a 1 5-fold attenuation in benzene concentration between 83Z-2 and 83U-2.
The benzene attenuation at 83U-2 compared with 83U-3 would be 50-fold.

The concentration of oxygen in the sample 83U-2, taken 0.5 ft below the sediment surface, was
higher than the concentration in sample 83U-3, taken 4.5 ft below the sediment surface. Tidal
action may reverse the hydraulic gradient in the area proximate to the creek. This would produce
a  reciprocating flow  of oxygenated creek water  into the sediments and  mix  oxygen into the
contaminated  ground water. Benzene may  have been degraded  aerobically. In  any case,
benzene and  the other BTEX  compounds did  not discharge  to  the stream at detectable
concentrations (see sample 83U-1).
Kinetics of Bioattenuation in  Ground Water at  Eglin AFB

First-order rate constants were calculated by correcting the downgradient concentration for
dilution. The rate constants were calculated as:

Rate =  Infcorrected cone, downgradient/conc. upgradient)
               residence time

Based on this relationship, rates were calculated for flow path segments from samples 80H-3
to 83Z-2 and 83Z-2 to 83U-3. The residence time in each segment was assumed to be 5 wk.

The rates of anaerobic bioattenuation (Table 2)  were some of the fastest that have ever been
encountered by staff of the Robert S. Kerr Environmental Research Laboratory, probably due to
the high water temperatures. There seemed to be preferential removal of toluene and o-xylehe
in the segment close to the spill. Rates of removal of  ethylbenzene and m+p-xylene increased
after toluene and o-xylene were  depleted.
Table 2. First-Order Rate Constants for Bioattenuation of BTEX Compounds in a Plume of
         Ground Water Contaminated by JP-4
Compound

Benzene
Toluene
Ethylbenzene
p-Xvlene
m-Xylene
o-Xylene
80H-3 to 83Z-2
83Z-2 to 83U-3
(per week)
None
-0.94
-0.21
-0.14
-0.14
-1.5
None
-0.38
-0.38
-0.57
-0.73
Cannot calculate
64

-------
Stoichiometry of Bioattenuation at Eglin

There was very little oxygen, nitrate, orsulfate
in the  plume of contaminated ground water
                                          AFB

                                          available for respiration of the BTEX compounds
                                           (see Table 1). There was an increase in the
concentration of methane and iron and a corre; ponding decrease in aromatic fuel hydrocarbons
along the flow path, however. Assuming the following  stoichiometry for methanogenesis from
BTEX compounds,

2 CH +  3/2 H2O  ->  5/4 CH4  + 3/4 C(D2,

approximately 1.0  mg of methane is produced for each 1.3 mg of BTEX destroyed. If 80H-3
and 83Z-2 are compared, the concentration olj methane at 83Z-2, corrected for dilution, would
be 43.5 mg/L. The increase in methane from 80H-3 to 83Z-2 would be 39.8 mg/L. Corrected
for dilution,  the concentration of BTEX plus TMB in 83Z-2  is 7.69  mg/L, a decrease of 1 6.4
mg/L compared with 80H-3. This decrease in aromatic  petroleum-derived hydrocarbons would
be expected  to produce 12.6 mg/L of methane. Some  of the methane sampled at 83Z-2 may
have come from the natural degradation of the peat. In  any case, the accumulation of methane
is sufficient to rationalize the destruction of the aromatic fuel hydrocarbons along the flow path.
Intrinsic Bioremediation of Ground Wate
                                           at Hill AFB,  Utah
Hill AFB is situated on a bird's foot delta formed  by the Weber River  in Pleistocene Lake
Bonneville. Leaking distribution pipes from an underground storage tank released JP-4 jet fuel
to the water table aquifer under the POL. An area with oily phase JP-4 extends approximately
1,000 ft downgradient of the spill  (Figure 3). The oily phase hydrocarbons and the plume of
contaminated ground water are confined to the channel sands of the delta deposits (Figure 4).
The elevation  of the water table drops 56 ft across the length  of the spill. Slug testing of
monitoring wells indicates a hydraulic conductivity near  8.5  ft per day, corresponding to an
interstitial seepage velocity of  1.6 ft per day.
Continuous cores were taken near the source
(82D), at the lower edge of the spill (82C), a
                                          of the spill (821 in Table 3), near the midpoint
                                          nd just beyond the edge of the spill  (82B). The
continuous cores started in clean material above the spill and extended through the spill into
clean material underneath the interval conta
subdivided into core samples representing 0.3
appeared in the capillary fringe (Table 3). The
to 5.2 ft below the watertable, with TPH conce
                                          ning hydrocarbon. The continuous cores were
                                          vertical ft of the subsurface. Near the spill, JP-4
                                          concentration maximum of 14,800 mg/kg was
located 0.5 ft above the watertable. A second interval that contained JP-4 extended from 3.8
                                          itrations that ranged from 1,290 to 3,830 mg/L.
At the midpoint of the spill (82D), only one core sample contained significant concentrations of
hydrocarbons. That core sample came from 1.0 ft below the watertable. At the lower edge, two
core samples representing 0.6 vertical ft had significant concentrations of hydrocarbons. Again
these .core samples were at or just below the watertable. At the midpoint and lower edge of the
spill, hydrocarbons were not detected in core
reported in Table 3 (detection limit 1 0 mg/kg).
and lower edge of the spill were less than 70

Monitoring wells were installed in the borehol
monitoring wells were installed downgradient
                                          material collected above or below the samples
                                          Concentrations of hydrocarbons in the midpoint
                                          D mg/kg.

                                          3S used to acquire the core samples. Additional
                                          of the lower edge of the spill.
                                                                                   65

-------
                                                                 EXTENT OF JP-4 SPILL
                                                                                   SCALE
                                                                                   300
                                                                                           600
                                           -STORM DRAIN

                              LEGEND
                    «MW-10  MONfTORING WELL LOCATION
                    B	B'  LINE OF HYDROGEOLOGIC SECTION
       Map Showing Sample Locations
       and Flow Path of Ground Water
         from JP-4 Spill to Point of
        Potential Discharge In Storm
               Water Drain
            HILL AIR FORCE BASE
Figure 3.   Flow path of ground water through a JP-4 spill at Hill AFB to the point of discharge
             in a storm drain:  plan view.
                           SW  B

                           J. 4690-
                           -=.4680-
                           J.4670-
                           -J4660
                           Sj 4650-
                           3 4640-
                           V 4630-
                           | 4620
                           = 4610
                           | 4600
                           5 4590
                           z 4580
EPA-82-D
                                   EPA-82-A
                                   EPA-82-F-
                           LEGEND
                        Poorly Sorted  Sands    £g] Silt and Silly Clay
                        Silty or Clayey Sands    0  Storm Sewer
                        Ground Water  Level                0150300
                                                          FEET
                                                                                   B' NE
              EPA-82-1
            Hydrogoologlc
             Section B-B'
          HILL AIR FORCE BASE
Figure 4.   Flow path of ground water from a JP-4 spill  at Eglin AFB to the point of discharge
             in a storm drain:  cross section.
66

-------
Table 3. Vertical Distribution of Oily Phase Hydrocarbons at Hill AFB
Core

Elevation
(ft)
TPH
Befnzene

821, near the spill area, water table elevatic
821-5
821-4
821-3
821-2
821-1
821-27 to
821-19
821-17
821-37
821-32
4,665.7
4,665.4
4,665.0
4,664.6
4,664.3
4,663.9 to
4,661.0
4,660.7
4,659.3
4,659.0
4,330
3,770
14,800
5,870
398
<300
1,290
1,370
3,830
Toluene
Ethyl-
benzene
1,2,4-
TMB
(mg/kg)
n 4,664.53 ft
0.0326
0.517
4.55
OJ.401
^0.01


c|.653
0.712
0,
0136
0.0266
0.235
2.73
12.6
0.142

0.591
0.182
0.032
14.5
4.8
47.7
17.5
0.556

3.39
2.72
1.24
49.9
42.7
167
69.8
4.43

2.34
5.38
8.37
82D, downgradient of spill, water table eleyation 4,631 .7 ft
82D-24
82D-23
4,631.0
4,630.7
77.1
572.0


q.271

<0.01

1.48

3.11
82C, lower end of spill, water table elevation 4,603.4 ft
82C-20
82C-19 '
4,603.5
4,603.2
593.0
638.0
<|o.oi
0
0062
0.0176
<0.01
0.00618
0.0180
1.03
1.16
82B, below the oil spill, water table elevation 4,608.3 ft
82B-20
82B-19
4,608.0
4,607.7
0.7
1.0
<0.01
<|o.oi
<0.01
<0.01
<0.01
<0.01
<0.01
<0.01
                                           jnd water from the area containing oily phase
Correcting for Dilution at Hill AFB

The TMBs were remarkably persistent in  gro
hydrocarbons  (Table 4). Concentrations of the TMBs in water from the lower edge of the spill
varied from 50  percent  to  147  percent  off the concentration  at the  source,  while  the
concentrations of benzene, toluene, ethylbenzene, p-xylene, m-xylene, and  o-xylene were
reduced to 0.02, 0.9, 5.6, 5.5, 3.4, and 0.23 percent of the initial concentration. The saturated
thickness of the channel sands was at most a fdw feet, and, therefore, there was little opportunity
                                                                                     67

-------
 for dispersive dilution of the contaminated ground water into clean ground water underneath
 the  plume.  Once  the  ground  water moved  past the  spill, it  was remediated  rapidly.
 Concentrations of all the aromatic hydrocarbons are low in well 82B just past the lower edge
 of the spill.
Table 4.   Changes in  Ground-Water Chemistry Along a Flow Path Through  a JP-4 Spill
           Undergoing Sulfate Reduction

Compound

Benzene
Toluene
Ethyl benzene
p-Xylene
m-Xylene
o-Xylene
1 ,3,5-TMB
1 ,2,4-TMB
1 ,2,3-TMB

BTEX and TMB
Oxygen
Nitrate +
Nitrite - N
Sulfate
Methane
Iron II
Location Along Flow Path to Surface Water
821
MW-11
82D
82C
82B
(«g/L)
2,740
327
486
784
1,370
1,140
162
495
240
336
90
139
230
635
204
71
165
69
96
10
147
149
383
103
129
183
89
4.9
3.1
27
43
47
2.6
238
324
120
<1
4.3
<1
<1
<1
<1
1.1
1.4
<1
(mg/L)
7.7



0.68

2.1
0.1
0.4
98
0.022
0.05
1.3
1.3
0.5
193
<0.001
1.7
2.1
0.5
0.1
50
0.002
0.84
0.001
1.2
0.4
64
<0.001
0.11
Stoichiometry  of  Bioattenuation at Hill AFB

Similar to the plume at Eglin AFB, neither oxygen nor nitrate is available for respiration of the
BTEX compounds within the spill area (Table 4). Oxygen concentrations as high as 8.0 mg/L
occur outside the spill. Unlike the case at Eglin, there was little accumulation of iron II and
68

-------
practically no methane production in the spill. Sulfate concentrations were high throughout the
spill. Assuming the following stoichiometry,
8 CH + 5 SO4.2  --> 5 S'2 + 8 CO2  + 4
                                           H20,
4.6 mg of sulfate would be required to degra Je 1.0 mg of BTEX and 1MB. The 7.7 mg/L of
BTEX and TMB present at the source would have a total theoretical demand of 35 mg/L of
sulfate. Concentrations in excess of the theorstical  demand remain in  the water after it has
moved away from the spill. There  is an adecuate supply of sulfate to  remediate the plume
through sulfate reduction alone.
                                      grou id
Concentrations of sulfate were higher in
lower concentrations of sulfate near the lower
exerted. It is just as possible that they represert
Hydrogen sulfide did not accumulate in ground
AFB. Unimpacted sediments at the site are tan
are black. Iron minerals in the aquifer matrix
precipitated as iron II sulfide.
                                            water undergoing intrinsic remediation at Hill
                                          or brown colored, while contaminated samples
                                          Must act to precipitate sulfide. Sulfide very likely
The concentrations of electron acceptors were
from the lower edge of the spill. Well 82B is
82E are 300, 400, and 500 ft downgradient o
table aquifer (Figure 4); well 82A is adjacentt
underneath the water table aquifer.
                                          compared in ground water at various distances
                                         jist outside the lower edge; wells 82F, 82H, and
                                           the spill. These wells are screened in the water-
                                           82F but is screened in the first confined aquifer
None of the wells outside the spill had signif
(Table 5). Water adjacent to the spill (82B)
                                        was
Intrinsic Bioremediation of Oily Phase JP
Two samples of floating oil from monitoring
analyzed by gas chromatography/mass spectra
mean molecular weight of the weathered fuel.
on duplicate samples.
The concentration of individual hydrocarbons
was estimated using Raoult's Law. With data f
petroleum hydrocarbon (mg/kg) was divided
 divided by 1 60 to express TPH in moles/kg o
  water near the source. It is possible that the
edge of the spill represent the sulfate demand
 natural variations in sulfate concentrations.
cant concentrations of aromatic hydrocarbons
  depleted in oxygen and nitrate.
                                          •4 Jet Fuel at Hill  AFB
                                         Well MW-1 0, near the midpoint of the spill, were
                                          netry(GC/MS) to determine the number-average
                                          Values of 1 56 and 1 60 daltons were determined
                                          in ground water in contact with oily phase JP-4
                                          •om Table 3, the concentration of an individual
                                           its molecular weight to express its concentration
 in moles/kg  core material.  The concentration of total petroleum hydrocarbons (TPHs) was
                                           re material. The mole fraction of the individual
 hydrocarbon was calculated by dividing moles pf individual hydrocarbon per kilogram by moles
 of TPH per kilogram.  Then the mole fraction was  multiplied by the water solubility of the
 individual petroleum hydrocarbons to estimate the equilibrium concentration in ground water.
                                                                                    69

-------
 TableS. Changes in Ground-Water Chemistry Along a Flow Path Downgradient of a JP-4 Spill

Compound

BTEX
TMB

Oxygen
Nitrate +
Nitrite - N
Sulfate
Methane
Iron II
Location Along Flow Path to Surface Water
82B
82F
82E
82H
82A
(«g/L)

-------
Table 6.  Comparison of the Concentration of
Benzene, Toluene, and Ethylbenzene Measured
         in Ground Water to the Concentration Predicted Using Partitioning Theory From the
         Mole Fraction of Those Compounds in Extracts of Core Material
Core

Elevation
(ft)
Benzene
Toluene
Ethyl-
benzene
1,2,4-
TMB
Measured Concentration («g/L)*
Predicted Concentration («g/L)
821, near the spill area, water table elevatio
11/93
821-5
821-4
821-3
821-2
821-1
821-27 to
821-19
821-17
821-37
821-32

4,665.7
4,665.4
4,665.0
4,664.6
4,664.3
4,663.9
to
4,661.0
4,660.7
4,659.3
4,659.0
2,740*
22.8
513
1,150
254
103
<300
1,870
1,940
13.2
n 4,664.53 ft
372*
5.5
55.6
165
1,923
320

. 410
119
7.5
486*
769
289
740
684
321

603
456
74.3
495*
900
. 884
881
929
869

135
307
171
82D, downgradient of spill, water table elevation 4,631 .7 ft
7/93
11/93
82D-23
4,631.7
4,631.7
4,630.7
96*
174*
1,769
10*
4.6*
15.6
147*
30.8
594
183*
119*
460
82C, lower end of spill, water table elevaticjn 4,603.4 ft
7/93
11/93
82C-20
82C-19
4,603.4
4,603.0
4,603.5
4,603.2
4.9*
<1*
<63
36.3
3.2*
8.4*
26.6
13.9
27*
6.8*
2.4
6.5
324*
68.9*
135
142
                                                                                  71

-------
Kinetics of Bioattenuation at Hill  AFB

There is no straightforward approach to calculate the kinetics of bioattenuation in ground water
in contact with oily-phase material. For the ground-water plume at Eglin AFB, the distance
between wells along a flow path and the estimated interstitial seepage velocity of ground water
was used to estimate residence time of the contaminants. At Hill AFB, the contaminated ground
water is in  contact with  oily-phase material.  The major  fraction of individual  contaminant
hydrocarbons in the aquifer is partitioned to the oily phase, which is moving slowly if it is moving
at all. Travel time of ground water is not related to residence time of contaminants, making it
impossible to determine the kinetics of attenuation from data at different locations collected at
the same time.  Kinetics  must be inferred from long-term monitoring data, which are not
available at the present time.

If we assume that the spill started 30 years ago, that the JP-4 at the lower edge was the first oil
spilled,  and that the  mole fraction of benzene in JP-4  used  at Hill AFB has not changed
appreciably overtime, a comparison of the mole fraction of benzene in 831-3 near the source
and 83C-19 near the lower edge of the spill indicates  a reduction in the mole  fraction of
benzene to 3 percent  of the  original concentration. If kinetics are first order on time, the rate
would be -0.1 I/year.
References

1.     Cozzarelli, I.M.,  R.P. Eganhouse, and  M.J.  Baedecker. 1990. Transformation of
       monoaromatic hydrocarbons to organic acids in anoxic ground-water environment.
       Environ. Geol. Water Sci. 1 6(2) = 135-141.

2.     Wilson, J.T., D.H.  Kampbell, and J.  Armstrong.  1 994. Natural bioreclamation of
       alkylbenzenes (BTEX)  from a gasoline spill in  methanogenic ground  water. In:  R.E.
       Hinchee,  B.C.  Alleman,  R.E.   Hoeppel, and  R.N.  Miller,  eds.  Hydrocarbon
       bioremediation. Ann Arbor, Ml: Lewis Publishers.
72

-------
A  Natural Gradient Tracer Experiment in
In Situ Biodegradation Rates: A Case for
a  Heterogeneous Aquifer With Measured
Natural Attenuation
Thomas B. Stauffer and Christopher P. Antwor h
Armstrong Laboratory, Tyndall AFB, FL
J. Mark Boggs
Engineering Laboratory, Tennessee Valley Lab

William G. Maclntyre
School of Marine Science, College of William
oratory, Morris, TN
and Mary, Gloucester Point, VA
Biodegradation  rates of organic  compounds  have been  measured in a  heterogeneous
unconfined aquifer at Columbus Air Force Base (AFB), Columbus, Mississippi, during a pulse
release experiment. Reaction rate calculations were based on a kinetic model that includes the
hydrologic characteristics of the aquifer. Degracation kinetics were approximately first order with
the following rate constants: benzene, 0.0066
d'1; o-dichlorobenzene, 0.0059 d'1.
J'1; p-xylene, 0.0141 d'1; naphthalene, 0.0063
transport models.  Field experiments for dete
Introduction

Biodegradation rates  of organic contaminants  in aquifers are needed for use  in fate and
rmination of in situ biodegradation rates  are
desirable because laboratory measurements may not relate to conditions in an aquifer. Madsen
(1)  notes in a recent review "that determination of microbial activity in disturbed, displaced
environmental samples incubated in the laboratpry is likely to be quantitatively, even qualitatively
different from the same determination in situ." Reliable estimation of an in situ biodegradation
rate requires  introduction of a  known mass of contaminant at a defined time zero,  and
observation of contaminant  concentration variation in space  and time. Confirmation of
biodegradation requires maintenance of mass b alances and determination of organic compound
degradation products. To our knowledge, no ex
rates and confirm biodegradation in an aquifc
>eriments to accurately measure biodegradation
r are found in the literature.
The  objective of this  research  was to measure in situ biodegradation  rates of organic
compounds. Accordingly, an experimental pulse injection of tritiated water and organic solutes
has been conducted at Columbus AFB at the macrodispersion experiment (MADE) site. The
injection was into the saturated zone of the heterogeneous unconfined aquifer formed by fluvial
                                         I I .L'
sedimentation.  Kinetics  of  in situ biodegradation of benzene,  p-xylene,  naphthalene, and
o-dichlorobenzene in the Columbus aquifer are reported here and related to the structural and
hydrologic properties of the aquifer. The Columbus aquifer material has a wide range of particle
sizes and large spatial variation in horizontal hydraulic conductivity (Kh), from <10"4 to 1 cm/s.
The  heterogeneity is in contrast to the generally lower (10~4 to TO"3 cm/s) values  and less
variable Kh field at the Borden Canadian Forces Base, Ontario  (2).
Asummary of hydrogeologic properties of the Columbus aquiferand a description of the MADE
site are given by Boggs etal. (3). Rehfeldtetal. 4) measured the spatial distribution of hydraulic
Symposium on Intrinsic Bioremediation of Ground Water
                                         73

-------
conductivity at the MADE site using borehole flow meters and other techniques. Figure 1  shows
the Kh distribution over a vertical section directed along the test plume axis, and indicates large-
scale heterogeneity and structures that Rehfeldt et al. (4)  refer to as channels. Close to the
injection wells Kh is relatively low, with values near TO"3 cm/s from the phreatic surface to the
lower confining layer and  extending about 40 m downgradient from the  injection location.
Immediately beyond this region, Kh increases to values near 1 O'1 cm/s in the upper 3 m of the
aquifer, which are maintained out to 200 m, while Kh in the lower portion of the aquifer remains
low. Thus, a near surface  channel crosses the intermediate and far field  at the site.  These
observations suggest that solute transport and distribution  might be analogous to the behavior
of a hypothetical leaking reactor located at the injection wells, with the near field region of the
aquifer representing the reactor vessel and the upper portion of the far field serving as the leak.
A leaky reactor kinetic model including radioactive decay and biodegradation was developed
for use with field data to obtain in s/fu  biodegradation rates for organic compounds  in the
Columbus aquifer. Radiocarbon measurements using 14C-labeled p-xylene were done to confirm
biodegradation of p-xylene to  its degradation products.
Experimental Methods

The experiment (MADE2)  is generally similar in design to the Borden site test described by
Mackay et al. (5). A 2-day pulse injection of water containing 3H2O and organic solutes was
begun on June 26, 1990. Injection wells were closely spaced on a line normal to the flow
direction at locations given in Boggs et al. (3). The intersection line formed by the well plane
and a vertical plane along the flow path is shown in Figure  1. Locations of injection  and
sampling wells are shown in plan view in Figure 2.
    65


    60

|=

J  55
 03
LLJ
    50
10°    10'1
                         10-2
10-"     (cm/s)
                                                           FLOW
                                                        4-
         25     50     75     100    125   150    175    200    225    250
                             Distance (m)
                                                                                  275
Figure 1.  Distribution of hydraulic conductivity over a vertical section containing the center line
          of the plume.

-------
     300
     250
     200
 i?150
    100
     50
    -50
               N
               Injectlor
                point
      -100
-50
                             i-
                                     X
Figure 2.  Plan view of the injection and sampling wells at the MADE2 site.
                                         barcad   A

                              multilevel sampler    •

                                   Injection well   a


                                       inset
                           15

                           10

                           5

                           0
                                                     A A
                                                          A A
                                                            A6&
                                             -15 -10  -5   0  5  10  15
50
On)
100
150
200
                                                                            75

-------
Experimental parameters for the MADE2 test were:  injection volume, 9,600 L; injection time,
47.5 hr; injection wells screened over a 0.6-m interval at 4 m below the phreatic surface in the
saturated zone of the aquifer; and five injection wells in a line normal, to the hydraulic gradient
and spaced at 1 -m intervals with equal flow to each well. Injection concentrations were: tritium,
55.6 nCi/mL; benzene, 68.1  mg/L; naphthalene, 7.23 mg/L; p-xylene, 51.5 mg/L containing
ring-labeled 14C p-xylene, 2.77 nCi/mL; o-dichlorobenzene, 32.8 mg/L. Tritium and 14C were
analyzed by  liquid  scintillation counting of water  samples. Concentrations of the organic
compounds in water were determined by solvent extraction into pentane containing toluene as
an  internal standard, and analysis of the extract was made by gas chromatography (GC) with
flame ionization detection.  Most sampling wells contained multilevel samplers, but BARCAD
water samplers were used in a few wells. Water was collected from the sampling wells and
analyzed to provide three-dimensional snapshots of solute concentrations at 27,1 32,224,328,
and 440 d after injection. Statistical moments for each snapshot were calculated by establishing
a triangular grid  followed by vertical integration of concentrations at each well, and spatial
integration was calculated overthe grid by the method of Garabedian (6), which was previously
applied at the MADE site by Adams and Gelhar (7). The  agreement between the calculated
zeroth  moments  and the  mass of 3H2O injected  confirmed mass balance for 3H2O.  Selected
wells from high and low solute concentration regions of the plume were analyzed for p-xylene
degradation by 14C counting methods 421 d after injection. Each watersample was subsampled
in the field. A 2-mL  subsample was mixed with cocktail and counted to obtain total 14C in the
water as either p-xylene or its degradation products. A 20-mL subsample was extracted with
2 mLof unlabeled p-xylene, and 1  ml of the xylene layer was transferred to scintillation cocktail
and counted for 14C p-xylene, which remained in the organic layer. Finally, a 10-mL subsample
was made basic by adding 1 ml 3N NaOH and 1 ml 3N Ca(NO3)2, and centrifuged to isolate
a precipitate containing Ca14CO3. The washed precipitate and 2 mL of supernatant water were
placed in separate scintillation vials and  counted. These analyses determined the amount of 14C
carbonate produced by complete degradation of 14C p-xylene and the amount of 14C in water-
soluble organic intermediate  products,  respectively.
Leaky Reactor  Model

Organic solutes  remaining in the low Kh region near the injection wells may be regarded as
semiconfined in  a reactor zone that is stirred by advection  and dispersion. This reactor is
considered to leak by advection through a near surface channel with relatively high Kh, with
replacement water supplied by ground-water flow across the upstream boundary of the reactor
zone. The conceptual model of this situation  is shown in Figure 3. This model compares well
with the Kh distribution shown in Figure 1. Reactions occurring in the reactor zone are tritiated
water (Tr) decay, with a  rate constant (kd) of 1.548 x 10"3 d"1, and biodegradation of organic
compounds, assumed to be first order with rate constants kb, kx, kn, and kc  for benzene (B),
xylene (X), naphthalene  (N), and o-dichlorobenzene (C), respectively. The  leak rate for each
solute from a well-mixed reactor should be first order, with the same constant, k|, for all solutes.
Sorption-desorption processes at aquifer material surfaces are assumed to be very fast relative
to the other processes considered here, and to be rate-limited by physical transport of organic
solute molecules to and from the solid-solution interface. Biodegradation and leakage occur
simultaneously in the reactor zone by the scheme given in Figure 3. The overall kinetics can be
expressed by the following set of linear differential equations with initial conditions: att = t0 =
0, [Tr]0  = 0.539 Ci, [B]0 = 660 g, [X]0 = 402 g,  [N]0 = 70 g, [C]0  = 31 8 g, where square
brackets represent solute mass and t  is time following injection.
76

-------
                    INJECTION
                     WELL

GROUND
WATER
FLOW

/
f
f
/ ^
V

,„„„„„„,, 	 J 	
REAC1
KB[B] '
KX[X]
KN[N]
KC[C]
T 1
ronz(
^DEG
DNE f
RADATION
IATES f
I
	 ^

CHANNEL
K (LEAK RATE CONSTANT)
^
Figure 3. Diagram of the leaky reactor model.
Analytic  solutions of equations 1 to  5 with  these  conditions are the  integrated rate laws
(equations 6 to 1 0) describing solute concentrcitions as functions of time. The leak rate constant
k| is calculated by solving equation 6, using the known kd of tritium. Constants kb, k,,, kn/ and k,.
are then determined by substitution of k( and solving equations 7 through 1 0.
            dt
                 = kb[B] -, kt[B]
           dt
           dt

= kn[N\ + k^N]


= kc(C] + k$C\
-k.
                           (1)

                           (2)


                           (3)

                           (4)


                           (5)
                                             (6)
                                                                                      77

-------
      In V'*l +  lnU?T
                                              (7)
*,-
                                ,
                                              (8)
            InV1'*/  .in,--, .*
                                              (9)
                   / p>]|
,     In l17^/    In U^
fc  = 	 + 	  + k,
                                             (10)
Organic solutes were retarded slightly relative to 3H2O by weak sorption on the aquifer material
(8),  but this effect was negligible  when  compared  with concentration  changes  due to
biodegradation. Laboratory batch sorption coefficients on a composite sample of Columbus
aquifer material for naphthalene, o-dichlorobenzene,p-xylene, and benzene were 0.085,0.065,
0.048, and 0.059 L/kg, respectively. Accordingly, this  model does not incorporate terms for
sorption of solutes in the reactor zone.
Results

The distribution of 3H2O in sampling wells on the plume axis 224 d after injection is presented
in Figure 4a. It is apparent that a large fraction of the 3H2O mass remained near the injection
wells  after 224  d. Biodegradation of organic solutes occurred primarily in a  reactor zone
approximately delineated by the p-xylene distribution at 224 d after injection. Figure 4b shows
that  p-xylene  was approximately confined within  20 m of the injection wells.  The same
distribution pattern also held for the other organic solutes. Confinement of organic solutes to
the reactor zone occurred because their biodegradation was rapid relative to the leak rate from
the reactor zone.

The spatial distributions of Kh in  Figure 1 and of 3H2O in Figure  4a are similar,  which is
expected  because hydraulic  conductivity  is  the  controlling  parameter for transport of an
unretained solute. High 3H2O concentrations remained in the neighborhood of the injection wells
224  d after injection, with lower concentrations in the upper portion  of the aquifer from about
30 m to 250 m downgradient. There was little 3H2O downgradient in the  lower portion of the
aquifer between 30 m and 1 75 m depth. This distribution indicates that solutes were transported
downgradient from the low Kh near the injection wells through an upper channel whose size,
location, and Kh govern the solute loss rate from the source region. Very little transport occurred
78

-------
 in the lower portion of the aquifer. The observed transport of 3H2O during this experiment was
 consistent with the leaky reactor model.
     65

"E"  60

J  55
"a
J>  50
 a?

     45
                                       tritium
                                                                      (a)
                 25    50    75    100   125   150   175   200   225   250   275
 E 60

J 55
"a
.£ 50
 us
                                     p-xy!
                                          ene
10°     10"'
                         C/Co
                                          10
                                            -4
                                                                     (b)
           0     25    50    75    100    25   150   175   200   225   250    275
Figure 4. Distribution of (a) normalized tritium and
         over a vertical section containing tf
                                             (b) normalized p-xylene concentration zone
                                         e center line of plume motion.
                                                                                79

-------
 Leakage of solutes was determined by the loss of tritium from the reactor zone, defined as the
 portion of the plume volume within 10m downstream from the injection wells. The amount of
 tritium in the reactor zone at a given time was calculated by spatial integration. Total tritium in
 this zone decreased exponentially with time, as shown in Figure 5. The tritium values have been
 corrected for radioactive decay. This plot indicates that solute leakage from the reactor is first
 order, with leakage rate  constant, k,  = 5.45 x 1 O'3 d'1. A mass balance for tritium calculated
 using spatial integration  over the entire  plume showed that the injected tritium mass was
 accounted for at the times given in Figure 5. This implies that the tritium decrease in the reactor
 zone was indeed due to  a process analogous to leakage.

 Biodegradation rates within  the reactor zone are approximately first order, as shown in Figure
 6. Departure of the curves from linearity is attributed to microbial processes. There  is an
 apparent lag period for  microbial activity soon after the injection, during which degradation
 rates are low. This  lag  period is followed  by degradation  at a maximum rate, which  is
 characteristic of the microbial metabolism. Finally, the rate decreases late in the reaction, as the
 solutes  (substrates) are depleted.

 Biodegradation  rate constants for the  organic  solutes in  the reactor  zone (kmax(reactor)),
 determined from the maximum slopes in Figure 6, are presented in Table 1 . The maximum rate
 constants (kmax(corrj(reactor)) in the second column have been corrected fororganic solute leakage
 from the reactor zone. Maximum biodegradation rate constants  for each organic solute were
 also calculated by spatial integration over the entire well field, and are included in Table 1 as
          field). These values are independent of the leaky reactor model.
Degradation rate calculations for p-xylene are based on analyses of p-xylene  by GC. Ring-
labeled 14C p-xylene was included in the injection solution to demonstrate that reductions in p-
xylene concentration were a result of biodegradation. Microorganisms mineralize 14C p-xylene
to water-soluble labeled intermediates and 14CO2 (predominantly as H14CO3" at ground-water
pH). 14C counting of whole water samples does not distinguish between degraded and intact p-
xylene, and thus does not measure degradation of 14C-labeIed organic compounds.  Detection
of these products provides a strong  indication that  p-xylene has biodegraded in the aquifer.

Results of measurement of 14C p-xylene degradation are given in Table 2. Total 14C in the water
sample and the amount of 14C in the water after extraction were used to calculate the fraction
of p-xylene converted to all products. Total 14C in the water sample and the 14C in  the carbonate
precipitate were used to calculate the fraction of p-xylene converted  to CO2.

These p-xylene conversionfigures compare well with mass-balance- based conversions calculated
from  the total p-xylene remaining in  the plume at day 421, which were interpolated  from
snapshot data. Mass-balance-based conversions are based on GC analysis and  on the known
mass of p-xylene injected. They are  included in the last column of Table 2.
80

-------
              -3
                      100     200

                      DAYS  AFTER
Figure 5. Tritiated water content in the reactc r organic as a function of time
            -1 -
            -2-
            -4-
            -5'
                     300    400

                     INJECTION
                                                    500
D BENZENE
A NAPHTHALENE
O p-XYLENE
A o-DICHLOROBENZENE
              0      100     200

                     DAYS AFTER



Figure 6.  Degradation curves for the compo
                     300      400

                     NJECTION
500
                     jnds in the reactor zone.
                                                                         81

-------
Table 1. Maximum Biodegradation Rate Constants From the MADE2 Site

Benzene
p-Xylene
Naphthalene
o-Dichlorobenzene
I^Jreactor)
(d-1)
0.0120
0.0196
0.0118
0.0114
kmaxicor^reactor)
(d-1)
0.0066
0.0141
0.0063
0.0059
k^whole field)
(d-1)
0.0104
0.0187
0.0104
0.0100
Table 2.  Degradation of p-Xylene in Water Samples Taken After 421 Days, Expressed as
          Weight Percent Converted

[xylene]>l ppm
(n-8)
[xylene]>l ppm
(n=10)
% 14C p-Xylene
Converted to All
Products
85.1 (± 6.3)
82.6 (± 6.1)
% 14C p-Xylene
Converted to
C02
73.3 (± 11.1)
74.2 (± 6.2)
Mass Balance
Based % p-Xylene
Converted
98
98
The values in the second column are on the high, but normal, side for 14CO2 release.  This
implies that most of the p-xylene went to energy production and was not converted to biomass.
The difference between degradation in the high and low concentration regions of the aquifer
is not significant. The difference of the means for p-xylene converted to all products and p-xylene
converted to CO2 may imply that some intermediate products are present, but this difference
may also be attributable to analytical anomalies (i.e., loss of Ca14CO3 during precipitate transfer
prior to counting). Agreement between the first and third columns of Table 2 indicates the
consistency of p-xylene degradation measurements by GC and 14C counting methods. The small
difference between these numbers is due to incorporation of 14C  into biomass and insoluble
carbonates. Most p-xylene degradation products apparently remained in the local ground water.
Discussion

Biodegradation rates given in Table 1 are based on field observations of solute behavior and
are an  essential input in  modeling organic contaminant fate and transport in  this aquifer
material. The relative rates are as expected, with p-xylene most rapidly biodegraded. Figure 6
indicates that biodegradation in the Columbus aquifer is a first-order process. Simkins and
Alexander (9) have indicated that biodegradation can be expected to follow Monod kinetics for
a microbial population not limited by nutrients and with a sufficient substrate concentration, but
first-order kinetics are observed at low substrate concentrations. Organic solute concentrations
82

-------
concentration can often be expressed as a firs
in the MADE2 test were quite low, so this conclusion  is consistent with experimental results.
Larson (10) notes that the rate  of  decreas 3 in concentration  as  a function  of  substrate
-order equation, and that first-order kinetics are
generally expected for biodegradation at low organic substrate concentrations. Applicability of
the leaking reactor model is site specific, and its success with the MADE2 test data is a fortuitous
circumstance dependent on the arbitrarily selected  positions of the injection  wells in the
Columbus aquifer. It is encouraging that the two methods of calculating  biodegradation  rates
gave similar results, thus providing confidence
hat these rates are useful for predictive purposes.
Organic solutes in MADE2 biodegraded under aerobic conditions in the aquifer. The organic
solute concentrations injected were chosen TO be too low to significantly deplete dissolved
oxygen  in the reactor zone of the aquifer. Maintenance of oxic conditions was confirmed by
monitoring dissolved oxygen and redox potential of water samples during the experiment. Thus
the biodegradation rates reported here were not affected by oxygen limitation. For very large
releases of similar organic compounds, the conditions near the source might rapidly  become
reducing due to the biological oxygen demand. Kinetic models to determine biodegradation
rates in  this situation would be complex due to the need to include oxygen transport terms and
the potential for degradation by anaerobic bacteria. It is therefore wise to use  small injection
amounts in field experiments to determine biodegradation rates.

In MADE2, the organic solutes degraded quickly. These results suggest that, for similar solutes,
aquifer  remediation activities should be  restricted to the source region, which might include
pumping to remove nonaqueous-phase liquids or excavation of contaminated aquifer material.
The source would be  reduced  but not eliminated  by this treatment.  In an  aquifer with
approximately steady flow, the plume of organic solutes from the reduced source would reach
a steady state,  with the boundary  determined by the hydrology of the site, sorption, in  situ
biodegradation, and oxygen and nutrient sup Dly.
Conclusions

Controlled-release experiments similar to the MADE2 test are needed to determine accurate
biodegradation rates for use in ground-water contaminant fate and transport models of aquifer
situations. The MADE2 study has demonstrated  the  practicality of these experiments and
obtained in situ degradation rates forfour organic contaminants in the Columbus aquifer. These
rates will be used in the design and modeling stages of a new field test at the MADE site, which
is now in preparation.
References

1.   Madsen, E.L. 1 991. Determining in situ biodegradation: Facts and challenges. Environ. Sci.
     Technol. 25(10):!,663-1,673.

2.   Robin, M.J.L., E.A. Sudicky, R.W. Gillham, and R.G. Kachanoski. 1 991. Spatial variability
     of strontium distribution coefficients and their correlation with hydraulic conductivity in the
     Canadian Forces Base Borden aquifer. Water Resour. Res. 27(1 0):2,61 9-2,632.
                                                                                    83

-------
3.   Boggs,  J.M., S.C. Young, and L.M. Beard.  1992. Field  study  of dispersion in a
     heterogeneous aquifer, 1.  Overview and site description. Water Resour. Res. In press.

4.   Rehfeldt, K.R., J.M. Boggs, and L.W. Gelhar.  1992.  Field  study of dispersion in a
     heterogeneous aquifer, 3. Geostatistical analysis of hydraulic conductivity. Water Resour.
     Res. In press.

5.   Mackay, D.M., D.L Freyberg, and P.V. Roberts.  1 986. A  natural gradient experiment on
     solute transport in a sand aquifer, 1. Approach and overview of plume movement. Water
     Resour.  Res. 22(13):2,01 7-2,029.

6.   Garabedian, S.P., D. LeBlanc, L.W. Gelhar, and M.A. Celia. 1991. Large-scale natural
     gradient tracer test in sand and gravel, Cape Cod, Massachusetts, 2. Analysis of spatial
     moments for a nonreactive tracer. Water Resour. Res. 27(5):911 -924.

7.   Adams,  E.E., and L.W. Gelhar. 1 992. Field study  of dispersion in a heterogeneous aquifer,
     2. Spatial moments analysis. Water Resour. Res. In press.

8.   Maclntyre, W.G., T.B.  Stauffer, and  C.P. Antworth.  1991. A comparison of sorption
     coefficients determined by batch, column,  and  box methods  on a low carbon aquifer
     material. Ground Water 29(6):908-913.

9.   Simkins, S., and M. Alexander. 1 984. Models for mineralization kinetics with the variables
     of substrate concentration and population density. Appl. Environ. Microbiol. 47(6) :1,299-
     1,306.

10.  Larson,  RJ.  1979. Role of biodegradation  kinetics in predicting environmental fate. In:
     Maki, A.W., et  al.,  eds.  Biotransformation and fate  of chemicals in the  aquatic
     environment. Washington DC: American  Society for Microbiology.
84

-------
Traverse City: Distribution of the Avgas Spill
David W. Ostendorf
Civil and Environmental Engineering Departm
  ent, University of Massachusetts, Amherst, MA
Abstract
The capillary tension/liquid saturation equati
Parker (2) provide a reasonably accurate desc
phase aviation gasoline (avgas) in solid core
Station in Traverse City, Michigan.
Introduction

The depth of a continuous liquid below the c
tension in the absence of vertical velocity, sin
forms a vertically uniform hydraulic head unde
idns
  of Parker and Lenhard (1) and Lenhard and
iption of the vertical distribution of the separate
samples taken from the U.S. Coast Guard Air
  round surface is directly related to its capillary
  :e the sum of the potential and pressure heads
  static conditions. Thus, the theory (1) describing
capillary tension/saturation relations for immi: cible, continuous fluids also finds expression as
vertical profiles of water and avgas in the absence of infiltration.  Lenhard and Parker (2)
 recognize this equivalence and propose profi
  es for total liquid and avgas saturation that are
 reasonably borne out by Ostendorf et al. (3) in rheir analysis of solid core data from five stations
 in the avgas plume at Traverse City (Figure 1).
 implications of the distribution.
  We summarize the theory here and discuss some
 2. The water is more strongly attracted to the
 Vertical Distribution of Free Avgas

 The vertical distribution of water and free avgas in the contaminated soil is idealized in Figure
  solid grains than the avgas and fills the smaller
 pores in the soil, forming an interface with free avgas. The free avgas in turn faces the soil gas
 in the large pores, which control the overall saturation of the soil as a consequence. Residual
 avgas may be trapped as discontinuous bubbles within the water phase due to hysteresis and
 a fluctuating water table.
 The volumetric water 0W, residual avgas 0LR,
  and  irreducible water 0WR contents combine to
 define an effective apparent water saturation Sw and comprise the total saturation S when they
 are added to the  free avgas content 0LF (1).
 Symposium on Intrinsic Bioremediation of Ground Water
                                                                                      85

-------
                                             Interdiction'
                                              Well Line
                                X
                                                 /'.  Ni
                /Flow
            Direction
               /
                                       Smith
                             Building
                                606
                                                       • SOBT
           /SUBS     /  Xpl"me
           50CL* 50CE     Boundary
50 m ,   /
                                                        Administration
                                                        Building
                                                   Spill
                                                   Origin
 Figure 1. Site plan, U.S. Coast Guard Air Station.
                                                                      Ground
                                                                      Surface
                                     Volumetric Liquid Content
                                                                   Air
                                                                 Water
                                                               Monitoring
                                                                 Well
Figure 2. Vertical distribution of liquids in soil and a  monitoring well.
86

-------
                                          ri  - (9,
                              S =
                                          n
                                                WR
                                              " WR
                                                                                 (la)
                                                                                 (Ib)
with porosity n. The total saturation is contrail sd by the tension across the avgas/air interface
and varies with depth b below the ground surface in the absence of dynamic effects associated
with infiltration
                         S = S
                               w
                    S  = {1 + rA(bL-b)]<*}<]
                         S = 1
distribution to separate phase contamination
elevation and extends above the  avgas table
                                                          bj
                                                                                 (2a)
                                                                                 (2b)
(2c)
Lenhard and Parker (2) derive equation 2 as an application of van Genuchten's (4) pore size
                                           n soil. Figure 2 displays the minimum depth
     JM
of free  avgas occurrence, along  with the avgas  table bL and  water table bw existing in  a
hypothetical monitoring well. Note that the avgas does not occupy all the pores at a given
                                            in the soil  due to capillary tension. Thus, the
monitoring well levels, though necessary to determine liquid tensions, do not explicitly determine
the vertical distribution of avgas in the soil. Th 3 pore size distribution of the soil is also needed
to specify the profile.
These latter data are characterized by the pore
                                           size uniformity exponent a and the scaling factor
/?L appearing in equation 2b. The latter parameter may be expressed in terms of the mean pore
radius 7 by noting that the avgas/air surface tension a^ relates fluid tension to interfacial radius
                                                                                   (3)
with gravitational acceleration g, avgas
1 lists parameter values calibrating the data
(circles) and theory (curve) for a typical total
pore radius of 5.4 x 10~5 m is implied by equal
based model of 7 checks this estimate. We
pore size flowing full, and equate this to a mo
                                            2a,
                                              LA
                                      density pL, and an assumed zero contact angle. Table
                                         from five stations at Traverse City (3). Observations
                                         saturation profile are sketched in Figure 3. A mean
                                           ion 3. A simple extension of a classical grain size
                                          n 3te that 7/2 is the hydraulic radius of the mean
                                           Jified Fair and Hatch (5)  estimate of the quantity
                                                                                       87

-------
                                      r
                                     "2
                                           V.
voros
                                            ^SOLIDS
                                    (4a)
                                   r =
                                                                                 (4b)
 Table 1. Total and Avgas Saturation Profile Parameters at Traverse City
Symbol
n
C'WR
PL
a
A.
Aw
A
Y
Parameter
Porosity
Irreducible moisture content
Avgas density
Pore size uniformity
Avgas scaling factor
Water scaling factor
Water table amplitude
Trapping factor
Value
0.367
0.059
7.07 kg/m3
3.00
8.20 rrr1
1 .53 m'1
0.35 m
40
The observed mean grain size of 3.8 x 10'4 m at Traverse City leads to a mean pore size value
of 5.7 x 10"5 m, in excellent agreement with the equation 3 value.

The water/avgas interface controls the water saturation in the  presence of the free avgas
      (b
                                                     w
                                                               bM)
(5a)
                          Sw-1
            bw)
(5b)
                                                   •'LA
                                              ''WL
                                                                                 (5c)
with water density p and scaling factor/?w predicated on the water/avgas surface tension <7WL.
The upper extent bM of free avgas may be estimated by equating the total saturation (equation
2b) and the water saturation (equation 5a) at this elevation. A water/air interface with a surface
tension <7WA governs the water saturation above this depth

-------
                                      _
                                  (_J   —
                 sw  =
                             ''WA
The free avgas saturation S^, in view of equa
                                      LF
                                           S - S
                            6 -
                             0.0
                                             -6
                                            L f^
                                      (6a)
                                      (6b)
ion 1, is simply given by
                                                w
                                       (7)
Figure 4 shows a typical free avgas profile atth 3 site, based upon the scaling factors of Table 1,
        1.0
Figure 3. Total saturation, Core 50BT.
                          5.5
                            0.0
          0.2
 Figure 4. Free avgas, Core 50CE.
                                                                                     89

-------
 Vertical  Distribution of Residual Avgas

 Residual avgas may be attributed to hysteretical trapping of product as it rises and falls through
 the water wet soil over a fluctuating water table. Free liquids at a given depth b experience
 progressively stronger capillary tensions as the water table falls due to  their higher position
 above this  reference level. After the table attains a maximum depth and begins to rise, the
 water/avgas  interface becomes steadily  larger due  to a decreasing capillary tension. The
 encroaching water occludes some of the avgas, giving rise to a discontinuous residual fraction
 within the water phase.

 Historical maximum and minimum effective water saturations SWMAX, SWM,N are established  by
 corresponding minimum (bWM|N) and maximum (bWMAX) water table depths, respectively
                             WMIN
                                                                                  (8a)
                                                                                  (8b)
                                                                                  (8c)
Ostendorf et al. (3) infer the historical water table excursion amplitude A cited in Table 1 from
water level variations in nearby Lake Michigan. Parker and Lenhard (1) suggest that the extreme
saturations induced  by extreme water positions trap residual avgas saturation SLR that is given
by
                             1-S,
                                 WMIN
                                                  (bWMAX>b>bWMIN)
                                       (9)
with empirical trapping factor Y. Ostendorf et al. (3) calibrate residual saturation profiles from
their five stations with the trapping factor cited in Table 1, with the typical results sketched  in
Figure 5.
                       E
                      .a
                            0.00     0.01
0.02    0.03
                                           LR
Figure 5. Residual avgas, Core 50CL.
90

-------
Discussion
The vertical distribution of free and residual avjas has important implications forthe estimation
of separate phase contamination from monitoring well observations. The depth integrated mass
MLF of free avgas may be estimated formally by integrating equations 2b, 5a, 6a, and 7 from
the water table to the upper limit of free avgqs occurrence, with the result
MO,=
                  Mu^-?1
                          'w
K»-[
                                            bw
                                              S^db
                                                (10a)
                                                                                (lOb)
                                                            ft.
                                                                                (IOC)
The integral function l(y) varies with the uniforrrpity exponent of the pore sizes, while its argument
is determined by the scaling factors and the p -oduct thickness in the monitoring well. Figure 6
displays the equilibrium variation of the depth
function of the monitoring well avgas thickness
this estimate, since the trapped avgas is not cor tinuously connected to product in the monitoring
well. Fora point of reference, Figure 5 implies
5.3 kg/m2; if this mass were free avgas, we
monitoring well.
           integrated mass predicted at Traverse City as a
           We note the exclusion of the residual mass from
          a depth integrated residual avgas mass of about
          would observe a 0.3-m product thickness in a
The methylene chloride extract chromatograms are composed primarily of known compounds
(6), so that the composition of the avgas can be examined at a given depth. We note distillation
of avgas as a function of depth in the solid core samples, as suggested by a profile from station
SOBS (Figure 7). The open circles correspond TO relatively volatile avgas compounds, with pure
phase vapor densities of 0.27 to 1.00 kg/m3 lot 12°C). This fraction becomes more important
in the lower elevations due presumedly to decreased volatilization losses in the wetter region of
the soil. The mid-range volatiles (closed circles), with vapor densities of 0.14 to 0.27 kg/m3, are
quite uniformly distributed, while the heavy compounds (triangles), with vapor densities less than
0.14 kg/m3 in magnitude, tend to dominate the blend of hydrocarbons in higher, drier soil due
to stripping away of the lighter fraction.
                                                                                    91

-------
                       CNJ
                       <
                       £


                       JP
                                      b —b  , m
                                      W   L
Figure 6. Depth integrated free avgas mass.
                       8- 5.5
                       Q
                                 O  High
                                 •  Medium
                                 V  Low
                                _!	i
                                10    20    30    40

                                  Avgas  Fraction, %
Figure 7. Avgas composition, SOBS.

-------
The partitioning of avgas into free and resic
relatively  high-permeability, saturated soil
                                          ual fractions has important  implications on  its
downgradient fate and transport as well. Figure 2 suggests that the free avgas occupies a
                                          region
                                          il I   r •
below  the  avgas table and a  lower-
permeability, unsaturated zone above the table. Since  both these regions share  a common
horizontal gradient due to the slope of the avgas table,  we anticipate slower and faster zones
of horizontal separate phase transport. This vertical profile of horizontal specific avgas discharge
is very nonuniform for sands with high values  like that  at the site. The phenomenon can be
approximated as a pseudosorptive process, with a linear balance between a mobile fraction and
a reversible, immobile fraction "sorbed" by capillary tension (7). Pursuing this analogy further,
the residual avgas can be thought of as an irreversibly  sorbed partition, lost from the mobile
fraction by hysteretical trapping. The net effect of these mechanisms is that the avgas travels with
the underlying ground water, but at a retarded velocity.
                                           the soil due to its direct interface with air. The
                                        remediate, since it is  surrounded by essentially
                                          initial period of relatively rapid remediation in
                                          owed by an asymptotically lower removal rate
The free avgas is relatively easy to strip out o
residual avgas is  much  more  difficult to
immiscible water. We accordingly expect an
response to soil venting  or air sparging, fol
exacerbated by avgas distillation.
References
1.      Parker, J.C., and R.J. Lenhard. 1 987. A model for hysteretic constitutive relations
        governing  multiphase flow,  1. Satu "ation  pressure relations. Water Resour. Res.
        23:2,187-2,196.

        Lenhard, R.J., and J.C. Parker. 1 990. Estimation of free hydrocarbon volume from fluid
        levels in monitoring wells. Ground Water 28:57-67.
2.
3.      Ostendorf, D.W., R.J. Richards, and
        Ground Water 31:285-292.
                                          :.P. Beck. 1 993. LNAPL retention in sandy soil.
        conductivity of unsaturated soils. Soil
        of water through sand. J. Am. Waterv
        residual aviation gasoline in sandy s
4.     van Genuchten, /vi.T. 1980. A clossd form  equation  for predicting the  hydraulic
                                          Sci. Soc. Am. J. 44:892-898.
5.      Fair, G.M., and L.P. Hatch. 1 933. Fu idamental factors governing the streamline flow
                                          •orks Assoc. 25:1,551 -1,565.
6.      Ostendorf,  D.W., LE. Leach, E.S.  Hnlein, and Y.F. Xie.  1991. Field  sampling of
 7.
                                          il. Ground Water Monitor. Rev. 11 =107-120.
        Ostendorf, D.W. 1 990. Long-term fate and transport of immiscible aviation gasoline in
        the subsurface environment. Water Sci. Tech. 22:37-44.
                                                                                     93

-------
 Traverse City: Geochemistry and Intrinsic Bioremediafion of BTX Compounds

 Barbara H. Wilson, John T. Wilson, Don H. Kampbell, and Bert E. Bledsoe
 U.S. Environmental Protection Agency, Robert S. Kerr Environmental Research Laboratory,
 Ada, OK

 John M. Armstrong
 The Traverse Group, Inc., Ann Arbor,  Ml
 Introduction

 Loss of petroleum products from underground storage tanks, pipelines, and accidental spills is
 a majorsource of contamination of unsaturated soils, aquifer solids, and ground water. Volatile
 aromatics such as benzene, toluene, ethylbenzene, and the xylenes (BTEX) are more soluble in
 water  than  the aliphatic and higher molecular-weight aromatic constituents of petroleum
 products (1). Once released to the subsurface, petroleum compounds are subject to aerobic
 microbial processes. The low-molecular-weight alkanes and aromatics are readily biodegraded
 in  oxygenated ground water,  depleting the  ground  water of available  oxygen  (2,3).
 Reoxygenation of the ground water may occurthrough reaeration from soil gases, ground-water
 recharge, and usually inefficient mixing with surrounding  oxygenated ground waters (4,5).
 Although ground waters nearthe perimeter of the contaminant plume may be reoxygenated, the
 interior of the plume  will remain anoxic for a  distance downgradient. Anaerobic biological
 processes can account for  most of the  removal  of  BTEX from  the   plume  (6,7).  The
 biogeochemical mechanisms that contribute to anaerobic processes in the subsurface, however,
 are not well understood.

 The impact of anaerobic microbial  processes  on the fate  of  monoaromatics, substituted
 aromatics, and chlorinated hydrocarbons in anoxic subsurface environments has been studied
 in laboratory and field situations (8-12).  Field evidence of biotransformation  of o-, m-,  and
 p-xylene was  observed in  methanogenic landfill  leachate  with their preferential  removals
 compared with otheralkylbenzenes present (1 3). Methanogenesis has been observed at two sites
 with  ground  water  contaminated  with creosote  (14,15).   Denitrifying,  iron-reducing,
 sulfate-reducing, and methanogenic activities found in ground water at the first site were highly
 correlated with the biodegradation of creosote; methane was only detected in the ground water
 that had been contaminated  with creosote. Intermediate products of methane fermentation,
 formate and acetate, were found in the ground water at the second site.

 In 1 969 the unsaturated soil and ground water underlying the U.S. Coast Guard Air Station at
 Traverse City, Michigan, were contaminated with  an.estimated 25,000 gal of aviation gasoline
 when a flange in an underground storage tank failed. Dissolution of the aromatics in the ground
 water resulted in concentrations of 36 mg to 40 mg of total alkylbenzenes per liter near the
 center of the plume. The subsurface contamination existed  as residual-phase hydrocarbon,
 dissolved-phase aromatics, and gaseous hydrocarbons resulting from volatilization (16-1 8). The
 plume of dissolved-phase aromatics extended from the air station into the East Arm of Grand
Traverse Bay  and  affected   numerous drinking water wells. The  area  near the leaking
 underground storage tank was used to store  degreasing solvents and to conduct degreasing
94
                                                    Symposium on Intrinsic Bioremediation of Ground Water

-------
operations during aircraft maintenance. A smc II plume of chlorinated solvents lies adjacent to
the gasoline plume.
Geochemical Characferization
Geochemical analyses of the water samples collected at Traverse City revealed waters of four
distinct geochemistries: 1) the heart of the plume, 2) an anaerobic zone of treatment, 3) an
aerobic zone of treatment, and 4) a pristine or renovated zone. The water from the heart of the
plume contained high concentrations of methane and BTEX, with no detectable oxygen. These
waters were surrounded by an anaerobic zone of treatment with greatly reduced concentrations
of dissolved aromatics, no oxygen, and substantial concentrations of methane. Surrounding the
anaerobic zone of treatment was an aerobic zone of treatment with measurable oxygen, small
quantities of methane, and very low concentre tions of the alkylbenzenes. The perimeter of the
plume was surrounded by a renovated or prist
detectable alkylbenzenes, and no methane.
   ne zone with high concentrations of oxygen, no
Gas chromatography/mass spectrometry (GC/MS) analyses of the waters confirm the presence
of BTEX at Traverse City. Also found were the chlorinated  compounds 1,2-dichloroethane
(1,2-DCA), tetrachloroethylene (PCE), and 1,1,1 -trichloroethane (TCA). The GC/MS analyses
of waters from wells R, S, and Q (Figure la) identified phenols and aromatic acids indicative
of anaerobic microbial action on  the soluble  aromatic constituents of petroleum products
(7,19,20). These compounds are found in portions of the plume with substantial concentrations
of methane and no detectable oxygen, and are probably precursors of the methane. Complete
information  on the geochemical characterization  of the ground waters  at Traverse  City,
collection of aquifer material, analytical methods used,  and  microcosm construction may be
found in Wilson et al. (21).
Laboratory Studies

To  confirm  field  evidence of  intrinsic  biorernediation, laboratory microcosm studies were
conducted on aquifer material from the U.S. Coast Guard Air Station. Material for aerobic and
anaerobic fate  studies was collected from three locations in the plume (Figure la). Aquifer
material from site A (1 1.7 m to 1 2.3 m below
was used to construct the microcosms  for the
   land surface), the zone of anaerobic treatment,
  anaerobic fate study. Aquifer material from site
B (9.6 m to 1 0.2 m below land surface), the aerobic zone of active biological treatment, and
from site C (6.6 m to 7.2 m below land surfa
  :e), the pristine or renovated zone, was used to
prepare the aerobic fate studies. Autoclaved controls were prepared from the site C material.
The compounds added to the microcosms were
and  chlorobenzene.
The initial compound concentrations and resu
are shown in Table 1. Benzene, toluene, p-:
aerobic and anaerobic aquifer material. The
each of the three geochemical zones studied,
aerobic zone of treatment, or the renovated
of the compounds in the anaerobic aquifer
p-xylene, and o-xylene had been reduced one
   benzene, toluene, p-xylene, o-xylene, TCA, TCE,
  ts of the fate studies at various incubation times
 •x/lene, and  o-xylene were biodegraded  in both
  •emovals were quite rapid for all compounds in
  whether in the anaerobic zone of treatment, the
(pristine) material. By the end of 8 wk of incubation
 rraterial, the concentrations of  benzene, toluene,
  order of magnitude. The biotransformation of the
                                                                                   95

-------
 four compounds in the aerobic zone of biological treatment occurred even more rapidly. By the
 end of 2 wk of incubation, the concentrations for all four compounds were decreased by two
 orders of magnitude. Similar losses were seen in material from the renovated/pristine zone.


 Table 1.  Behavior of Benzene, Alkylbenzenes, TCAa, TCEb, and Chlorobenzene in Aquifer
          Material From an Aviation Gasoline Plume

Subsurface
Material
A,
anaerobic
treatment,
11.7 mto 12.3 m
below land surface

B,
active aerobic
treatment,
9.6 m to 10.2 m
below land surface

c,
aerobic renovated,
6.6 m to 7.2 m
below land surface


c,
autoclaved,
6.6 m to 7.2 m
below land surface

Compound lftg/\- Pore Water)

Week
0

4

8
95C
0

2

4
14
0

2

4
14
0

2

4

Benzene
450

12

6
ndd
450

2

5
2
420

4

1
—
420

380

240

Toluene
420

56

40
nd
420

2

5
2
380

3

3
—
380

290

230
m +
p-Xylene
440

78

17
nd
390

1

2
1
370

3

1
—
370

200

170

o-Xylene
410

41

6
nd
390

2

1
nd
370

3

1
—
370

190

180

TCA
570

420

580
73
600

440

540
430
600

580

620
650
600

590

560

TCE
540

260

340
54
650

440

480
260
540

540

600
500
650

540

430
Chloro-
benzene
500

66

34
3
500

53

50
30
500

100

106
67 '
500

300

210
°1,1,1 -Trichloroethane
bTrichloroethylene
'Concentrations at this time interval were determined by GC/MS; concentrations at other time intervals were
determined by GC. All values are means of triplicate analyses.
dNot detected, detection limit of 0.1 fig/L


At the end of 4 wk of incubation, the respective concentrations for benzene, toluene, p-xylene,
and  o-xylene in the autoclaved  samples were 57 percent, 61  percent, 46 percent, and 49
percent of the original concentrations. Due to the removals of BTX in the controls, a second set
of autoclaved samples  was  prepared. The  concentrations after 1 6  wk of incubation  in the
duplicated controls were benzene, 83 percent; toluene, 66 percent; m- + p-xylene, 51 percent;
o-xylene, 50 percent; TCA, 86 percent; TCE, 73 percent; and chlorobenzene, 58 percent. The
96

-------
cause for the removal of organics in the contrc Is has not been determined; however, sorption
to aquifer solids probably occurred. Chlorobenzene was also biodegraded in each of the three
geochemical zones studied.  Decreases of  one  order of  magnitude were  observed for
chlorobenzene in both the aerobic and anaerobic zones of biological treatment after 4 wk of
incubation. Similar removal was seen in the renovated/pristine material after 14 wk of incuba-
tion. In the  autoclaved samples, 44 percent of the chlorobenzene  remained  after 4 wk of
incubation.
No significant biotransformation of TCA or TC
the end of 8 wk of  incubation.  Evidence of
:, compared with the controls, was observed at
reductive dechlorination of both compounds,
however, was indicated by GC/MS analyses of anaerobic microcosms at 95 wk of incubation
by the identification of 1,1-dichloroethylene
(22,23).
1,1-DCE) and 1,1-dichloroethane  (1,1-DCA)
Headspace  concentrations of methane were  measured  immediately  before  sampling to
determine the maintenance of methanogenic conditions in the microcosms. Methane was found
m all the anaerobic samples, with concentrations ranging from 50 ppm  to TOO ppm; no
methane was found in the headspace of the aerobic or autoclaved samples.
Correspondence  Between Laboratory and Field Data

As part of a settlement with the State of Michigan, the U.S. Coast Guard monitors alkylbenzene
concentrations in selected monitoring wells quarterly. Three of the monitoring wells (M30 near
site S, M31  near site Q, and M2  near site A in Figure 1) lie along a flow path. The time
required for water to move from one well to the next can be estimated by dividing the distance
between the wells by the flow velocity (approximately 1.5 m per day). This value was determined
directly from tracer tests, and is confirmed by calculations based on the hydraulic conductivity
of the aquifer and its hydraulic gradient (24). Water takes 1 0 wk to flow from S to Q, and 24
wk to flow from S to A. The first-order rate of biodegradation along a segment of aquifer
between the monitoring wells can be estimated by dividing the concentration in the well distal
to the spill by the concentration in the proximate well, taking the natural logarithm, then dividing
by the time required for water to flow between the wells.

Table 2 portrays the depletion of total BTX between S and Q and between S and A for the years
1 984 through 1 987. The rate constants are surprisingly consistent. A purge field was installed
and  put on  line in mid-1985 to prevent further migration of the  plume from Coast Guard
property. As soon as the purge field was put into operation, the water behind the field near site
A (Figure  la) became  stagnant,  and  concentrations of  BTX began to  drop.  Solution
concentrations of BTX dropped to low values by late 1 985 (data not shown). Because the water
was not moving, the decline in concentration overtime could be used to estimate the first-order
rate constant for anaerobic BTX biotransforma
3).
Ion in the part of the aquifer near site A (Table
                                                                                   97

-------
                                                                Figure 1a
                                                 U.S. COAST GUARD
                                                 AIR STATION BOUNDARY
                              PURGE WELL FIELD
 Figure I a. Locations of wells in an aviation gasoline plume atthe U.S. Coast Guard Air Station
           at Traverse City. Well A is located in the plume below the purge field; well B is
           located in the aerobic zone of treatment; well C is  located in the pristine region
           surrounding the plume; and wells'D, P, Q, R, and S are located in the plume above
           the purge field.
                                                                Rgure 1fo
                           TOLUENE   BENZENE
XYLENES
                   ua/L
                  30,000
                  10,000
                   3,000
                   1,000
                    300
                    100
                     30
                     10
                           P R
Figure 1 b.  Concentration of BTX in monitoring wells along a flow path down the central axis
           of the plume in the first quarter of 1 986.
98

-------
Table 2. First-Order Rates  of Anaerobic Biolransformation (per Week) of Total  BTX Along
         Segments in the Aquifer
Year
1984
1985
1986
1987
StoQ,
Quarter of Year
First
0.11
0.02
0.43
0.34

Third
0
0
17
27
oll4
olio
S to A,
Quarter of Year
First
0.10
0.07
0.20
0.33
Third
0.09
cnca
cnc
cnc
°Cannot calculate: the plume was interceptec
Table 3. Comparison of the First-Order Rates
 by a purge well field and did not reach site A.
per Week) of Anaerobic Biotransformation of BTX
Compound
Benzene
Toluene
m + p-Xylene
o-Xylene
All xylenes
Microcosms'"
0.5
0.3
0.4
0.5







Aquifer Segment
StoQ
0.05
1.3


0.03
Change Over Time at
Ac, 6/85 to 1 0/85
0.17
0.47


0.10
"Laboratory microcosm studies
bAlong flow path segments in the aquifer
cAt a site with stagnant ground water behind
a purge well field
Sediment samples for the microcosm study were acquired from site A in late 1 985, and this
study was conducted in the first quarter of 1 986. Figure 1 b depicts the concentration of BTX in
monitoring wells along a flow path down the central axis of the plume  at that time. Table 3
compares the depletion of  BTX along the aquifer segment S to Q during the first quarter of
1 986 and the depletion in  the static water a  site A in the last half of 1 985 with the  rates of
anaerobic BTX biotransformation in the microcosm  study.
The rate of disappearance of BTX compouncs
compare well with actual rates of aerobic BTX
in the field is controlled by mass-transport lim
are limited by reaction rates. The anaerobic fa1
  as measured in aerobic microcosms does not
degradation in the field. The rate of degradation
tations for oxygen (4,5), while laboratory studies
e study compared quite favorably with those rates
                                                                                    99

-------
measured by field data. Methanogenesis and other anaerobic processes are not limited by the
availability  of oxygen  in  either  microcosms  or  subsurface  materials.  Mass-transport
considerations are therefore not as critical to the comparison of microcosm  and field  data.
Anaerobic microcosms might prove to be a valuable tool to evaluate intrinsic biorestoration of
aquifers contaminated with petroleum products.
Conclusions

The  results  of the laboratory study confirm field  evidence of  both  aerobic and anaerobic
transformation  of alkylbenzenes and suggest that intrinsic aerobic and anaerobic in s/fu
biorestoration of ground water contaminated with petroleum products can occur. The anaerobic
transformations seen at this site and confirmed  by the laboratory study provide an attractive
alternative to aerobic restoration. The removals of the alkylbenzenes in the anaerobic material
were quite rapid and compared favorably with removals seen in the aerobic zone of treatment.
Comparison of first-order rates of disappearance in anaerobic microcosms with those calculated
from field data show acceptable agreement. Anaerobic processes in the subsurface are probably
limited by in situ reaction rates ratherthan by mass-transport limitations for nutrients. Potentially,
anaerobic microcosm studies could be useful in the evaluation of  intrinsic bioremediation of
petroleum-contaminated subsurface materials.

The  aerobic degradation  of alkylbenzenes in  subsurface environments  has  been well
documented (25,26)   and is  currently the  state-of-the-art for  restoration  of petroleum
contamination. Anaerobic biotransformation, however, can enhance in situ biorestoration in
oxygen-depleted regions of a plume where heavily contaminated ground water has excessive
oxygen demand. Naturally occurring anaerobic biological processes can potentially remediate
ground water contaminated with petroleum products and significantly increase the reliability of
existing  remediation technologies.
References

1.     Coleman, W.E., J.W. Munch, R.P. Streicher, H.P. Ringhand, and F.C. Kopfler. 1 984.
       The identification and measurement of components in gasoline, kerosene, and no. 2
       fuel oil that partition into the  aqueous phase after mixing. Arch.  Environ. Contam.
       Toxicol. 13:171-178.

2.     Atlas, R.M. 1 981. Microbial degradation of petroleum hydrocarbons: An environmental
       perspective. Microbiol. Rev. 45(1 ):1 80-209.

3.     Gibson, D.T.,  and  V.  Subramanian. 1984. Microbial degradation  of aromatic
       hydrocarbons. In: D.T. Gibson, ed. Microbial degradation of organic compounds. New
       York: Marcel Dekker. pp. 181-252.

4.     Borden, R.C., and P.B. Bedient. 1 986. Transport of dissolved hydrocarbons influenced
       by  oxygen-limited biodegradation, 1. Theoretical development. Water  Resour. Res.
       22(131:1,973-1,982.
100

-------
5.     Borden, R.C., P.B. Bedient, M.D. Lee,
       dissolved  hydrocarbons  influenced
       application. Water Resour. Res. 22(13)
9.
10.
11.
13.
14.
       of crude oil  in  a shallow aquifer,
       geochemical  facies. In:  Mallard, G.
                                         C.H. Ward, and J.T. Wilson. 1986. Transport of
                                          by oxygen-limited  biodegradation,  2.  Field
                                          :1,983-1,990.
6.     Baedecker, M.J., D.I. Siegel, P. Bennett, and I.M. Cozzarelli. 1 988. The fate and effects
                                          1. The distribution of chemical species and
                                          :., and  S.E. Ragone,  eds. Proceedings of the
       Technical Meeting of the U.S. Geological Survey Toxic Substances Hydrology Program,
       Phoenix, AZ (September 26-30). Water Res. Invest. Rep. 88-4220. pp. 13-20.

7.     Cozzarelli,  I.M., R.P. Eganhouse, and M.J.  Baedecker. 1 988. The fate and effects of
       crude oil in a shallow aquifer, II. Evide nee of anaerobic degradation of monoaromatic
       hydrocarbons. In: Mallard, G.E., and S.E. Ragone, eds. Proceedings  of the Technical
       Meeting of the U.S. Geological Survey Toxic Substances Hydrology Program, Phoenix,
       AZ (September 26-30, 1988), Water

8.     Kuhn, E.P., PJ. Colberg, J.L. Schnoor,
                                         Res. Invest. Rep. 88-4220. pp. 21-33.

                                         0. Wanner, A.J.B. Zehnder, and R.P. Schwarzen-
       bach. 1 985. Microbial transformations of substituted benzenes during infiltration of river
       water  to   ground  water:  Laboratory  column  studies.  Environ.  Sci.  Technol.
       19(10):961-968.

       Lovley, D.R., M.J. Baedecker,  D.J. Lonergan, I.M. Cozzarelli, E.J.P. Phillips, and  D.I.
       Siegei. 1 989. Oxidation of aromatic contaminants coupled to microbial iron reduction.
       Nature 339:297-300.

       Lovley, D.R. and D.J. Lonergan. 1993. Anaerobic oxidation of toluene, phenol, and
       p-cresol by the dissimilatory iron-reducing organism, GS-1 5. Appl. Environ. Microbiol.
       56(6):1,858-1,864.
       Suflita, J.M., S.A. Gibson, and R.E. B
       pollutant chemicals in aquifers. J. Ind
Beman. 1 988. Anaerobic biotransformations of
 Microbiol. 3:179-194.

 Rees. 1986. Biotransformations of selected
12.    Wilson, B.H., G.B.  Smith, and  IF
       alkylbenzenes and  halogenated  ali hatic hydrocarbons  in  methanogenic aquifer
       material: A microcosm study. Environ

       Reinhard, M., N.L. Goodman, and J.
       organic  chemicals  in  two  landfi
       18(12):953-961.
 Sci. Technol. 20(10):997-1,002.

  Barker. 1 984. Occurrence and distribution of
  leachate  plumes.  Environ.  Sci.  Technol.
       Ehrlich, G.G.,  D.F. Goerlitz, E.M. Godsy/and M.F.  Hult. 1982.  Degradation of
       phenolic contaminants  in  ground water  by anaerobic  bacteria:  St. Louis Park,
       Minnesota. Ground Water 20(6):703-710.
1 5.    Goerlitz, D.F., D.E. Troutman, E.M. Gpdsy, and B.J. Franks. 1 985. Migration of wood-
       preserving chemicals in contaminated
       Florida. Environ. Sci. Technol. 19(10
                                         ground water from a sand aquifer in Pensacola,
                                         =955-961.
                                                                                  101

-------
16.    Kampbell, D.H., J.T. Wilson, and D.W. Ostendorf. 1990. Simplified soil gas sensing
       techniques for plume mapping and remediation monitoring. In: Kostecki, P.T., and EJ.
       Calabrese, eds. Petroleum  Contaminated Soils.  Chelsea, Ml: Lewis Publishers, pp.
       125-139.

1 7.    Ostendorf, D.W. 1 990. Long-term fate and transport of immiscible aviation gasoline in
       the subsurface environment. Water Sci. Technol. 22:37-44.

18.    Rifai, M.S.,  P.B.  Bedient,  J.T. Wilson, K.M.  Miller,  and J.M. Armstrong.  1988.
       Biodegradation modeling at aviation fuel spill site. J. Environ. Eng. 114(5): 1,007-1,029.

1 9.    Grbic-Galic, D. 1 989. Microbial degradation of homocyclic and heterocyclic aromatic
       hydrocarbons under anaerobic conditions. Develop. Ind. Microbiol. 30:237-253.

20.    Grbi<5-Galic, D., and T.M. Vogel.  1 987. Transformation of toluene and benzene by
       mixed methanogenic cultures. Appl. Environ.  Microbiol.  53(2):254-260.

21.    Wilson, B.H., J.T. Wilson, D.H. Kampbell, B.E. Bledsoe, and J.M. Armstrong. 1990.
       Biotransformation of monoaromatic and chlorinated  hydrocarbons at  an aviation
       gasoline spill site. Geomicrobiol. J. 8:225-240.

22.    Barrio-Lage, G.,  F.Z. Parsons, R.S. Nassar, and  P.A.  Lorenzo. 1986. Sequential
       dehalogenation of chlorinated ethenes. Environ. Sci. Technol. 20(l):96-99.

23.    Klecka,  G.M., SJ. Gonsior, and D.A. Markham.  1990.  Biological transformations of
       1,1,1-trichloroethane in subsurface soils and ground water. Environ. Toxicol. Chem.
       9:1,437-1,451.

24.    Sammons, J. 1994. Personal communication between John Sammons, The Traverse
       Group,  Inc., and the authors.

25.    Lee, M.D., J.M. Thomas, R.C. Borden, P.B. Bedient, C.H. Ward, and J.T. Wilson. 1 988.
       Biorestoration of aquifers contaminated  with organic compounds. CRC  Crit. Rev.
       Environ. Control 1 8:29-89.

26.    Raymond, R.L. 1 974. Reclamation of hydrocarbon contaminated waters. U.S. Patent
       Office 3,846,290.
102

-------
Mathematical  Modeling of Intrinsic Bioremediation at Field Sites
Hanadi S. Rifai
Energy and Environmental Systems Institute, R
        ce University, Houston, TX
Introduction
Intrinsic bioremediation is  an important attenuation mechanism at contaminated  field sites
because it  limits  pollutant migration  and  reduces contaminant  mass  in  the subsurface.
Quantifying the impact  of intrinsic bioremediation at a  field  site involves  conducting a
comprehensive field sampling program in conjunction with extensive modeling of contaminant
transport and fate. This  paper  reviews the basic data requirements for modeling  intrinsic
bioremediation at field sites and discusses some of the available  models that can be used for
that purpose. The limitations of existing models for simulating biodegradation are presented, as
well as  the difficulties in validating and verify ng these models. Finally, a  few case  studies of
modeling intrinsic  bioremediation are reviewed.
Intrinsic Bioremediation  Processes
Intrinsic bioremediation refers to the reduction  of contaminant mass at a field site due to
biodegradation. This reduction can occur under aerobic or anaerobic conditions. When oxygen
is  utilized as the electron acceptor,  the process is referred to as aerobic respiration. When
oxygen is not  present (anox/c conditions),  microorganisms can use organic chemicals  or
inorganic onions as  alternate  electron  acceptors. Anaerobic biodegradation  refers  to
fermentative, denitrifying, iron-reducing, su/f^te-rec/uc/ng, or  mefhanogen/c  processes. To
quantify the  impact of intrinsic bioremediation on contaminant concentrations at a field site, one
needs to develop an accurate picture of the c istribution of electron acceptors both  in pristine
and contaminated areas. The concentrations of the electron acceptor in pristine areas provide
an indication of the biodegradation potential at the site. The  disappearance or decline of
electron acceptor concentrations
biodegradation may be occurring.
in
contaminated  areas  provides  an  indication that
Existing Biodegradation  Models

The biodegradation of contaminants in ground water is  mainly controlled by the rate of the
reaction and the availability of the electron acceptor. A mathematical expression that represents
the chemical reaction can be written to account for the effect that the rate of the reaction has
on biodegradation. This mathematical expression can then be combined with the  transport
equation to account for the electron acceptor limitation effect on the biodegradation process
in the subsurface.
Many biodegradation models have been
kinetic expression for biodegradation (see Tc
number of aerobic and anaerobic biodegrad
    developed in recent years, most of which utilize some
        ble 1). The models listed  in Table 1 simulate a
       altion processes subjectto specified conditions and
 Symposium on Intrinsic Bioremediation of Ground Water
                                                                                     103

-------
 assumptions. The difficulties involved in applying these models include 1) the data required as
 input to the model is lacking, such as the kinetic rate parameters or estimates of the hydraulic
 conductivity; 2) the majority of these models are proprietary and very few are public domain
 models; and 3) modeling in general.is complicated and time-consuming, and requires a certain
 level of expertise.
 Table 1.  Biodegradation Models
Name
-
BIOPLUME
-
BI01D
-
-
-
BIOPLUME II
-
BIOPLUS
ULTRA
-
-
Description
1 -D, aerobic, microcolony, Monod
1 -D, Monod
1 -D, analytical first-order
1-D, aerobic and anaerobic,
Monod
1 -D, co-metabolic, Monod
1 -D, aerobic, anaerobic, nutrient
limitations, microcolony, Monod
1-D, aerobic, co-metabolic,
multiple substrates, fermentative,
Monod
2-D, aerobic, instantaneous
2-D, Monod
2-D, aerobic, Monod
2-D, first-order
2-D, denitrification
2-D, Monod, Biofilm
Author(s)
Molzetal. (1)
Borden et al. (2)
Domenico (3)
Srinivasan and Mercer (4)
Semprini and McCarty (5)
Widdowson et al. (6)
Celia et al. (7)
Rifai et al. (8)
MacQuarrie et al. (9)
Wheeler et al. (10)
Tucker et al. (11)
Kinzelbach et al. (12)
Odencrantz et al. (1 3)
Modeling of  Intrinsic Bioremediation  at  Field Sites

Many case studies of simulating biodegradation at field sites exist in the general literature. In this
paper, four case studies are reviewed that combine field and laboratory investigation programs
with modeling studies.
104

-------
at the  Conroe  Superfund  site in Texas. The
operated as a wood-preserving facility from 1
ponds   were   composed   of  predominant
Conroe Superfund Site, Texas

Borden et al. (2) applied the first version of the BIOPLUME model to simulate biodegradation
United Creosoting Company (UCC) site was
946 to 1 972. Wastes disposed  in two unlined
y  polycyclic  aromatic  hydrocarbons  and
pentachlorophenol  (PCP). Monitoring  of  the  site has shown elevated levels  of  organic
contaminants in the soil and ground water, as well as elevated levels of chloride in the ground
water. The ground-water velocity at the UCC site is approximately 5 m/yr.

Oxygen exchange with the unsaturated zone was simulated by Borden et al. (2) as a first-order
decay in hydrocarbon concentration. The loss of hydrocarbon due to horizontal  mixing with
oxygenated ground water and resulting biodegradation was simulated by generating oxygen and
hydrocarbon distributions independently and then combining them by superposition. Simulated
oxygen  and hydrocarbon concentrations closely matched the observed values. The Conroe
Superfund site was one of the first sites to be modeled using a biodegradation expression in a
transport model.
Traverse City Site, Michigan

The Traverse City field site is a U.S. Coast Guc
in the northwestern portion of the lower penin: ula of Michigan. The ground water at the site is
contaminated  with organic chemicals from a leaking  underground storage tank. The main
contaminants at the site are benzene, toluene, and xylenes (BTX). The contaminant plume ranges
rd Air Station located in Grand Traverse County
from 1 50 ft to 400 ft wide and is about 4,000
ft long. A pumping wellfield system was installed
at the downgradient end of the dissolved plune to control offsite migration.

A modeling effort of natural attenuation at ths site was completed by Rifai et al. (8) with the
BIOPLUME II model. Modeling was performed for the  period before the pumping wells were
installed and also for the period after the wells were turned on. The data in Figure 1 show the
results of the model simulation along the center line of the plume for the period before the
wellfield was turned on. The model  predict
 concentrations at the monitoring wells reason ably well except in the vicinity of well M31. Rifai
 et al.  (8)  indicated that this  was because  he simulation did not account for anaerobic
 biodegradation, which was occurring in the interior of the plume.
ons by  Rifai et al.  (8) matched the  observed
                                                                                   105

-------
                                      a Obaervad Data
                                 	OBIOPUUME n  Model
                      M28
U26   TP4   TP3
                                                W30
                                                        M31
                                                              IN2W2
                                                              IN4W4
 Figure 1. BIOPLUME II model predictions for the Traverse City field site (8).
 Gas Plant Facility in Michigan

 Soluble hydrocarbon and dissolved oxygen (DO) were characterized in a shallow aquifer
 beneath a gas plant facility in Michigan by Chiang et al. (14). The distributions of BTX in the
 aquifer had  been  monitored in 42  wells for a  period of 3 years. The site geology  is
 characterized as  a  medium to coarse sand with interbeds of small gravel and  cobbles. The
 general direction of ground-water flow is northwesterly. The depth to water table ranges from
 10 ft to 25 ft below land surface, and the slope of the water table was estimated as 0.006.
 Based on  ground-water and soil sampling data, Chiang et al. (14) concluded that the flare pit
 was the major source of the hydrocarbons found in the aquifer,  while the slope oil tank was a
 secondary source.

 Chiang etal. (14) evaluated a first-order decay biodegradation  approach and the BIOPLUME
 II model for simulating biodegradation at the gas plant facility. Using the model and assuming
 first-order decay,  several simulations were made to match the observed benzene concentration
 distribution of 1/22/85 by setting the observed concentration distribution of 1 1/1/84 as the
 initial condition.  The variables  involved included  the distribution  of the leakage/spill  rates
 between the flare pit and the  slope oil tanks and macrodispersivities of the aquifer.

The BIOPLUME II model was used  to simulate the July 1987  data by setting the observed
concentration distribution of February 1 987 as an initial condition. The data  in Figure 2 show
the comparison between the measured and the simulated soluble BTX concentrations of July
 1 987. As can be seen from Figure 2, the model predictions for BTX were reasonable. The model
predictions for oxygen concentrations, however, were not as similar. The authors attributed the
106

-------
differences to the fact that the BIOPLUME II
for 1 ppm of benzene, whereas the actual rec

Cliffs-Dow Superfund Site
                                       model assumes a requirement of 3 ppm of oxygen
                                          uirement is in the range of 1 ppm to 3  ppm.
Ground water at the Cliffs-Dow site is contam
compounds. The aquifer sediments at the site
hydraulic  conductivity  ranges between  3.5
contaminants found at the site near the source
phenols, and naphthalene at concentrations re
                                          noted with low levels of phenolic and polycyclic
                                          consist of mostly coarse sands and gravels. The
                                          x 10"3 to 4.6  x  10"2  cm/sec. The  principal
                                           area include phenol, several methyl-substituted
                                          nging from 220/fg/Lto 860/fg/L. Based on the
analysis of samples obtained from monitoring wells, Klecka et al. (4) found that the levels of
organic contaminants are reduced to near ojr below the detection limit within a distance of
1 00 m downgradientfrom the source. Further analyses of the ground-water chemistry were used
to verify that biodegradation was occurring a
contaminants.
                                          • the site and causing the disappearance of the
The migration of organic constituents in th
                                      e ac
                                          uiferwas simulated using the BIO1 D model and
assuming  a first-order decay  expression.  Half-lives for the contaminants at the  site were
estimated  from the results of soil microcosm experiments based on the time required for 50-
percent disappearances of the parent compound. The velocity was varied over a range from 0.2
to 0.46 m/d, which is representative of the re nge of ground-water flow rates at the site.

Figure 3 illustrates the impact  of biodegradction on contaminant concentrations at the site.
Model simulations performed  using a  half-li
components were reduced by greater than 99
                                            of 2 d indicated that levels of the phenolic
                                          percent within a distance of 30 m downgradient
of the source. When the half-life was increased  by a  factor of 1 0, the concentrations were
reduced to a similar extent within 75 m. Because of the dominance of biodegradation, increases
in ground-water velocity from 0.2 to 0.46 m/d had minor effects on the level of attenuation
predicted with the model.
                                           developed overthe last 1 0 years. These models
Conclusions

A number of biodegradation models have bee
are generally similar in that they simulate the transport and biodegradation of a  number of
components in the ground water. The models differ in the  mathematical biodegradation
expressions that they use and in the numerical procedures used to solve the complicated system
of equations. Application of these models to field sites has proven to be complicated due to the
lack of biodegradation parameters that can be measured in the field for model input. As a
result, most modeling  applications  at the field scale  have resorted to first-order decay or
instantaneous  representation of the biodegradation process.
                                                                                   107

-------
                                  = 2 days

                             t1/2 = 10 days
                                  t1/2 = 20 days
 Retardation factor = 5
 Ground water velocity = 0.2 m/day
                                                            i  i  i   i  I   i  i   i  i
                                            100         150
                                             Distance (m)
                                                   (a)
               200
 250
                             t1/2 = 2 days

                             t1/2 = 10 days
                                  t1/2 = 20 days
Retardation factor = 5
Ground water velocity = 0.46 m/day
                                              J	I	I	I	I	I	I	L_l	I	I	I	I	!_
                                           100          150
                                            Distance (m)
                                                   (b)
              200
250
Figure 2. BIOID model predictions for the Cliffs-Dow Superfund site (15).
108

-------

000
000

000

000
000

000

000

000

000

000

0
0
158
63
[el
|o I
0

0

0

0

0

0
[3"
15.
LP_
[4"
1°
82
0

0
0
5
532
923
1503
0 0

0 0

0 0

0 0

0 0

0 0

0
0
fo]
[oj
5317
3421
10690
10544
lol
[OJ

0

0

0

0

0
0

0

0
0
38
0








0
0

0

0
fo

o
o"
c«

0

0

0






!2
0
Uj
[0]

0

0

0

0
0

0

0
0

0
1387
1573

0

0

0

0
0

0

0
0

0
5876
5024
1301
1398

0
fol
lol

0
0

0

0
0

0

0

0

0

0

0 0
0 0

0 0

0 0
0 0

0 0

0 0

0 0
Flare pit
0 0

0 0

000
0 0

000

000
000

000

000

000

000

000
FigureS. BIOPLUME II model  predictions fcr the gas  plant facility in Michigan  (14)—top
         number: observed data; bottom number: simulated data.
References
i.
2.
3.
4.
5.
Molz, F.J., MA Widdowson, and LD.
dynamics coupled to nutrient and oxyg
22(8):1/207-1,216.

Borden, R.C., P.B. Bedient, M.D. Lee,
dissolved  hydrocarbons  influenced
application. Water Resour. Res. 13:1,
Benefield. 1 986. Simulation of microbial growth
 n transport in porous media. Water Resour. Res.
C.H. Ward, and J.T. Wilson. 1 986. Transport of
 by oxygen-limited  biodegradation,  2.  Field
983-1,990.
Domenico, PA 1 987. An analytical madel for multidimensional transport of a decaying
contaminant species. J. Hydrol. 91:49-58.
Srinivasan, P., and J.W. Mercer. 1 988. Simulation of biodegradation and sorption
processes in ground water. Ground Water 26(4):475-487.

Semprini, L., and P.L. McCarty. 1 991. Comparison between model simulations and field
results for in situ  biorestoration  of  chlorinated  aliphatics,  1.  Biostimulation  of
methanotrophic bacteria. Ground Water 29(3):365-374.
                                                                                  109

-------
 7.


 8.


 9.


 10.
 6.     Widdowson, M.A., F.J. Molz, and LD. Benefield. 1 988. A numerical transport model
       for oxygen- and nitrate-based respiration linked to substrate and nutrient availability in
       porous media. Water Resour. Res. 24(9):1,553-1,565.


       Celia, M.A., J.S. Kindred, and  I.  Herrera.  1989.  Contaminant transport and
       biodegradation, 1. A numerical model for reactive transport in porous media. Water
       Resour. Res. 25(6): 1,141-1,148.

       Rifai,  H.S.,  P.B.  Bedient, J.T.  Wilson,  K.M. Miller, and  J.M.  Armstrong.  1988.
       Biodegradation modeling at aviation fuel spill site. J. Environ. Eng. 1 14(5):1,007-1,029.

       MacQuarrie, K.T.B., E.A. Sudicky, and E.O. Frind. 1 990. Simulation of biodegradable
       organic contaminants in ground water, 1. Numerical formulation in principal directions.
       Water Resour. Res. 26(2):207-222.

       Wheeler, M.F., C.N.  Dawson, P.B. Bedient, C.Y. Chiang, R.C. Borden, and H.S. Rifai.
       1 987. Numerical simulation of microbial biodegradation of hydrocarbons in ground
       water. Proceedings of the Solving Ground Water Problems With Models Conference,
       Denver, CO (February 1 0-12). Dublin, OH: National Water Well Association (NWWA).

11.    Tucker, W.A., C.T. Huang, J.M. Bral, and R.E. Dickinson.  1986. Development and
       validation of the underground leak transport assessment model (ULTRA). Proceedings
       of Petroleum Hydrocarbons and Organic Chemicals in Ground Water:  Prevention,
       Detection, and Restoration, Houston, TX (October-November). Dublin, OH: National
       Water Well Association (NWWA). pp. 53-75.

12.    Kinzelbach, W., W. Schafer, and J. Herzer. 1 991. Numerical modeling of natural and
       enhanced denitrification processes in aquifers. Water Resour. Res. 27(6):1,1 23-1,1 35.

13.  '  Odencrantz,. J.E., A.J. Valocchi, and  B.E. Rittman. 1990. Modeling  two-dimensional
       solute  transport  with  different  biodegradation  kinetics. Proceedings of Petroleum
       Hydrocarbons and Organic Chemicals in Ground Water: Prevention, Detection, and
       Restoration,  Houston, TX (October-November).  Dublin,  OH: National Water Well
       Association (NWWA).

14.    Chiang, C.Y., J.P. Salanitro, E.Y. Chai, J.D.  Colthart, and C.L. Klein. 1989. Aerobic
       biodegradation of benzene, toluene, and xylene in a sandy aquifer: Data analysis and
       computer modeling. Ground Water 6:823-834.

1 5.    Klecka, G.M., J.W. Davis, D.R. Gray, and S.S. Madsen. 1 990. Natural bioremediation
       of organic contaminants in ground water: Cliffs-Dow Superfund site. Ground Water
       4:534-543.
110

-------
Biogeochemical  Processes in an Aquifer Contaminated by Crude Oil:
An Overview of Studies at the Bemidji, Minnesota, Research  Site
Robert P. Eganhouse, Mary Jo Baedecker, an
U.S. Geological Survey, Reston, VA
  Isabelle M. Cozzarelli
Abstract
Crude oil inadvertently released from a pipeline in a remote area of north-central Minnesota has
altered the geochemistry of a shallow aquifer. Part of the oil was sprayed over a  large area to
the west of the pipeline, and another portion accumulated in an oil body that now floats on the
water table to the east of the point of discharge. Dissolution of oil components into the ground
water and microbial degradation of the oil have resulted in the formation of distinct geochemical
zones in which a variety of natural biogeochemical reactions can  be observed. Upgradientfrom
the oil body in the "spray zone," concentrations of total dissolved organic carbon  (TDOC), Ca,
Mg,  and HCO3" are greater and pH is lower than measurements observed in native ground
water. These differences reflect the transport of oil  constituents to the water table by recharge,
oxidation  under  aerobic   conditions,  and  dissolution  of  carbonates.  Beneath and just
downgradient from the oil  body, oxygen is depleted, and anaerobic degradation reactions (Fe
and  Mn reduction, methanogenesis) dominate. This is evidenced by increased concentrations
of Fe2+, Mn2+, CH4, and TDOC in ground water. Volatile hydrocarbons, mainly benzene and
its alkylated  derivatives, represent 26 percent of the TDOC in this zone. Microbially mediated
removal of these compounds within the anoxia zone is indicated  by the presence of structurally
related  oxygenated  intermediates (for  example, alkylated  benzenecarboxylic acids)  and
differences in the removal rates of isomeric a Iky I benzenes. Downgradient from the anoxic zone,
mixing of oxygenated water with plume constituents leads to removal of iron (via  precipitation)
and virtually all of the oil-derived organic constituents. Within a distance of 200 m downgradient
from the oil  body, the geochemistry of the ground  water is virtually indistinguishable from that
of native  ground  water.  Data  collected over an 8-yr period demonstrate that while the
contaminant plume has become increasingly reducing in character, its size has not changed
significantly. This attests to the efficiency of natural processes in removing/attenuating the oil-
derived contaminants of primary concern attiis site.
Introduction

Contamination of ground water by intentional or inadvertent releases of crude oil or refined
petroleum products is a widespread problem, and a great deal of effort (and money) is presently
being devoted to  remediation efforts. The efficacy of current engineering approaches is subject
to considerable debate. One thing is clear,  however. For future remediation efforts  to be
effective and useful, we must improve  our understanding of how nature responds to such
impacts. In principle, this should permit the manipulation of natural processes for purposes of
contaminant removal. Investigations that have
oil  spill site by the U.S. Geological Survey
understanding. The present paper discusses s
seen carried out at the Bemidji, Minnesota, crude
since 1 984 are  aimed at developing such an
ome of the major findings of those studies.
Symposium on Intrinsic Bioremediation of Ground Water
                                         111

-------
 Methods

 Sampling and Analysis

 Wells were installed at the study site using a hollow stem auger without lubricants or grease on
 the equipment (1). Augers for the drill rig and stainless steel well screens were steam cleaned
 prior to  use. The polyvinyl chloride casings were 5  cm in diameter. Two types of wells were
 installed. Watertable wells were emplaced so that stainless steel screens (1.5 m long) intersected
 the water table at the  approximate midpoint of the  screen;  locations are shown in  Figure 1.
 Deeper wells were installed below the water table with screen  lengths of 0.15 m or 0.61 m
 (Figure 2).

 Samples of water, sediment, and oil  were collected and subjected to a variety of chemical
 analyses. Water samples to be used for determination of inorganic constituents and  dissolved
 organic carbon were collected with submersible pumps, whereas a Teflon bailer was used to
 collect water for determination  of organic constituents. The analyses included 1)  water—
 dissolved oxygen (DO), volatile dissolved organic carbon (VDOC), nonvolatile dissolved organic
 carbon (NVDOC), methane, pH,  Eh, major cations (Ca, Mg, Na, K), alkalinity, NH4+, NO3", Ch,
 SO/', sulfide, iron  (total 6  Fe2+), Mn2+, Si, Ba, Al, Sr, <513Cmethane, c513CTIC(TIC=total dissolved
 inorganic carbon), volatile hydrocarbons (VHCs), extractable hydrocarbons, XAD resin isolates,
 and  low-molecular-weight organic  acids; 2)  oil—<313C,  elemental  analysis  (C,H,N,S),
 hydrocarbons, 13C-NMR, and heavy metals;  and  3) sediment—(513C, hydrocarbons, and
 elemental analysis (C,H,N,S). Details of the methods of sample collection and analysis are given
 by Eganhouse et al. (2,3) for hydrocarbons; by Leenheer and Huffman (4), Huffman and Stuber
 (5), and Thorn and Aiken (6) forXAD resin isolates; by Baedeckerand Cozzarelli (7), Baedecker
 et al. (8), and Bennett  et al. (9) for inorganic constituents, methane, and elemental analyses;
 and by Cozzarelli et al. (10, 11)  for low-molecular-weight organic acids.

 Site Description

 The study site is in north central Minnesota near the town of Bemidji (Figure 1). The aquifer is
 a pitted and dissected glacial outwash underlain by a poorly permeable till at about 24 m below
 land  surface. The  outwash  sediments  are heterogeneous  and composed  of moderately
 calcareous (6 percent carbonates), moderately to poorly sorted sands consisting primarily of
 quartz and feldspar of fine-to-medium  grain size (1, 12). Coring studies  have revealed that the
 sands are variably interbedded with gravel deposits and clay lenses. The water table is 6 m to
 10m below land surface, and ground-water flow is to the east-northeast (Figure 1), discharging
 into an unnamed lake approximately 300 m downgradient of the spill site. Estimates of the flow
 velocities nearthe watertable range from 0.05 meters/day (m/d) to 0.5 m/d (9) forfine-grained
 and coarse-grained sediment, respectively.
112

-------
                                                Sampling Sites
                                              • Ground water
                                              O Sediment
                                              X Oil
                                              X Oil and sediment
                                              X Oil and ground water
Figure 1.   Map of the field site near Bemidji  showing  location of the ruptured pipeline,
           approximate location of the oil bodv, area over which oil was sprayed, and sampling
           locations (2).
A pipeline rupture occurred in August 1 979, spilling 1,670 m3 of crude oil. A portion of the 410
m3 of crude oil unaccounted for after the clea
the water table. The oil body described  here is
interval of unconsolidated sediment above the
iup effort is present in a body of oil floating on
irregularly distributed over a 7-m to 8-m vertical
water table; by 1 990, it had spread to a  length
of 70 m to 80 m in the direction of ground-wa
received oil spray during the pipeline rupture,
This area, extending 140 m to 1 80 m to the w
approximately 6,500 m2, is hereafter referrec
erflow (13). An area upgradientof this oil body
and crude oil coated only the surface sediment.
est-southwest of the pipeline and encompassing
to as the "spray area" (see Figure 1).
                                                                                     113

-------
Results and Discussion

Effects of the Crude Oil on Ground-Water Geochemistry

Formation of Geochemical Zones. The inadvertent introduction of crude oil to this aquifer has
resulted in marked alteration of geochemical conditions in the ground water. The driving force
forthese changes has been the microbial degradation of metabolizable organic matter, in this
case represented by soluble constituents of the oil. After several years of research during the mid
to late 1 980s, it became evident that the saturated zone of this aquifer could be characterized
by geochemically distinct zones. These zones, depicted in a cross section centered on the main
contaminant plume in  the  direction  of  ground-water flow  (Figure 2), correspond to:
I—uncontaminated native ground water, II—ground water upgradient of the oil body but within
the "spray area," HI—anoxic ground water immediately beneath and downgradient of the oil
body, IV—suboxic transitional zone where anoxic ground water from zone III mixes with more
oxygenated  ground water further downgradient,  and  V—oxygenated ground  water  that
increasingly approaches conditions of the native ground water. The geochemistry of the ground
water in each zone reflects that of the water upgradient from it and processes (e.g., sorption,
dilution, degradation, dissolution, gas exchange) occurring within that zone.
         IB



         f
           435,-
           430
           425
            420
            415
                   Water table
                                             Oil body
((  ((
                 -200
                         -150
                                -100     -50      0      50
                                 Distance from center of oil body, meters
           100
                  150
                          200
Figure 2.   Cross section of the aquifer near Bemidji along main sampling transect. Locations
           of water table and deep wells indicated as filled bars (2). Fora detailed description
           of zonation, see Baedecker et al. (8).
The native ground water (zone I) is a dilute Ca-Mg-HCO3" water with total dissolved solutes
<400 mg/L, a Ca:Mg ratio of 2.2:1, a median DO concentration of 7.68 mg/L, and a  pH of
7.6 to  7.8. TDOC concentrations are  approximately  2 mg/L to 3 mg/L.  The chemical
composition is controlled by carbonate equilibria and, to a lesser extent, by dissolution of quartz,
feldspar, and clay minerals and by the degradation of naturally occurring organic material (8).
114

-------
Ground water within zone II is influenced by th
quantities were insufficient, however, for forma
e aerobic degradation of soluble oil constituents
originally deposited at land surface by the spraying  of oil from the ruptured high pressure
pipeline. In this area, oil is known to have penetrated  the upper 4 cm to 6 cm of the soil. The
ion of a discrete oil phase. Ground water in zone
II exhibits elevated concentrations of TDOC, Ga, Mg and HCO3"; lower pH (about 0.5 to 1 pH
unit);  and  a slightly reduced DO concentration.  Carbon dioxide  partial  pressures (PCO2)
calculated  by Bennett  et al. (9)  are as  high  as TO"1'5 atm  (as  compared with ambient
atmospheric values of 10"3-5atm). The TDOC concentrations are nearly an order of magnitude
greater than those found in the  native ground water. While the additional TDOC in zone II is
undoubtedly derived from the oil, virtually all of it represents partially oxidized transformation
products, not hydrocarbons (2).  No VHCs were detected. The few petroleum hydrocarbons that
are present have compositions that indicate ex ensive biodegradation. Presumably, some of the
degradation/oxidation  occurred in the  unsakirated zone followed  by  transport of soluble
transformation products to the water table during recharge. The slight depression  of oxygen
concentration in ground water  in this zone indicates that some degradation  is likely to be
occurring within  the ground water as  well.-If rremoval of the NVDOC is occurring via aerobic
respiration, however, the rate of supply of NVDOC must  exceed its rate of removal because
NVDOC concentrations tend to  increase with approach to zone 111. Bennett and others (9)  have
suggested that the large increases in Ca, MgJ and HCO3" with increasing PCO2/ as we'l as the
similarity of the [Ca  +Mg]/[HCO3-] mole ratio of ground water in this zone to that of the native
ground water, indicate thatthe principal reaction responsible forgenerating the plumes of alkali-
earth  solutes is  dissolution  of  carbonates By  reaction with carbonic acid. Because these
processes are occurring under oxic (as opposed to anoxic) conditions, the primary chemical
signatures of the contaminant plumes reflect tie products of organic remineralization (HCO3~)
and  transformation (TDOC) and those result
reactions.
ing  from acidification (pH, Ca, Mg), not redox
Zone III encompasses ground water near, dovmgradientfrom (about 75 m), and in immediate
contact with the oil to a depth of about 3 m below the water table. Dramatic changes in
ground-water  geochemistry  result from  the dissolution of soluble oil constituents and  their
metabolism, leading  to  complete consumption of oxygen  and the dominance  of  anoxic
degradation reactions, including iron and manganese reduction and  methanogenesis.  Sulfate
reduction and denitrification  are  not impo-tant in this system because of the very  low
concentrations of  SO42"  and NO3" found in  the native ground water.  In this  zone,  TDOC
concentrations rise to a maximum of 48 mg/., and a significant fraction of this  is VDOC (42
percent). The volatile compounds are dominated (approximately 63  percent)  by a  mixture of
saturated, aromatic,  and alicyclic hydrocarbons derived from the  oil, the  most important
constituents  of which  are benzene  and a complex assemblage of alkylbenzenes. Methane is
found in  the  ground  water, and  field  measurements of Eh  indicate  a  strongly reducing
environment. In 1 987 the average stable carbon isotope ratios forTDOC ((513CT|C) and methane
((513Cmethane) were -8.23 and -55.45 per mil,
native ground water is -12.55 per mil. The
 respectively. The average 513C value for TIC in
 heavier ratio for TIC in zone III, thus, reflects
fractionation resulting from methanogenesis. The most dramatic changes in inorganic chemistry
are seen as large increases in the concentrati
 >ns of dissolved iron and manganese that result
from mobilization of these redox species in response to the microbially mediated oxidation of
hydrocarbons.  Silica concentrations also increase due to  enhanced dissolution  of  silicate
minerals. Other compounds present in the ground water of zone III, but not found in the oil or
(except in trace amounts) ground water in zor es I and II, are a complex mixture of oxygenated
products of hydrocarbon degradation, including low-molecular-weight organic acids (10, 1 1]
The acids are structurally related to coexisting
monoaromatic hydrocarbons from the oil. Those
                                                                                    115

-------
organic  acids  that  have  been  identified correspond to aromatic  hydrocarbons  whose
concentrations  decrease most rapidly within the anoxic zone, whereas no  potential  acid
intermediates were found for aromatics which appear to be more stable within  zone III. These
results signal the partial oxidation of hydrocarbons to more soluble metabolites.

Zone IV is a transition zone characterized by small but.detectable quantities of oxygen. Dissolved
iron and silica  decrease to below detection limits at the  boundary of this zone due  to
precipitation reactions. Low-molecular-weight organic acids are at or below detection limits, and
the concentrations of all oil-derived  hydrocarbons are much lower than  in the ground water of
zone III. Ca, Mg, and Sr also decrease but rather gradually. Bennett and others (9) have
hypothesized that this  is the result of dispersive mixing and that transport of these constituents
is conservative.

Evolution of the Contaminant Plume. Figure 3 depicts the concentrations of dissolved Fe2+,
Mn2+, methane, and <513CnC from ground water taken from zone III nearthe downgradient edge
of the oil body for the years 1 984 to 1 992 (8, 14). At this site, concentrations of methane and
Fe2+ increased  by factors of 100 and 25, respectively, during the first 5 years. Thereafter, the
concentrations  of methane and Fe2+ have  virtually leveled off. Over the  same  period,
manganese concentrations first increased and then  declined, whereas  (513CTK;  has increased
continuously. These variations in the contaminant plume chemistry reflect evolutionary changes
in the biology and geochemistry of this perturbed system with continued supply of hydrocarbons
and depletion of terminal electron acceptors.
                     i	1	1	T	T	1
        0.001
                  1984       1986      1988      1990       1992
Figure 3.  Concentrations of dissolved ferrous iron, manganese, methane (in millimoles), and
          <513C in per mil of TDOC for the years 1984 to 1992 (14).
116

-------
In particular, it would appear that manganese
•eduction was an important degradative process
at the  earliest stages of plume evolution, b jt as time has progressed iron  reduction and
methanogenesis have become dominant, presumably due to depletion of the supply of reducible
manganese. As of 1 992, the dominant process at this location appears to be methanogenesis.
This is supported by the steady increase in (513OT|C . During this same period, the distribution and
size  of the contaminant plume, as  measured  by the  concentrations of  monoaromatic
hydrocarbons  (2),  have not changed  appreciably. While our understanding of the factors
controlling plume evolution remains limited and the size and shape of the contaminant plume
itself has not changed, it is clear that the bioc eochemistry within the plume is dynamic.

Transport and Fate of Oil-Derived Contaminants

The following discussion considers the transport and  fate of the monoaromatic hydrocarbons
within zones 111 and  IV. Benzene is the dominan  individual VHC in anoxic ground water near the
leading edge of the oil body (70  percent), reacning concentrations in excess of 9,300 ^ag/L. The
C, .4 alkylated benzenes are second in abundarce, representing approximately 14 percent of the
total VHCs. Alkanes, consisting of a complex m xture of normal and branched hydrocarbons with
four to seven carbon atoms, account for T2 percent of the VHCs.  Cyclic hydrocarbons
(cyclopentanes and cyclohexanes), cycloaromatics (e.g., alkylated indans), and heteroatomic
species (e.g., tetrahydrothiophenes) are relatively minor constituents (total less than 4 percent).
Thirty-one  meters  farther downgradient from this site (but still  within zone III), the  VHC
composition is markedly different. The  relative  abundance of alkanes (and toluene; see Figure
4) is reduced, and benzene plus monoaromatic: hydrocarbons with two to four a I kyl carbon atom
substituents dominate (96 percent). Benzene represents 90 percent of the total VHCs. These
compositional differences are also found among individual C^ (alkylated) benzenes. There are
four isomeric C2-benzenes, eight isomeric C3|benzenes, and 22 isomeric C4-benzenes. All of
these compounds,  with the exception of t-butylpenzene, which was not detected in the crude oil,
are  present in the  contaminated  ground  water within zone 111. The  composition of the
monoaromatic hydrocarbons changes systematically with distance downgradient from the oil
body and with increasing depth in the saturated zone (15, 16).  Most importantly,  isomeric
alkylbenzenes show dramatically different ap
parent removal rates downgradient from the oil
 body. Because these isomers have similar physical properties, the attenuation of VHCs (and
therefore VDOC) is attributable to biological,
 rather than physical, processes.
                                                                                   117

-------
               1
               §
               1
               u
               u
                    10s
                    104
105
                    102
                   10°
                   10''
                   io-2
                             Toluene
                     20     40     60      80    100    120    140
                             Distance from Center of Oil Body (meters)
                                                  160
Figure 4.  Concentrations of benzene, toluene, and ethylbenzene (ag/L) along the flowpath of
           the contaminant plume. Also shown is the concentration of a conservative solute with
           an assumed starting concentration of 10,000 fig/L. Model assumptions discussed
           in Baedecker et al. (8); figure from  Mallard and Baedecker (1 7).
This hypothesis is reinforced by data illustrated in Figure 4. Here concentrations of benzene,
toluene, ethylbenzene, and a conservative tracer (see Baedecker et al. [8] for discussion) are
shown as a function of distance downgradient from the center of the oil body. Two features of
these data are readily apparent. First, the concentration changes observed for these aromatic
hydrocarbons greatly exceed what would be predicted on the basis of conservative transport.
Second, because  benzene, toluene, and ethylbenzene  form an homologous series with
increasing Kows (octanol-water partition coefficients), one would expect that a systematic pattern
of removal rates would be found if sorption was the dominant process. This obviously is not the
case, as toluene is very rapidly removed by comparison with either benzene or ethylbenzene.
Clearly,  biodegradation is the  most important  process  limiting the  transport  of  these
hydrocarbons.
Conclusions

At the Bemidji field site, the introduction of crude oil to the subsurface has resulted in dramatic
changes in  the  geochemistry of  the  ground  water.  Even  so,  natural  processes,  with
biodegradation being  most important,  have effectively  limited the  transport of  oil-derived
contaminants to a distance of 200 m from the source. Continued monitoring of the contaminant
plume has revealed the dynamic nature of the coupled biological and geochemical processes
operative at this site. An understanding  of these  processes and the factors that affect plume
evolution will be essential if we are to develop environmentally responsible remediation methods
in the future.
118

-------
References
i.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Hult,  M.F.  1984.  Ground-water  contamination  by  crude oil  at  the  Bemidji,
Minnesota, research site: An introduction. In: Proceedings of the  U.S. Geological
Survey toxic-waste technical meeting, Tucson, AZ (March 20-22). Water Res. Invest.
Rep. 94-4188. pp. 1-15.
Eganhouse, R.P., M.J. Baedecker,
Dorsey. 1993.  Crude  oil  in  a s
          geochemistry. Appl. Geochem. 8:551-567.
 M. Cozzarelli, G.R. Aiken, K.A. Thorn, and T.F.
 hallow sand and  gravel  aquifer,  II. Organic
Eganhouse, R.P., T. Dorsey, C. Ph
C6-C10  aromatic hydrocarbons
chromatography. J. Chromatog. 6

Leenheer, J., and E.W.D. Huffman
 nney, and S. Westcott. 1 993. Determination of
 in  water  by  purge-and-trap  capillary  gas
 28:81-92.

  1 979. Analytical method for dissolved organic
carbon fractionation. Water Res. Invest. Rep. No. 79-4.

Huffman, E.W.D., and H.A. Stube'. 1985. Analytical methodology for elemental
analysis of humic substances. In: Aiken, G.R., D.M. McKnight, R.L. Wershaw, and
P. MacCarthy, eds. Humic substances in soil, sediment and water. New York, NY:
John Wiley and  Sons. pp. 433-456.

Thorn, K.A., and G.R. Aiken.  1 989. Characterization of nonvolatile organic acids
resulting from the  biodegradation of crude oil by nuclear  magnetic  resonance
spectrometry.  In: Mallard, G.E., and S.E. Ragone, eds. Proceedings of the technical
meeting of the  U.S. Geological jSurvey  Toxic Substances  Hydrology  Program,
Phoenix, AZ (September 26-30, 1 988). pp. 41-52.
Baedecker, M.J., and I.M. Cozzare
constituents in contaminated ground
Ground-water quality and analysis
Dekker. pp. 425-461.
illi. 1 992. The determination and fate of unstable
   water. In: Lesage, S., and  R.E. Jackson, eds.
 at hazardous waste sites. New York, NY: Marcel
Baedecker, M.J., I.M. Cozzarelli, D.I.  Siegel, P.C. Bennett, and R.P. Eganhouse.
1 993. Crude oil in a shallow sand and  gravel aquifer, III. Biogeochemical reactions
and mass balance modeling in anoxic ground water. Appl. Geochem. 8:569-586.

Bennett, P.C. D.I. Siegel, M.J. Baedecker, I.M. Cozzarelli, and M.F. Hult. 1 993.
Crude oil  in a shallow  sand and |gravel aquifer, I. Hydrogeology and  inorganic
geochemistry. Appl.  Geochem. 8:529-549.

Cozzarelli, I.M., M.J. Baedecker, R.P.  Eganhouse, and D.F. Goerlitz. 1994. The
geochemical evolution  of low-molecular-weight organic acids derived from the
degradation of petroleum contaminants in ground water. Geochim. Cosmochim.
Acta 58:863-877.
                                                                                 119

-------
n.
12.
13.
14.
15.
16.
17.
Cozzarelli, I.M., R.P.  Eganhouse, and M.J. Baedecker.  1 990. Transformation of
monoaromatic hydrocarbons to organic acids in anoxic ground-water environment.
Environ. Geol. Water Sci. 16:135-141.

Franzi, D.A. 1 985. The surficial and subsurface distribution of aquifer sediments at
the Bemidji research site, Bemidji, Minnesota. Presented at the Second Toxic-Waste
Technical Meeting, Cape Cod, MA (October 21 -25).

Essaid, H.I., W.N.  Herkelrath, and K.M. Hess. 1 991. Air, oil, and water distributions
at a crude-oil spill site,  Bemidji, Minnesota. In: Mallard, G.E., and  D.A. Aronson,
eds.  Proceedings  of the technical  meeting of the U.S. Geological Survey Toxic
Substances Hydrology Program, Monterey, CA (March 11-15). pp.  614-620.

Baedecker, M.J.,  and  I.M. Cozzarelli. 1994. Biogeochemical  processes  and
migration of aqueous constituents in ground water contaminated with crude oil. In:
Button, A.R., ed. Toxic substances and the  hydrologic sciences. Minneapolis, MN:
American Institute  of Hydrology, pp. 69-70.

Cozzarelli, I.M., R.P. Eganhouse, and M.J.  Baedecker. 1 989. The fate and effects
of crude oil  in a shallow aquifer, II.  Evidence of  anaerobic degradation of
monoaromatic  hydrocarbons. In:  Mallard,  G.E.,  and S.E.  Ragone,  eds.  In:
Proceedings of the technical meeting of the U.S. Geological Survey Toxic Substances
Hydrology Program, Phoenix, AZ (September 26-30, 1 988). Water Res. Invest. Rep.
88-4220.  pp. 21-34.

Eganhouse, R.P., T.F. Dorsey,  C.S. Phinney, M.J. Baedecker, and I.M.  Cozzarelli.
1 987. Fate of monoaromatic hydrocarbons in an oil-contaminated aquifer: Evidence
for the importance of microbial activity. In: Proceeding of the Geological Society of
America 1987 annual meeting, Phoenix, AZ. 19:652.
Mallard, G.E., and M.J. Baedecker. 1 993. Hydrocarbon transport and degradation
in ground  water:  U.S.  Geological  Survey investigations.  In: Pare, K.M., ed.
Proceedings of the Air Combat Command environmental quality 1 993 symposium,
Langley AFB, VA (March 1 -5) pp. 102-108.
120

-------
Simulation of Flow and Transport Processes at the Bemidji, Minnesota,
Crude-Oil  Spill Site
Hedeff I. Essaid
U.S. Geological Survey, Menlo Park, CA
Abstract
The Bemidji, Minnesota, field site has provided an opportunity to study in detail the natural
processes that occur following a crude-oil spill. Detailed field studies have characterized the
subsurface oil distribution and the characteristics  of the contaminated ground-water plume.
Numerical models that simulate flow and transport processes are useful tools for integrating
information collected in the field, fortesting hypotheses, and forstudying the relative importance
of simultaneously occurring processes in complex,  real-world systems. Numerical modeling of
multiphase flow iat the Bemidji site has  illustrated the importance of  spatial variability on the
movement and distribution of oil in the subsurface. Solute-transport modeling  that includes
aerobic and anaerobic degradation processes is being used as a tool to study the field-scale,
naturally occurring solute-transport and degrcdation processes occurring at the site.
Introduction

On August 20, 1 979, a buried ojl pipeline
(1 1,000 barrels)  of crude oil (Figure 1). The
outwash plain. Depth to the water table rangi
lear Bemidji broke, spilling about 1.7 x TO6 L
site is located  in a pitted and dissected glacial
s from 0 m to 8 m below land surface, and the
flow through the aquifer is generally horizontal and northeastward towards an unnamed lake
300 m  downgradient from the point of pipe ine rupture. An  estimated 1.2 x 1 O6 L (7,800
barrels) of the spilled oil was removed by pumping from surface pools, trenching, burning, and
excavating soil (1). The petroleum in the pipeline was under pressure, causing oil to be sprayed
over approximately 6,500 m2  when the  pipeline  broke. The oil collected in topographic
depressions and trenched areas where large volumes  of oil  infiltrated into the subsurface,
forming two main bodies of oil floating on the water table. The subsurface oil bodies  provide
a long-term, continuous source of hydrocarbor
with the flowing ground water.
 components that dissolve in and are transported
Numerical models that simulate flow and transport processes are useful tools for integrating
information collected in the field, fortesting hypotheses, and forstudying the relative importance
of simultaneously occurring processes in compl sx, real-world systems. Many researchers working
at the Bemidji site have focused considerable effort on characterizing the subsurface distribution
of the oil and the nature of the resulting contaminated ground-water plume. This paper reviews
this work, then briefly summarizes the results  of numerical simulations of multiphase flow and
of ground-water transport and biodegradation at the site.
Symposium on Intrinsic Bioremediation of Ground Water
                                         121

-------
                  95 5'40"
                                                                          95 5'4"
            47 34'32"
             47 34'14"
                                             Direction of
                                           ground-water flow
                        Trace of north pool
                         simulated section
                                              SCALE
                                            0  50  100  150 Meters
                                               I    I   I
Figure 1. The north oil pool at the Bemidji crude-oil spill site. Plus symbols are locations of
          boreholes where samples were collected forsaturation analyses, and circles are wells
          sampled for ground-water concentrations. The line A-A' is the trace of the multiphase
          flow simulation section, and B-B' is the trace of the transport simulation section.
The Observed  Subsurface  Oil  Distribution

The Bemidji field site has provided an opportunity to study in detail the subsurface oil distribution
following a spill. Field cores were collected for the purpose of determining the oil-saturation
distribution (the fraction of the  pore space that  is  occupied  by oil) in the subsurface.
Determination of the subsurface fluid-saturation distributions required the implementation of a
sampling technique that could recover relatively undisturbed core samples from the unsaturated
and saturated zones while maintaining the in situ pore-fluid distribution (2). To improve field
sample collection, a freezing-tip core  barrel was developed and used for sample collection (3).
To  allow visual inspection of the cores  in the field,  clear polycarbonate  liners, 47 mm in
diameter and 1.5 m long, were used within the core barrel.

Following retrieval, the cores were frozen and cut into 78-mm long subsamples using a circular
saw fitted  with  a  masonry  blade. The oil  saturation of each core was  determined  in the
laboratory  using  a  porous  polyethylene (PPE) technique (2,4).  In  this  process,  strips of
hydrophobic PPE are placed into a slurry created by adding water to a core sample.  The PPE
absorbs the oil from the sample but does not take up  water. The amount of oil present in the
core is calculated from the change in weight of the oily PPE strips. The sample is then dried in
an  oven, and water saturation is determined  gravimetrically. Air saturation  can  then be
calculated by subtracting the sum of oil and water saturation from unity. Following the saturation
analysis, each sample was sieved to obtain the  particle-size distribution.
122

-------
To date, samples have been collected and analyzed from 1 0 boreholes (550 subsamples) at the
site of the north oil pool (Figure 1). The three-dimensional distribution of oil saturation at the
north pool obtained by kriging the observed data is shown in Figure 2 (5, 6). The oil distribution
at this site is complex. A considerable amount of oil remains in the unsaturated zone at locations
where oil  infiltrated from a trench excavated
                                          following the spill. The maximum oil saturation
measured at this site (0.74) was located downgradient from the zone of oil infiltration. The body
of oil floating on the watertable is not lens-sha'ped. The oil-saturation contours follow a sinuous
path that is roughly parallel to the direction of ground-water flow. Thin silt lenses at the north
pool site appear to have considerable influence on the observed oil-saturation distribution.
These layers result  in high residual oil saturations in the unsaturated zone and cause the shape
of the oil body floating on the water table to be complex and irregular rather than lens-shaped.
Geostatistical and multiphase flow simulatior
difficulty of measuring sediment  hydraulic
particle-size distribution data were  used to es
                                           have been used to assess the effect of spatial
variability of hydraulic properties on the oil-saturation distribution at the site. Because of the
                                          properties  for  oil-contaminated  samples, the
                                          imate the permeabilities (k) of the field samples
and the retention curves (7). The mean and irariance of the log(k) distribution (k in m2) were
-1 1.2 and 0.25, respectively, at the north pool. Figure 3 shows a cumulative probability plot
of log(k) at the site. A linear relation indicates a lognormal distribution of k, and the slope of
the line is related to the variance of k. The north pool plot of log k appears to consist of two
linear segments, a distribution that suggests that the permeability distribution at this site consists
of two  lognormal  populations:  a  coarse
                                          fraction  (log(k)>-l 1.64)  and  a fine  fraction
                          Kriged North Pool Oil Saturations
          430
   Elevation
       (m)
           422
                                  0    Oil Saturation  .75
Figure 2. Three-dimensional oil saturation
          is at an elevation of about 423.5
                                        di stribution at the north oil pool site. The watertable
                                          n above sea level.
                                                                                    123

-------
            99.99
       ra
       •§
       .1
       3
       o
             0.01
               -15.
-14.
-13.
-12.
                                            Log (k)
-10.
-9.0
Figure 3.  Cumulative probability plot of log (permeability) at the north pool site.
To obtain a regular grid of k values needed for the multiphase flow simulations, geostatistical
simulation techniques (8) were used to generate k distributions for the north pool site that were
conditioned on the values estimated from the core samples. A permeability realization was
obtained that reproduced the geometry of the fine and coarse fractions and also reproduced
the variability structure within these fractions. The details of this process are explained by Dillard
(9) and Dillard et al. (10).
Multiphase Flow Modeling

Atwo-dimensional numerical model was developed and used to simulate multiphase flow along
a longitudinal vertical transect parallel to the direction of flow at the north oil pool. The model
solves a  mass balance equation for the oil and water phases, assuming that the air phase is
maintained at atmospheric pressure. An important feature of the model is that it incorporates
hysteretic relations between capillary pressures and fluid saturations. The  details of the model
and the approaches used are  given by Essaid et al. (7).

The model was used to simulate subsurface flow from the time of the spill in August 1 979 until
the samples were collected in  June 1990. The.simulated section was 120 m long and 10m
deep. The initial condition for the simulation was a hydrostatic water pressure  distribution
corresponding to the measured water-table elevation at the time of sampling. The lateral and
bottom boundaries were assumed to be hydrostatic water-pressure boundaries. Oil was assumed
to infiltrate through five constant oil pressure nodes at the top boundary representing the trench
124

-------
that was excavated following the spill. Oil infiltration was stopped when the cumulative oil
infiltration was equal to the estimated oil mass in the observed transect. More details of the
north pool simulation are given by Dillard
01
                                           and Dillard et al. (10).
Data  from the first sample transect (Figure 4a) were compared with the two-dimensional
simulated oil-saturation distribution. In the first simulation, a uniform mean k of 5 x 10"12 m2 and
uniform mean capillary pressure/saturation fur ction parameters were used. A uniform k resulted
in a symmetric, lens-shaped oil body (Figure 4 D) that did not reflect the features of the observed
    distribution.  In  the  second  simulation, spatial  variability of hydraulic properties was
introduced, resulting in variability in oil saturations within the lens (Figure 4c). Very little oil was
trapped in the unsaturated zone. The shape o
complex and irregular, with zones of low oil  s
silt layers.  In the third simulation,  hysteresis
                                           the oil body floating on the water table became
                                          aturation corresponding to the low permeability
                                           was introduced,  causing  more oil to  become
entrapped in the unsaturated zone (Figure 4d). The modeling results suggest that the silt layers
and spatial variability exert a strong control on the oil distribution in the subsurface.
The Observed Ground-Water  Plume

Numerous researchers working at the Bemidji
microbial activity and degradation  of petrolei
                                          :rude-oil spill site have documented evidence for
                                          m hydrocarbons in the field (1 1 -1 8). As a result
of these studies, five geochemical zones in ground water (Figure 5) have been identified at the
site (1 1,1 2,1 9). Zone 1 consists of oxygenated, uncontaminated native ground water. Zone 2,
which is belowthe area where the land surface was sprayed by oil following the pipeline rupture,
   characterized by reduced  oxygen  conce itrations  and  the presence of refractory high
is
 molecular-weight hydrocarbons. Zone 3, beneath and  immediately downgradient from the
 separate-phase oil body, consists of an anoxic plume  of ground water that  contains high
 concentrations of hydrocarbons and methane.  Zone 4 is the transition zone from anoxic
 conditions to fully oxygenated conditions, and|concentrations of hydrocarbons decrease rapidly
 as a result of aerobic degradation processes. Zone 5 consists of oxygenated water downgradient
 of  the  contamination  plume that contains slightly  elevated  concentrations  of  dissolved
 constituents. Long-term monitoring of the plume since 1 984  has shown that, near the water
 table, the concentration of total dissolved organic carbon (TDOC) and dissolved oxygen (DO)
 downgradient from the oil body has remained relatively stable with time. In the anoxic zone
 (Zone 3), concentrations of reduced manganese (Mn)  and iron (Fe) and of methane have
 increased with time,  indicating a  sequence of Mn  reduction followed by Fe reduction and
 methanogenesis.
                                                                                    125

-------
                                 430
                            Elevation
                               (m)
                                 420
                      b    430
                     Elevation
                        (m)
                           420
                      C    430
                     Elevation
                        (m)
                          420
                      d    430
                     Elevation
                        (m)
                          420
                                                  Distance (m)
100
                                                 Distance (m)
       120
                                                 Distance (m)
       120

                                                 Distance (m)
       120
                                         0      Oil Saturation      1'°
Figure 4.  Oil saturation distributions along the sample transect: a) observed oil saturations, b)
           simulated oil saturations with uniform mean properties, c) simulated oil saturation
           with spatially variable properties, d) simulated oil saturation with spatially variable
           properties and hysteresis.
126

-------
                          100
                                       RECHARGE ZONES
— < 	 > — < 	 > — *•
to 424
re
1U
8
S_ 420
LU
Q
t
x\ x i A \ ^
* V! Zone 2 \ \c Zc
Zono 1 ^ 	 ^ "•'•^
X ^^-^
EXPLANATION j* ^^
X midpoint o( well x j
screen (
Datum is soa lovol /
Oil Body ^ "^
^UUIJIU^ x x x \ x 'X x x
ne 3 x *is * \. X
X /
v Zone 4 .. /
~-~-— _Ji— -^ x
x Dicoclion ol Ground-Water Flow
I I
                                      DIST/
   200

   , IN METERS
                                                               300
                                                                                  400
Figure 5.  The simulated cross section showing the five geochemical zones in ground water.
Ground-Water Transport and  Biodegradation Modeling

A two-dimensional, multispecies solute-transport model that incorporates  biodegradation is
being developed and applied to the ground-water system at the Bemidji site. The model is being
used to quantify the field-scale degradation processes and to identify the important factors
affecting the distribution of solute species in the field. The model simulates the aerobic and
anaerobic degradation processes that have been observed in the contaminated ground-water
plume at the spill site. The U.S. Geological Survey's Method of Characteristics transport model
(20, 21) was expanded to handle multiple salutes and to include biodegradation terms. The
approach of Kindred and  Celia  (22) was used to represent the biodegradation terms in the
transport equation.  Details of the model are given by Essaid et al. (23).
A  vertical  cross section of unit  width  that is  approximately  parallel  to  the  direction of
ground-water flow was simulated using the trc ns.port model for the period from the time of the
spill in ~\979 until September 1990 (Figure 5). Ground-water samples from numerous wells
along this section have been analyzed overtims (11,12,1 5). Steady-state flow, no sorption, and
isothermal conditions were assumed. Foursolu
TDOC was split into two fractions: degradable
tes and two microbial populations were modeled.
dissolved organic carbon (DDOC) and refractory
To  represent the aerobic and anaerobic deg
populations of bacteria were included in the
dissolved organic carbon (RDOC). The remair ing two solutes modeled were DO and methane.
radation processes, aerobic and methanogenic
simulations.  Competitive inhibition was used to
represent the suppression of methanogenesis by oxygen. In this manner, as oxygen in the ground
water is consumed and an anoxic zone develops, the methanogens begin to flourish, resulting
in increased methane production. Iron and manganese reductions were not included because
of the complexity of incorporating the rock-water interactions of dissolution and precipitation into
the transport model.

-------
 For simulation  purposes, the system was  represented  by an  initially clean aquifer with
 background dissolved organic carbon concentrations and fully oxygenated water. Following the
 oil spill, it was assumed that DDOC and RDOC dissolved and entered the aquifer with recharge
 water. The estimated values of initial concentrations and recharge water concentrations for each
 solute are given in Table 1. The oil present in the pore space within the oil body reduces water
 flow through this zone. The magnitude of reduction of water flow  is a complex function  of the
 oil distribution. As a first approximation of this effect, the hydraulic conductivity and recharge
 rate in the  zone of the oil  body were reduced to  25 percent of the aquifer values.  No
 measurements or estimates  are  available for  many of the  transport and  biodegradation
 parameters under natural field conditions. Therefore, reasonable estimates of these values were
 used in the simulations. The details of the simulation parameters and boundary conditions are
 given by Essaid  et al. (23).
Table 1. Initial and Recharge Water Concentration (mg/L)"
Solute
DDOC
RDOC
TDOC
DO
Methane
Initial
Concentration
0.0
2.0
2.0
9.0
0.0
Recharge Zone
A
0.0
2.0
2.0
9.0
0.0
B
10.0
20.0
30.0
3.0
0.0
c
100.0
30.0
130.0
0.0
0.0
D
0.0
2.0
2.0
0.0
0.0
E
0.0
2.0
2.0
9.0
0.0
*Recharge zones A through E are shown in Figure 5.
Transport Simulation  Results

Observed and simulated profiles of TDOC, DO, and methane near the water table are plotted
in Figure 6. The observed and simulated concentration profile of TDOC is shown in Figure 6a.
The observed points show a concentration distribution at the water table that is relatively stable
with time. The simulation has captured this feature, as can be seen by the similarity between the
simulated 1 986 and 1 990 concentration profiles. There is an increase in TDOC concentration
in the upgradient spray zone, followed by a rapid  increase in TDOC  concentration in the zone
of the oil body. Downgradient from the oil body, the TDOC concentration decreases gradually
to the background concentration. This decrease is a result of microbial and physical processes.
There is anaerobic degradation of DDOC within the anoxic zone near the oil body and aerobic
degradation of DDOC at the margins of the plume, where  oxygenated recharge and ground
water are encountered. Also, there is dilution of TDOC  downgradient from the oil body as a
result of the physical processes of displacement and  mixing  of flowing ground water with
recharge water.
128

-------
                     TDOC
E 80
_J
DC
UJ

1 60

CD

| 40


1
5 20
     111
     o
                 100     200     300
                   DISTANCE, IN METERS
                                                             METHANE
                                              tr
                                              UJ
                                              K

                                              EC
                                              UJ
                                              Q-
                                            30
                                   400
                                              S 20
                                              Q 10
                                              I
                                         UJ
                                         o
                                               O
                                               O
                 100     200     300
                   DISTANCE, IN METERS
Figure 6.  Graphs of simulated and observed
          b) DO, c) methane.
100     200     300     400
 DISTANCE, IN METERS
                                                           EXPLANATION

                                                   — 1986 Simulated   * 1987 Observed
                                                   — 1990 Simulated   -i- 1888 Observed
                                                    A  1S86 Observed   x 1990 Observed
                                   400
                                     concentrations at the water table: a) TDOC,
The simulated DO concentration (Figure  6b)  decreases  in the spray zone because of the
assumed decrease in DO  concentration  in  recharge  water,  caused  by degradation  of
hydrocarbons in the unsaturated zone and by the consumption of oxygen by degradation in the
ground water. Nearthe oil body and immediately downgradientfrom it, the DO of the recharge
water is assumed to have been completely co isumed in the unsaturated zone. An anoxic zone
develops in this area. Farther downgradient, DO begins to increase as oxygenated recharge
water enters the system.
                                                                                   129

-------
 Methane is  produced in the anoxic  zone that develops  in  the  immediate vicinity  and
 downgradient of the oil body. The methane peak is displaced downgradient from the center of
 the oil body because of the input of methane-free ground water from the upgradient area. The
 predicted decline in methane concentration at a distance of 230 m (Figure 6c) is a result of the
 upwelling of oxygenated water caused by the upward bending of flow lines around the oil body.
 The simulated profiles show a marked increase in methane production from 1 986 to 1 990 as
 the population of methanogens increases. This increase in methanogenesis  results in a slight
 decrease in TDOC concentrations from 1986 to 1990  (Figure  6a). The  rate of increase in
 methane production was quite sensitive to the biodegradation parameters used in the simulation.

 To examine the effect of degradation on DDOC in the aquifer, two-dimensional distributions of
 DDOC for three different simulations are shown in Figure  7. In the first simulation, there is no
 degradation (Figure 7a); in the second simulation, degradation occurs (Figure 7b);  and in the
 third simulation, degradation occurs and the hydraulic conductivity distribution is heterogeneous
 (Figure 7c). The distribution of DDOC for the case with no degradation (Figure  7a) reflects the
 physical processes of dispersion,  diversion of flow around the oil body, and the depression of
 the plume beneath the water table because of the deflection of flow lines by incoming recharge
 water.
                                          RECHARGE ZONES
C  z
    -.  420
   <   41G
   EXPLANATION
 X midpoint of wallscroon
       zono boundary
_ 1 	 Isoconconlralion lino
   Datum is soa lovol
                             100
                                                200

                                         DISTANCE, IN METERS
                                                     300
                                                                         400
Figure 7.  Simulated two-dimensional distributions of DDOC: a) with no degradation, b) with
          degradation, c) with degradation and spatial variability.
130

-------
In the second  simulation, the anaerobic a
contaminant  plume  that  is narrower  than
                                          id aerobic degradation processes  result  in a
                                          the  plume in  the  first simulation and  whose
concentration gradients are comparatively she rp at the edges (Figure 7b). In this simulation, 46
percent of the total DDOC mass entering the aquifer is degraded: 14 percent by anaerobic
degradation and 32 percent by aerobic degradation.
Previous work has shown that the hydraulic properties of the aquifer are spatially variable (7,9).
To make the simulation more realistic, a heterogeneous hydraulic conductivity distribution was
created using the methods of Dillard et al. (1 (D) and was used in the transport model. Because
of the complex flow field, an irregularly shaped plume develops  (Figure 7c). The variability in
flow paths and flow velocities results in increased mixing and dispersion of ground water. This,
in turn, results in increased biodegradation. In this simulation, of the total DDOC mass entering
the aquifer, 60 percent is degraded: 21 percent by anaerobic degradation and 39 percent by
aerobic degradation.

The simulations  represent a highly simplified  representation of the true field conditions and
neglect Fe and Mn reduction. Also, the parameters used in the simulations are highly uncertain.
Nevertheless, the results do reproduce the general features of the observed contaminated
ground-water plume.  In addition to the kinetics of the biodegradation processes, important
factors that affect the magnitude of degradation and the distribution of the solutes in the field
are the recharge influx and the degree of dispersion and mixing in the ground-water system
caused by heterogeneity of the hydraulic con
                                          Juctivity.
                                           opportunity to study in detail the processes that
Summary

The Bemidji crude-oil spill site has provided a
occur following a spill of an organic immiscible fluid that is slightly soluble in water. Detailed
field studies have characterized the subsurface oil distribution and the  characteristics of the
contaminated ground-water plume. Numerical  modeling  of multiphase flow has illustrated the
importance of spatial variability on the movement and  distribution of oil in the subsurface.
Solute-transport modeling that includes aerob
used as a tool to study the field-scale solute-t
                                          c and anaerobic degradation processes is being
                                          •ansport and degradation processes. In addition
to the kinetics of the biodegradation processes, important factors that affect the distribution of
the solutes in the field are the recharge influx
ground-water system.
References
          Hult,  M.F.  1984.  Ground-water
          Minnesota,  research  site:  An  in
          contamination by crude oil at the
          Invest. Rep.  84-4188. pp. 1-15.
          Hess, K.M., W.N. Herkelrath, and
          fluid contents at a crude-oil spill site
                                          and the degree of dispersion and mixing in the
                                           contamination by  crude oil  at the Bemidji,
                                          reduction.  In:  Hult, M.F., ed.  Ground-water
                                           Bemidji, Minnesota, research site. Water Res.
                                          H.I. Essaid. 1992. Determination of subsurface
                                           . J. Contam. Hydrol. 10:75-96.
                                                                                    131

-------
3.        Murphy,  F., and W.N.  Herkelrath. 1994. A sample-freezing drive  shoe for a
          wireline-piston core  sampler.  In: Morganwalp,  D.W.,  and  D.A.  Aronson, eds.
          Proceedings of the Technical Meeting of the U.S. Geological Survey Toxic Substances
          Hydrology Program, Colorado Springs, CO (September 20-24, 1993). Water Res.
          Invest. Rep. 94:4014.

4.        Cary, J.W., J.F. McBride, and C.S. Simmons. 1991. Assay of organic liquid contents
          in  predominantly water-wet unconsolidated  porous media.  J. Contam. Hydrol.
          8:135-142.

5.        Essaid, H.I., W.N. Herkelrath, and L.A. Dillard. 1 994. Field and modeling studies of
          multiphase fluid flow at the Bemidji, Minnesota, crude-oil spill site. In: Dutton, A.R.,
          ed. Toxic Substances and the Hydrologic  Sciences. Minneapolis, MN: American
          Institute of Hydrology, pp. 52-68.

6.        Herkelrath, W.N., H.I. Essaid, and L.A. Dillard. 1 994. Multiphase fluid distributions
          in a shallow water-table aquifer contaminated by crude oil. In:  Dracos, T.H., and F.
          Stauffer, eds. Proceedings of the IAHR/AIRH Symposium on Transport and Reactive
          Processes in Aquifers, Zurich,  Switzerland (April 1 1-15).  Rotterdam, Netherlands:
          A.A. Balkema. pp. 537-542.

7.        Essaid, H.I., W.N. Herkelrath, and K.M. Hess. 1 993.  Simulation of fluid distributions
          observed at a crude-oil spill site incorporating hysteresis, oil entrapment, and spatial
          variability of hydraulic properties. Water Resour.  Res. 29(6):1,753-1,770.

8.        Deutsch, C.V., and A.G. Journel. 1 992.  GSLIB: Geostatistical Software Library and
          user's guide. New York: Oxford University Press.

9.        Dillard,  L.A. 1993.  Multiphase flow modeling  of a crude-oil spill  site  using
          geostatistical simulation  of soil  hydraulic  properties. M.S. thesis.  Stanford, CA:
          Stanford University.

10.       Dillard, L.A., H.I. Essaid, and W.N. Herkelrath. 1 994. Multiphase flow modeling at
          the Bemidji, Minnesota,  crude-oil spill site using geostatistical simulation  of soil
          hydraulic properties. In: Morganwalp, D.W., and D.A. Aronson, eds.  Proceedings of
          the Technical Meeting of the U.S. Geological Survey Toxic Substances Hydrology
          Program, Colorado Springs, CO (September 20-24, 1 993). Water Res. Invest. Rep.
          94-4014. In press.

11.       Baedecker, M.J., I.M. Cozzarelli, D.I. Siegel, P.C.  Bennett, and  R.P. Eganhouse.
          1 993. Crude oil in a shallow sand and gravel aquifer, 3. Biogeochemical reactions
          and mass balance modeling in anoxic ground water. Appl. Geochem. 8:569-586.

12.       Bennett, P.C., D.I. Siegel, M.J.  Baedecker, and M.F. Hult. 1993. Crude oil in a
          shallow sand and gravel aquifer, 1. Hydrogeologyand inorganic geochemistry. Appl.
          Geochem. 8:529-549.
132

-------
13.
14.



15.



16.


17.
18.
19.
20.
21
22.
Chang,  F.H., H. Wang,  B.  Denz
n, and J. Buller.  1991.  Kinetics of crude-oil
         biodegradation by bacteria indigerous to sediment and ground water. In: Mallard,
         G.E., and D.A. Aronson, eds.  Proceedings of the Technical Meeting of the U.S.
         Geological Survey Toxic Substancjes  Hydrology Program, Monterey, CA  (March
         11-15, 1991). Water Res. Invest. Rep. 91-4034. pp. 646-649.
Cozzarelli, I.M., R.P. Eganhouse,
and M.J. Baedecker.  1 990. Transformation  of
monoaromatic hydrocarbons to organic acids in anoxic ground-water environment.
Environ. Geol. Water Sci. 1 6:135^141.
Eganhouse, R.P., MJ. Baedecker,
Dorsey. 1993.  Crude  oil in a
geochemistry. App. Geochem. 8:5
 M. Cozzarelli, G.R. Aiken, K.A. Thorn, and T.F.
hallow sand  and gravel  aquifer,  2.  Organic
51-567.
Hiebert, F.K., and P.C. Bennett. lj?92. Microbial control of silicate weathering in
organic rich ground water. Science 258:278-281.
Hult, M.F., M.K. London, and H.O.
models  of  mobilization and  tran
unsaturated zone near Bemidji, Min
Proceedings of the Technical Meetir
Hydrology Program, Monterey, CA
4034. pp. 621-626.
Pfannkuch. 1 991. Field validation of conceptual
sport of volatile petroleum derivatives  in  the
lesota. In: Mallard, G.E., and D.A. Aronson, eds.
ig of the U.S. Geological Survey Toxic Substances
March 1 1 -15,1 991). Water Res. Invest. Rep. 91 -
Lovley, D.R., M.J. Baedecker, D.J. lonergan, I.M. Cozzarelli, E.J.P. Phillips, and D.I.
Siegel.  1989. Oxidation of aronatic contaminants coupled  to  microbial  iron
reduction. Nature 339:297-299.

Baedecker, M.J., D.I. Siegel, P.  Bennett, and  I.M. Cozzarelli. 1989. The  fate and
effects of crude oil in a shallow aquifer, I. The distribution of chemical species and
geochemical facies. In: Mallard, G.E., and S.E. Ragone, eds. Proceedings of the
Technical  Meeting of the U.S.  Geological Survey Toxic  Substances  Hydrology
Program, Phoenix, AZ (September 26-30, 1 988). Water Res. Invest. Rep. 88-4220.
pp. 13-20.

Goode, D.J., and LF. Konikow.  1 989. Modification of a  method-of-characteristics
solute-transport model to incorporate decay and equilibrium-controlled sorption or
ion exchange. Water Res. Invest. Rep. 89-4030. p. 65.

Konikow,  L.F., and J.D. Bredehoeft. 1 978. Computer model of two-dimensional
solute transport and dispersion in ground water. U.S. Geological Survey Techniques
          of Water Resources Investigations.
                                Book 7.
 Kindred, J.S., and M.A. Celia. 1989. Contaminant transport and biodegradation, 2.
 Conceptual model and test simulations. Water Resour. Res. 26(6):1,149-1,1 60.
                                                                                  133

-------
 23.       Essaid, H.I., MJ. Baedecker, and I.M. Cozzarelli. 1 994. Use of simulation to study
           field-scale solute transport and biodegradation at the Bemidji, Minnesota, crude-oil
           spill site. In:  Morganwalp, D.W., and DA Aronson, eds.  Proceedings  of  the
           Technical Meeting  of the  U.S.  Geological Survey Toxic Substances Hydrology
           Program, Colorado Springs, CO (September 20-24, 1 993). Water Res. Invest. Rep.
           94-4014 (in press).
134

-------
An Overview of Anaerobic Transformation  of Chlorinated Solvents
Perry L. McCarty
Department of Civil Engineering, Stanford Uqiversity, Stanford, CA
Abstract

Intrinsic cometabolic transformation of chlorinated solvents  commonly occurs  at sites where
co-contaminants are present as primary substi
activities of transforming bacteria. The extent
relative concentration  of  primary substrates
conditions present. Reduction of tetrachloroeth
ates to support the energy needs and metabolic
of transformation that occurs depends upon the
 and  the microorganisms  and environmental
sne (PCE) and trichloroethene (TCE) to ethene has
occurred at many sites, although transformations are often not complete. Evidence for intrinsic
biotransformation of chlorinated aliphatic hydrocarbons (CAHs) is provided by the presence of
CAH  transformation products  and indicators of anaerobic biogical  activity,  such as the
disappearance of dissolved oxygen, nitrates, qnd sulfates, and the production of methane and
soluble iron (II).
Introduction
Chlorinated solvents and their natural transformation products represent the most prevalent
organic ground-water contaminants in the cou itry. These solvents, consisting primarily of CAHs,
have been  used widely for degreasing of aircraft engines,  automobile parts,  electronic
components, and clothing. Only during the pcist 15 years has it become recognized that CAHs
can  be  transformed  biologically  (1).  Such transformations  sometimes occur under the
environmental conditions present in an aquifer in the absence of planned human intervention,
a process called intrinsic biotransformation (2)1 Conditions underwhich this is likely to  occurwith
CAHs and the end products that can be expected are discussed in this  paper.
The  major  chlorinated   solvents  are  carbon   tetrachloride  (CT),   PCE,  TCE,  and
1,1,1 -trichloroethane (TCA). These compounds can be transformed by chemical and biological
processes in soils to form a variety of other CAHs, including chloroform (CF), methylene chloride
(MC), cis- and trans-1,2-dichloroethene (c-DCE, t-DCE), 1,1 -dichloroethene (1,1 -DCE), vinyl
chloride (VC), 1,1 -dichloroethane (DCA), and chloroethane (CA). In CAH transformation, the
microorganisms responsible cannot obtain energy for growth from the transformations. The
transformations are brought about through co-metabolism or through interactions of the CAHs
with enzymes or  cofactors produced  by tne microorganisms  for  other  purposes.  In  co-
metabolism, other organic chemicals must be present to serve as primary substrates to satisfy
the energy needs of the microorganisms. Chemical  transformations of some CAHs can also
occur within the timeframe of interest in ground water. Transformations that are likely,  and the
environmental conditions  required, are discussed below.
 Symposium on Intrinsic Bioremediation of Ground Water
                                          135

-------
 Chemical Transformation

 TCA is the only major chlorinated solvent that can be transformed chemically in ground water
 under all conditions likely to be found and within the one- to two-decade time span of general
 interest, although chemical transformation of CT through reductive processes is a possibility.
 TCA chemical transformation occurs by two different pathways, leading to the formation of
 1,1-DCE and acetic acid (HAc):
       CH3CC13
         TCA
                                      CH2=cci2 + H+ + cr
                                       1,1-DCE
(elimination)     (1)
                                       CH3COOH + 3H+ + 3CT
                                        HAc
  (hydrolysis)    (2)
The rate of each chemical transformation is given by the first-order reaction:
               C  =  C,e
                        -kt
                (3)
where C is the concentration of TCA at any time t, C0 represents the initial concentration at
t = 0, and k is a transformation rate constant. The overall rate constant for TCA transformation
(kTCA) is equal to the sum of the individual rate constants (kDCE + kHAc). The transformation rate
constants are functions of temperature:
               k  =  Ae-E/0.008314K
                (4)
where A and E are constants and K is the temperature in degrees Kelvin. Table  1 provides a
listing of values reported forAand E for TCA abiotic transformation by various investigators, and
calculated values for the TCA transformation rate constant for 1 0°C, 1 5°C, and 20°C using
equation 4. Also given is the average calculated TCA half-life based upon tl/2 = 0.69/k. The
temperature effect on TCA half-life is quite significant.
Table 1. Reported First-Order TCA Abiotic Transformation Rates (kTCA)
A
yri
3.47 (10)20
6.31 (10)20
1.56 (10)20
Average half-life (yr)
E
Id
118.0
119.3
116.1


10
°C
0.058
0.060
0.058
12
I
-------
Ciine and Delfino (4) found that kDCE equaled
found it to be 22 percent. This means that a
7 fig/L compared with 200,Mg/L forTCA. Wh<
about 21 percent of KTCA/ and Haag and Mill (3)
most 80 percent of the TCA is transformed into
acetic acid. The 20-plus percent that is converted to 1,1 -DCE, however, is of great significance
because 1,1-DCE  is considered more toxic than TCA, with a maximum contaminant level of
neverTCAis present as a contaminant, 1,1 -DCE
can also be expected. In general, TCA is probably the main source of 1,1 -DCE contamination
found in aquifers.

Chloroethane, formed through biological transformation of TCA, can also be chemically
transformed with a half-life on the order of months by hydrolysis to ethanol, which can then be
biologically converted to acetic acid and nor harmful products (6).
Biological Transformations

CAHs can be oxidized or reduced, generally
.ground waters, intrinsic reductive transformat
presence of intermediate  products that are
through co-metabolism, as noted in Table 2. In
ons are most often noted, perhaps because the
formed  provide strong evidence that reductive
transformations are taking place.  Intrinsic aerobic transformation of TCE is also possible,
although if it did occur, the intermediate products are unstable and more difficult, analytically,
to measure. Thus, convincing evidence for the latter is difficult to obtain. Also, aerobic co-
metabolism of TCE would only occur if sufficient dissolved oxygen and a suitable electron donor,
such as methane, ammonia, or phenol, were present. Since circumstances under which the
proper environmental conditions for significant aerobic co-metabolism are unlikely to occur
often, intrinsic aerobic co-metabolism of TCE is probably of little significance. Evidence is ample,
however, that anaerobic reductived transformations of CAHs occur frequently, and this process
is important to the transformation of all chlorinated solvents and their transformation products.
The major environmental requirement is the presence of sufficient concentrations of other
organics that can serve as electron donors for energy metabolism, which often is the  case  in
aquifers. Indeed, the extent to which  reductive dehalogenation occurs may be limited by the
amount of such co-contaminants present. Theoretically, only a 0.4-g chemical oxygen demand
(COD) equivalent of primary substrate would
but much more is actually required because
be required to convert 1  g of PCE to ethene (7),
f the co-metabolic nature of the transformation.
Figure 1 illustrates the potential chemical and biological transformation pathways for the four
majorchlorinated solvents underanaerobicenyironmental conditions (6). Freedmanand Gossett
(8) provided the first evidence for conversion of PCE and TCE to ethene, and de Bruin et al. (9)
reported completed reduction to ethane. Table 3 indicates that while some transformations, such
as CT to chloroform and carbon dioxide, may take place under mild reducing conditions such
as those associated with denitrification, complete reductive transformation to inorganic end
products and  of PCE and TCE to ethene generally requires conditions suitable for methane
fermentation. Extensive reduction, although perhaps not complete, can also occur under sulfate-
reducing conditions. For methane fermentation to occur in an aquifer, the presence of sufficient
organic co-contaminant is required to reduce the oxygen, nitrate, nitrite, and sulfate present.
Some organics will be required to  reduce the CAHs, and  perhaps Fe(ll) as well, if  present in
significant amounts. If the potential for intrinsic transformation of CAHs is to be evaluated, then
the concentrations of nitrate, nitrite, sulfate, Fe(ll), and methane, and of organics (as indicated
by COD ortotal organic carbon [TOC]) should be determined. Unfortunately, such analyses are
not considered essential in remedial investigations, but it is evident that they should be.
                                                                                   137

-------
Table 2. Biodegradability of Chlorinated Solvents Under Aerobic or Anaerobic Conditions and
        Through Use as a Primary Substrate for Energy and Growth or Through Co-metabolism

Carbon
Tetrachloride
(CT)
Tetrachloro-
ethylene
(PCE)
Trichloro-
ethylene
(TCE)
1 ,1 ,1 -Tri-
chloroethane
(TCA)
Aerobic Biotransformation
Primary substrate
Co-metabolism
No
No
No
No
No
Yes
No
Perhaps
Anaerobic Biotransformation
Primary substrate
Co-metabolism
Hazardous intermediates
Chemical Transformation
No
Yes
Yes
Perhaps
Perhaps
Yes
Yes
No
Perhaps
Yes
Yes
No
No
Yes
Yes
Yes
          CCI2 =CCI2  PCE
CT
                  CH 3 CH3
           ANAEROBIC  TRANSFORMATIONS  OF  CHLORINATED  SOLVENTS
Figure 1.  Anaerobic chemical and biological transformation pathways for chlorinated solvents.
138

-------
Table 3.  Environmental Conditions Genera
          Chlorinated Solvents
ly Associated With Reductive Transformations of

Chlorinated Solvent
Carbon tetrachloride
1 ,1 ,1 -Trichloro-
ethane
Tetrachloroethylene
Trichloroethylene
Redox Environment
All

TCA-*1,1-DCE
+ CH3COOH


E
c
(




)enitrifi-
ation
T-»CF



Sulfate
Reduction
CT->C02+C|-
TCA^1,1-DCA
PCE^1,2-DCE
TCE^1,2-DCE
Metha no-
genesis

TCA-*C02+d-
PCE-s-ethene
TCE->ethene
Case Studies

Major et al. (10)  reported field evidence for intrinsic bioremediation of PCE to ethene and
ethane at a chemical transfer facility in North Toronto. PCE was stored at the site 1 0 years prior
to the study and contaminated the ground water below with both free and dissolved PCE. In
addition to high concentrations of PCE (4.4 mg/L), high concentrations of methanol (810 mg/L)
and acetate (430  mg/L) were found in the contaminated ground water; methanol and acetate
are co-contaminants that served as the primary substrates forthe transforming organisms. Where
high PCE  was  found, TCE (1.7 mg/L), cis-DJCE (5.8 mg/L), and  VC (0.22 mg/L) were also
found, but little ethene (0.01 mg/L) was found. At one downgradient well, however, no PCE or
TCE was found, but cis-DCE (76 mg/L), VC (9.7 mg/L), and ethene (0.42 mg/L) were present,
suggesting thatsignificantdehalogenation had occurred. Otherdichloroethylenes (1,1 -DCEand
trans-DCE) were  not significant in  concentration,,  indicating that cis-DCE was the  major
transformation intermediate. Microcosm studies also supported that biotransformation was
occurring  at the site, with complete disappearance of PCE, TCE, and cis-DCE and production
of both VC and ethene. The conversions were accompanied by significant methane production,
indicating that suitable redox conditions were present for the transformation.
Fiorenza et al. (1 1) reported on PCE, TCE, T
3A, and dichloromethane (DCM) contamination
of ground  water at two separate locations at a  carpet-backing manufacturing  plant in
Hawkesbury, Ontario. The waste lagoon was pie major contaminated area, with ground water
containing 492 mg/L of volatile fatty acids and 4.2 mg/L of methanol, organics that appeared
to provide the co-contaminants that served as primary sources of energy forthe dehalogenation
reactions. Here, the sulfate concentration was nondetected, but the concentration in  native
ground water was about 1 5 to 18 mg/L. Total dissolved iron was quite high (1 9.5 mg/L) and
well above the upgradient concentration of 2.1 mg/L. Methane was present, although quite low
in concentration (0.06 mg/L). These parameters are all supportive of conditions suitable for
intrinsic biodegradation of the chlorinated solvents. Whilesome chemical transformation of TCA
was indicated  (0.4 mg/L), biotransformation was quite extensive, as indicated by a 1,1-DCA
concentration of 7.2 mg/L compared with the TCA concentration of 5.5 mg/L. Some CA was
also present (0.1 9 mg/L). Transformation was also indicated for PCE and TCE, which remained
at concentrations of only 0.01 6 mg/L and 1.5 mg/L, respectively, while the cis-DCE, VC, and
                                                                                  139

-------
ethene concentrations were 56, 4.2, and 0.076 mg/L, respectively. Only traces of ethane were
found. Trans-DCE concentration was only 0.57 mg/L, again providing evidence that cis-DCE
is the most common transformation intermediate from TCE and PCE. Downgradient from the
lagoon, the dominant products were cis-DCE (4.5 mg/L), VC (5.2 mg/L), and  1,1 -DCA (2.1
mg/L). While good evidence for intrinsic biotransformation is provided for this site, the ethene
and ethane concentrations appear very low compared with VC concentration, suggesting that
biotransformation was not eliminating the chlorinated solvent hazard at the site, although it was
producing compounds that may be more susceptible to aerobic co-metabolism.

Evidence for intrinsic biotransformation of chlorinated  solvents has also been provided from
analyses of gas from municipal refuse landfills where active methane fermenation exists. A
summary by McCarty and  Reinhard (12) of data  from Charnley et al. (13) indicated average
gaseous concentrations in  parts per million by volume from eight refuse landfills to be: PCE,
7.15; TCE, 5.09; cis-DCE, not measured; trans-DCE, 0.02; and VC, 5.6. While these averages
indicate that, in general, transformation was not complete, the presence of high VC indicates
the transformation was significant. ForTCA, gaseous concentrations were: TCA, 0.1 7; 1,1 -DCE,
0.10; 1,1 -DCA, 2.5; and CA, 0.37. These data indicate that TCA biotransformation was quite
extensive, with the transformation intermediate, 1,1 -DCA, present at quite significant levels, as
is frequently found  in ground water.

Perhaps the most extensively studied and reported intrinsic chlorinated solvent biodegradation
is that at the St. Joseph, Michigan, Superfund site (7, 14-17). Ground-water concentrations of
TCE as high as 100 mg/L were found present, with extensive transformation to cis-DCE, VC,
and  ethene. A high but undefined COD  (400 mg/L) in ground water, resulting from waste
leaching from a disposal lagoon, provided the energy source for the co-metabolic reduction of
TCE.  Nearly complete conversion of the  COD to methane provided  evidence of the  ideal
conditions for intrinsic bioremediation (7). Extensive analysis near the source of contamination
indicated that 8 percent to 25 percent of the TCE had been converted to ethene, and that up
to 15  percent of the  reduction  in  COD  in  this zone was  associated with reductive
dehalogenation (15). Through more extensive analysis of ground  water farther downgradient
from the contaminating source, Wilson et al. (1 7) found a 24-fold reduction in CAHs across the
site.  A review of the  data  at individual sampling points indicated  that conversion of TCE to
ethene was most complete where methane production was highest and removal of  nitrate and
sulfate by reduction was most complete.
References

1.        McCarty, P.L., and  L. Semprini.  1994. Ground-water treatment for chlorinated
          solvents. In: Morris, R.E., ed. Handbook of bioremediation. Boca Raton, FL: Lewis
          Publishers, Inc. pp. 87-11 6.

2.        Council, N.R. 1 993. In situ bioremediation: When does it work?. Washington,  DC:
          National Academy Press.

3.        Haag, W.R.,  and T.  Mill.  1988. Effect of subsurface sediment on  hydrolysis of
          haloalkanes and epoxides. Environ. Sci. Technol. 22:658-663.
140

-------
4.
5.
6.
7.
8.
9.
10.
n.
 12.
 13.
 14.
 15.
Cline, P.V., and J.J. Delfino. 1 989.
to the stable product 1,1 -dichloroe
water treatment. Chelsea, Ml: Lew^s

Jeffers, P., L. Ward, L. Woytowitch
Transformation kinetics of 1,1,1 -trichloroethane
hene. In: R.A. Larson, ed. Biohazards of drinking
  Publishers, Inc. pp. 47-56.

 and L. Wolfe.  1 989.  Homogeneous hydrolysis
          rate constants for selected chlorine ted methanes, ethanes, ethenes, and propones.
          Environ. Sci. Technol. 23(8): 965-969.
Vogel, T.M., C.S. Griddle, and P.L
aliphatic compounds. Environ. Sci,
vlcCarty. 1 987. Transformations of halogenated
 Technol. 21:722-736.
McCarty, P.L, and J.T. Wilson. 1 992. Natural anaerobic treatment of a TCE plume,
St. Joseph, Michigan, NPL site. In: U.S.  EPA. Bioremediation of hazardous wastes.
EPA/600/R-92/126. Cincinnati, OH.  pp. 47-50.
Freedman, D.L., and J.M. Gossett.
tetrachloroethylene and trichloroeth yl
Appl. Environ. Microbiol. 55(9):2,
  1 989. Biological  reductive  dechlorination  of
  ene to ethylene under methanogenic conditions.
144-2,151.
de  Bruin, W.P., et al.  1992. Complete biological reductive fransformation  of
tetrachloroethene to ethane. Appl.

Major, D.W., W.W. Hodgins, and
 Environ. Microbiol. 58(6):1,996-2,000.

 3.J. Butler. 1 991.  Field and laboratory evidence
of in situ biotransformation of tetrachloroethene to ethene and ethane at a chemical
transfer facility in North Toronto. In: Hinchee, R.E., and R.F. Olfenbuttel, eds. Onsite
bioreclamation. Boston,  MA: Butterworth-Heinemann. pp. 147-171.

Fiorenza, S., et al. 1 994. Natural anaerobic degradation of chlorinated solvents at
a Canadian manufacturing plant. In: Hinchee, R.E., A. Leeson, L. Semprini, and S.K.
Ong, eds. Bioremediation  of chliorinated and  polycyclic aromatic hydrocarbon
     pounds. Boca Raton, FL: Lewis Publishers, Inc. pp. 277-286.
          com
McCarty, P.L., and M. Reinhard. 1993. Biological and chemical transformations of
halogenated aliphatic compounds  in aquatic and terrestrial environments. In:
Oremland, R.S., ed. The biogeocnemistry of global change: Radiative trace gases.
New York, NY: Chapman & Hall, Inc.

Charnley, G., E.A.C.  Crouch, L.G. Green, and T.L. Lash. 1988. Municipal solid
waste landfilling: A review of environmental effects.  No. Meta Systems, Inc.
 Hasten, Z.C., P.K. Sharma, J.N. Black, and P.L. McCarty. 1 994. Enhanced reductive
 dechlorination of chlorinated ethenes. In: U.S. EPA.  Bioremediation of hazardous
 wastes. San Francisco, CA.

 Kitanidis, P.K.,  L. Semprini, D.rj.  Kampbell, and  J.T.  Wilson.  1993. Natural
 anaerobic bioremediation of TCE ctthe St. Joseph, Michigan, Superfund site. In: U.S.
 EPA. Symposium on bioremediation of hazardous wastes: Research, development,
 and field evaluations  (abstracts).  EPA/600/R-93/054. Washington, DC (May).
 Cincinnati, OH. pp. 47-50.
                                                                                  141

-------
 1 6.      McCarfy, P.L., et al. 1 991. In situ methanotrophic bioremediation for contaminated
          ground water at St. Joseph, Michigan. In: Hinchee, R.E., and R.G.  Olfenbuttel, eels.
          Onsite bioreclamation processes forxenobiotic and hydrocarbon treatment. Boston,
          MA: Butterworth-Heinemann. pp. 1 6-40.

 17.      Wilson, J.T., J.W. Weaver, and  D.H.  Kampbell. 1994. Intrinsic bioremediation of
          TCE in ground water at an NPL  site in St. Joseph, Michigan.  Presented at the U.S.
          EPA Symposium on Intrinsic Bioremediation of Ground Water, Denver, CO (August
          30 to September 1).
142

-------
Contamination of Ground Water With Trichloroethylene at the Building 24 Site at
Picatinny Arsenal, New Jersey
Mary Martin and Thomas E. Imbrigiotta
U.S. Geological Survey, West Trenton, NJ
Abstract

Ground water at the Building 24 site at Pic
jtinny Arsenal in Morris  County, New Jersey, is
contaminated withtrichloroethylene (TCE). Estimated average linearground-waterflowvelocities
are 0.3 to 1.0 m/d, and estimated travel time
from the source area to Green Pond Brook is 2 y
to 5 y. The total mass of dissolved TCE within the 1 30,000-m2 plume area is estimated to be
970 kg. About 65 percent of the mass is in areas where TCE concentrations exceed 10,000
ftg/t, whereas about 30 percent is in areas vhere TCE  concentrations are 1,000 to 10,000
The average flux of TCE discharged to Green Pond Brook from the plume area, estimated from
measured TCE concentrations in samples of water from the unconfined aquifer and measured
base-flow discharge in Green Pond Brook, s 1 to  2 mg/sec. Biotransformation is the most
important mechanism by which dissolved TC: leaves the ground-water system.
Introduction

Picatinny Arsenal, a U.S. Army armament re
glaciated valley in north-central New Jersey  (
and a new metal-plating facility and industria
1 960 to 1 981,  the wastewater treatment s
search and development center, is located in a
Figure 1). In 1960, Building 24 was remodeled,
 wastewater treatment plant were installed. From
stem  discharged tens of thousands  of  liters of
wastewater daily into two 2.5-m deep, sand-bottomed settling lagoons behind the building (1).
The metal-plating wastewater contained trace metals, such as cadmium, chromium, copper,
lead,  nickel, tin, vanadium, and zinc, and othpr ions used in plating solutions, such as sodium,
potassium, sulfate, chloride, and cyanide (2). From 1 973 to 1 985, an improperly installed relief
system of the degreasing unit allowed pure chlorinated solvents to condense in an overflow pipe
and discharge to a 1 -m deep dry well in front of Building 24. The condensate from the overflow
system contained TCE and, after 1 983, 1,1,
 -trichloroethane (2).
The infiltration of wastewater from the lagoons and of chlorinated solvents from the dry well has
created a plume of contaminated  ground water downgradient from Building 24. Most of the
contamination is limited to unconfined sediments, where estimated ground-water flow velocities
(based on estimated horizontal hydraulic conductivities, measured head gradients, estimated
porosities, and results of calibrated solute-transport model simulations) generally range from 0.3
to 1.0 m/d in the plume area. On the basis
within the unconfined aquifer of a conservative solute entering the aquifer near Building 24 and
discharging at Green Pond Brook is 2 y to 5
of these estimated velocities, the residence time
Symposium on Intrinsic Bioremediation of Ground Water
                                         143

-------
                                                                           A. Areal extent of trichloroethylene
                                                                             contaminant plume
                                                                                               EXPLANATION
                                                                              Ky*:|:'x)   Area in which trichloroethylene concentration
                                                                              l^ili'a    exceeds 10 micrognuns per liter

                                                                            	10	LINE OF EQUAL TRICHLOROETHYLENE
                                                                                       CONCENT'RATION-Shows trichloroethylene
                                                                                       ennCCTTff tjon, in nucrognu&s per liter.
                                                                                       Dashed where approximate

                                                                             A    "A'  Line of section

                                                                              *41 -9    Sampling site location and local identifier
        BOO 400  200   0  200  400  600  800 1.000 1.200 1.40O 1.6O01.800 2.000

                  DISTANCE FROM BUILDING 24 CB-24). IN FEET
                                                                          B. Vertical distribution of trichloroethylene
                                                                            concentrations
                                                                                              EXPLANATION

                                                                             -100	LINE OF EQUAL TRICHLOROETHYLENE
                                                                                      CONCENTRATION-Shows trichloroeibyiene
                                                                                      conccxumion, in nucrogninu per liter.
                                                                                      Dashed where approximate

                                                                                     Weil sacea ani tricblorodhylene conccnmtlion.
                                                                              Utoo

                                                                               NS
                                                                                     Not sampled

                                                                                     Less than

                                                                                     Localioa of well and local identifier
Figure 1. (A)  Location of Building 24 study area at Picatinny Arsenal, and areal extent of TCE
            plume and altitude of water table, January 1 993, and (B) vertical distribution of TCE
            concentrations, October to November 1 991.  (Location of section A-A'  is shown in
            Figure 1A.)
144

-------
Ground-water contamination in the unconfinec
et al. (3), Fusillo et al. (2), and Imbrigiotta et c
   aquifer has been discussed previously by Sargent
   I. (4,5). Results of water-quality sampling  during
1 986 to 1 991 confirm that TCE remains the dominant contaminant in the unconfined aquifer
and that the extent of the plume of TCE-contaminated ground water has changed little since
September 1 986. These water-quality measurements also show that 1)  Building 24 is the source
of the TCE plume; 2) the highest TCE concen
less than 6 m and within 10  m downgradienl
from Building 24 to Green Pond Brook and fo
   rations near Building 24 are found at depths of
   from the dry well; 3) the plume extends 500 m
   lows the ground-water head gradients as it flows
downward through the unconfined aquifer, then upward toward Green Pond Brook; 4) the
plume disperses as it moves downgradient, is
   about 350-m wide where it enters Green Pond
Brook, and has an areal extent of about 1 30,000 m2; 5) the highest concentrations of TCE are
found at the base of the  unconfined aquifer
Brook; and 6) the highest TCE concentrations
than 6 m (Figure 1).
   midway between Building  24 and Green Pond
   at Green Pond Brook are found at depths of less
ard
The U.S. Geological Survey is conducting an
contamination by the chlorinated solvents
Picatinny Arsenal. The objectives of the
biological processes that affect the movement
in the subsurface; 2) determine the relative
predictive models of contaminant transport.
      interdisciplinary research study of ground-water
        other contaminants at the Building 24 site at
       are to 1) describe the chemical, physical, and
      and fate of these contaminants, particularly TCE,
      mportance of these processes; and 3) develop
This paper describes a conceptual model o
chlorinated solvents at Picatinny Arsenal resea
in the  unconfined  aquifer at the site
solute-transport simulations.
I tie
         processes that affect the fate and transport of
      •ch site. A preliminary solute mass balance of TCE
is presented and compared with results of numerical
Conceptual Model
A conceptual model of the physical, chemical,
and  mass balance  of TCE within the plume
(Figure 2). TCE is present at the site in severe
phase (soil gas), and matrix phase (sorbed o
also may be present as a dense nonaqueous
of TCE from Building 24 to the base of the
of high concentrations of TCE at the base of
Results of chemical analyses of samples of
the plume  area indicate that  the fate and
chemical, and biological processes at the site
affect the movement of dissolved TCE and
discharge  to  Green  Pond Brook.  TCE als
biotransformation (reductive dehalogenation
saturated-zone sediments. Volatilization of TCJE
by  Smith and others  (7).  Transport in the
principally   by molecular  diffusion.  The
cis-1 -2-dichloroethylene (cis-DCE) and vinyl c
   and biological processes that affect the transport
   was developed by Imbrigiotta and Martin (6)
   forms: solute phase (dissolved in water), vapor
   ito solid surfaces or associated with biota). TCE
   -phase liquid (DNAPL). Nonaqueous-phase flow
ur confined aquifer is hypothesized to be the cause
   the unconfined aquifer (5).
ground
           water, soil gas, and aquifer sediments from
      fransport of TCE are affected  by the  physical,
      . Advection and dispersion in the saturated zone
      cause it to be removed  from the system in the
           removed from the system by anaerobic
  o  is
  , volatilization at the water table, and sorption to
    to the unsaturated-zone soil gas was measured
  imsaturated zone was determined to be driven
  detection  of  the  biotransformation  products
  hloride (VC) in waterfrom 75 percent of the wells
                                                                                  145

-------
                                        VOLATILIZATION
                    Bnildini 24              (0.1 mg/s)
                        \	      W«er table
             LONG-TERM
             DE3ORPTION
             (1S-8S mg/s)
                SHORT-TERM
                DESOHPTION
               (Not estimated)
                             ADVECTIVE
                           TRANSPORT TO
                         QREEN PONO BROOK
                             (1-2 mg/s)

                                       Land surface
                                TH>'C.H:LORO  T.«YUEJ*E
                                    DISSOLUTION
                                    (Not estimated)
                                    ANAEROBIC
                                BIOTRANSFORMATION
                                    (1-3O mg/s)

                                                            Estimated top of confining unit
          NOT TO SCALE
            GAINS
                            TRICHLOROETHYLENE MASS BALANCE COMPONENTS
                                 [mg/s. milligrams par second; <. lass than]
                                                 LOSSES
            LONG-TERM DESORPTION
            INFILTRATION
            DISSOLUTION
            SHORT-TERM DESORPTION
  15-85 mg/s
  < 0.1 mg/s
Not estimated
Not estimated
ANAEROBIC BIOTRANSFORMATION   1-30 mg/s
ADVECTIVE TRANSPORT TO
GREEN POND BROOK             1-2 mg/s
VOLATILIZATION                 0.1 mg/s
SHORT-TERM ADSORPTION     Not estimated
Figure 2. Conceptual model and preliminary mass-balance estimates of TCE fluxes resulting
          from processes that affect the fate and transport of TCE in the ground-water system
          at Picatinny Arsenal.
screened within the plume is indicative of biologically mediated, reductive dehalogenation of

TCE. The occurrence of methanogenesis is indicated by the detection of dissolved methane in

water from 85 percent of the wells in which cis-DCE was detected and in water from 94 percent
of the wells in which VC was detected.


Desorption from contaminated sediments is a source of TCE in the ground-water system. If TCE

is present as a DNAPL, dissolution will result in a source of dissolved TCE in the ground water.

Relatively constant TCE concentrations measured  near Building 24  and at the  base of the

unconfined aquiferfrom 1986 to 1991  indicate that slow desorption of TCE from contaminated
146

-------
aquifer sediments, dissolution of DNAPL TCE  n the aquifer, or both are a continuing source of
dissolved TCE in the ground-water system.
Estimated  Mass Distribution of TCE
For a given sample set, the total mass of dissolved TCE below the water table is estimated to
be 970 kg or about 660 L of liquid TCE. This estimate was calculated by using results of six sets
of water-quality analyses  made from  1986  to 1991 (Figure  3).  Each  measured  TCE
concentration is assumed to represent the TCE concentration of a volume of ground water
extending  half the  horizontal and  vertical d stance to the  adjacent sampling points. These
volumes are limited vertically by the extent OT the unconfined aquifer and horizontally by the
1 0-jug/L maximum TCE concentration  line. Aquifer porosity  is assumed to be 30 percent. The
total volume of dissolved TCE  within  the ground water outside the 10-^g/L maximum  TCE
concentration line is estimated to be about 1 L.

The estimated total volume  of TCE appears  to be unrelated to the total number of samples
collected.  Four sets of data plotted in Figure 8 (A, C, E, and F) show that about 65 percent of
the total volume of TCE  is found in ground water that contains TCE concentrations greater than
1 0,000 jug/L, and about 30 percent is founc
1,000 to 1 0,000 pg/L The estimate of the
mostly on the number of samples that conta
volume of water each sample is assumed to
The relative amounts of TCE in the dissolved
 in areas where TCE concentrations range from
volume of TCE in the water appears to depend
n relatively high  concentrations of TCE and the
represent.

 vapor, and matrix phases in a block of aquifer
and unsaturated zone immediately downgradientfrom Building 24 were estimated on the basis
of measured TCE concentrations in samples cf all three phases (8) (Figure 4). Most of the mass
of TCE in the system near Building 24 is in the matrix phase and is associated with the sediments
in both the unsaturated and saturated zones. The mass distribution shown in Figure 4 is based
on a representative sample of each TCE phas 5 in the area of Building 24. The ratio of the mass
of TCE sorbed to the soil to the mass of disso ved TCE in six sets of samples collected near the
water table throughout the site ranged from 2:1  to 5:1. One sample, collected downgradient
from the wastewater lagoons, had an unusual
TCE at the site has not been estimated.
y high ratio of about 20:1. The amount of DNAPL
Preliminary  Solute  Mass Balance

Desorption of TCE from soils that have undergone long-term adsorption (years) and dissolution
of DNAPL TCE probably are the processes by
because the direct release of TCE from Build
DNAPL TCE has not been estimated, the rate of TCE dissolution cannot be estimated. Three
first-order rate constants  of TCE desorption
calculated by Koller et al. (8) ranged from 0.
measured in laboratory flowthrough columns
which most TCE enters the ground-water system,
ng 24 stopped in 1 985. Because the amount of
 from shallow aquifer sediments at the  arsenal
003 to 0.01 5 per week. The rate constants were
using uncontaminated water as the influent fluid.
Desorption rates in the field, where ground water containing TCE is flowing past the desorbing
sediments, probably would be lower. By using an estimated mass of TCE sorbed to the aquifer
sediments of three to four times the mass of TCE in the dissolved state, the estimated fluxof TCE
into the ground-water system through desorpti
on is 15 to 85 mg/sec. This flux estimate is made
                                                                                 147

-------
               900
            £  800

            1U
               700 -
            UJ

            g  600 |-
            UJ
            O
            cc

            3
            X
            O
            £
            i-
               500
400
300
            O

            111  ZOO


            13
            _!
            o  100
           Sample set    A

     Number of samples   96
                     B

                    39
C


40
D

59
e

50
F


42
                                                 EXPLANATION

                                (jig;L, mlcrograms pur llt«r: s, less than or oqual to: >, groator than)

                                                         Sample
                                                           sal
     Volume of Irichloroathylana calculated
     using samples with Irichloroathylana
                    concentrations of:

                      m  1 to £100u.g/L

                      gjjgg  >100 to s1.000ng/L

                             >1,000 to 510.000|ig/L


                             >10,000|ig/L
                                                                      Data of collection...
                                            1|300



                                            1,200



                                            1.100



                                            1,000



                                            900



                                            800



                                            700



                                            600



                                            500
                                                  ^.
                                                  O

                                            400    K
                                                  1-

                                            300    £



                                            200    CO



                                            100



                                            0
                                            A    April and August-Soptambor 19S6

                                            B    August 1987

                                            C    June 1989

                                            0    November-December 1989

                                            E    April-March 1S90

                                            F    October-November 1991
Figure 3.  Estimated volume of dissolved TCE in ground water at the Building 24 site at Picatinny

            Arsenal,  1986-91.
                                              24    Mass distril3ution of tricnoloroethylene near building 24
                                                                  [kg, kilogram; <, less than]
                                        NOT TO SCALE
                                                                       Phase
                                                                                  Percent

                                                                    Mass (kg)     of total
10 feet
10 faat
Unsaturated
zone
Saturated
zone
Vapor
Matrix
Solute
Matrix
0.001
3
1
4
99.9
20
80
Figure 4. Mass  distribution of TCE in  the  saturated and  unsaturated  zones  immediately

           downgradient from  Building 24 at Picatinny Arsenal.
148

-------
by assuming that, over long periods (years), the short-term desorption rate (weeks and months)
is equal to the short-term adsorption rate, anc
adsorption, which is no longer occurring.
 that soils previously have undergone long-term
Preliminary estimates of the flux of TCE into and out of the ground-water system at the Building
24 site for each of the mass-balance components are shown in Figure 2. The estimated flux of
TCE discharged to Green Pond  Brook from the plume area, calculated  on the basis  of
measured TCE concentrations in ground wateirand measured base-flow discharge in the brook,
is 1 to 2 mg/sec. The flux of TCE volatilized fj-om the water table is estimated to be about 0.1
mg/sec on the basis of measured soil-gas TCE concentration gradients and estimates of the
physical characteristics of the unsaturated zone.
Biotransformation probably is the mechanism by which most of the dissolved TCE leaves the
ground-water system. First-order rate constan
0.001 to 0.02 per week were estimated by Wi
microcosm studies of soil from five sites within
s for TCE transformation ranging from less than
son et al. (9) on the basis of results of laboratory
he plume area. By using these rate constants and
the estimated mass of  dissolved TCE,  the  rate  of TCE loss  from the  plume  through
biotransformation is calculated to be about 1 to 30 mg/sec. Analogous first-order rate constants
for TCE  biotransformation calculated  by |Ehlke  et al.  (10)  from  field-measured  TCE
concentrations and time-of-travel data generally were higher than those measured in the
laboratory experiments. Thus, the actual flux
greater than that shown in Figure 2.
A reactive multispecies transport model of a two-dimensional vertical section along the central
axis of the plume is being used to analyze the
chemical transport characteristics and th
calibration, and sensitivity analysis have been
volatilization, and microbial degradation of T
 of TCE lost through biotransformation may be
aboratory and field estimates of the physical and
timated TCE mass balance. The model design,
described by Martin (1 1). Transport, desorption,
IE are simulated. The formation and transport of
the degradation products cis-DCE and VC also are simulated. Specified-flux boundary conditions
are used to represent ground-water recharge and flows  across the horizontal and vertical
boundaries of the cross-sectional area. Constant-concentration nodes and desorption within the
plume area are solute sources to the  simulated system.  The constant-concentration solute
sources are assumed to represent high rates of desorption ordissolution of TCE near the settling
lagoons, at the overflow dry well, and  near the  base of  the  unconfined  aquifer 230 m
downgradient from Building 24.

The simulations were designed to represent cverage steady-state flow conditions and virtually
steady-state transportconditions after 1 985. S
the calibrated mode! are shown in Figure 5.
imulated concentrations of TCE and cis-DCE from
Most concentrations were simulated to within an
order of magnitude of average concentrations measured in micrograms per liter in water
samples from each well. Because the degradation of VC was not simulated, concentrations are
higher than measured concentrations and are not shown.
The calibrated model does not provide a unique estimate of the magnitudes of the various
mass-balance components of the plume of TCE-contaminated ground water at the Building 24
 site; however, sensitivity simulations were use
 ground water. Results of a series of sensitivity
 general conceptual model as defined by the estimated solute mass balance presented above.
 to test hypotheses concerning the fate of TCE in
simulations discussed by Martin (11) support the
                                                                                   149

-------
                                                                  (A) Trichloroethylene
  185 •  —
          1OO     0     100    200    300     400    SOO

             DISTANCE FROM BUILDING 24 (B-24). IN METERS
                                                                 (B) cis-l-2-Dichlordethylene
  IBS -   	
             DISTANCE FROM BUILDING 24 (B-24). IN METERS
                                           EXPLANATION

                           —100	LINE OF EQUAL SOLUTE CONCENTRATION-
                                     -Shows solute concentration, in micrograms per liter.

                             Q      Location of well screen

                            CAF-7    Location of well and local identifier
Figure 5.  Simulated   concentrations  from  calibrated  model  with  simulated  desorption,
           volatilization, and microbial degradation:  (A) TCE and (B) cis-DCE.  (Location of
           section A-A' is shown in Figure 1A.)
150

-------
Results of sensitivity simulations made with various desorption and degradation rates typically
showed that use of the laboratory estimates resulted in reasonable simulated concentrations.
Although volatilization is not a major mass-be
an important mechanism for removing solutes
the water table. The overall flux of TCE into
                                         lance component, this process was shown to be
                                         and thereby affecting solute concentrations near
                                          and out of the system  was not simulated well
because the total simulated mass of TCE in the system was too low. Increasing the simulated
solute-source area near the base  of the urconfined aquifer might result in  a  reasonable
simulated TCE mass balance.
                                          chlorinated solvents from the dry well at Building
Summary

Infiltration of wastewaterfrom the lagoons anc
24  at Picatinny Arsenal, New Jersey, has created a plume of contaminated ground water
downgradient from the building. TCE is the predominant contaminant in the 1 30,000 m2 plume,
which extends 500 m to Green Pond Brook. Ground-water velocities typically range from 0.3
to 1.0 m/d.
Results of water-quality sampling  conducted
                                          from 1986 through  1991 show that the TCE
contaminant  plume has changed little since  September 1986, and that the  highest TCE
concentrations are found near the water table near Building 24 and  near the base of the
unconfined aquifer about midway between Building 24 and Green Pond Brook TCE is present
at the site in several phases: dissolved in wa
solid surfaces or associated with biota. TCE a
of dissolved TCE below the water table is est
of TCE in the system is associated with the se

A conceptual model of the physical, chemical,
and mass balance of TCE within the plume ir
the  ground-water system by  discharge  to
volatilization; and 3) gain of dissolved TCE
sediments and possibly from dissolution of D
                                         er, as a vapor in the soil gas, and sorbed onto
                                         so may be present as a DNAPL. The total volume
                                         mated to be about 660 L, but most of the mass
                                         diments in the saturated and unsaturated zones.

                                         and biological processes that affect the transport
                                         eludes  1) transport of TCE from the Building 24
source area to Green Pond Brook by advection and dispersion; 2) loss of dissolved TCE from
                                          Green Pond  Brook,  biotransformation, and
                                         by slow desorption from contaminated  aquifer
                                         IAPLTCE in the aquifer. Preliminary estimates of
the flux of dissolved TCE discharged to Gredn Pond Brook from the plume is  1 to 2 mg/sec.
Most dissolved TCE leaves  the ground-water system by means of biotransformation.  The
estimated flux of TCE out of the system by th s process may be about an order of magnitude
greater than the flux of TCE discharged to Green Pond Brook. Although the estimated flux of
TCE out of the ground-water system by  volatilization is estimated to be about an  order of
magnitude less than the flux of TCE  discharge to Green  Pond Brook, volatilization is an
important mechanism for removing solutes and thereby affecting solute concentrations near the
water table.
                                                                                  151

-------
 References

 1.      Benioff, PA, M.H. Bhattacharyya, C. Biang, S.Y. Chiu, S. Miller, T. Fatten, D. Pearl,
        A.  Yonk,  and C.R. Yuen. 1 990.  Remedial investigation concept plan for Picatinny
        Arsenal, Vol. 2. Descriptions  of and sampling  plans for  remedial investigation sites.
        Argonne,  IL: Argonne National Laboratory, Environmental Assessment and Information
        Sciences Division, pp. 22-1 to 22-24.

 2.      Fusillo, T.V., T.A. Ehlke, M. Martin, and B.P. Sargent. 1 987. Movement and fate of
        chlorinated solvents in ground water: Preliminary results and future research plans. In:
        Franks, B.J., ed. Proceedings  of  the  U.S.  Geological  Survey Program on Toxic
        Waste—Ground-Water Contamination, Pensacola, FL (March 23-27). U.S. Geological
        Survey Open File Rep. 87-109. pp. D5-D12.

 3.      Sargent, B.P., J.W. Green, P.T. Harte, and E.F. Vowinkel. 1 986. Ground-water-quality
        data for Picatinny Arsenal, New Jersey,  1 958-85. U.S. Geological Survey Open File
        Rep. 86-58. 66 pp.                                                    ;

 4.      Imbrigiotta, I.E., M. Martin, B.P. Sargent, and L.M. Voronin. 1 989. Preliminary results
        of a study of the chemistry of  ground water at the Building 24 research site, Picatinny
        Arsenal, New Jersey. In: Mallard, G.E., and S.E. Ragone, eds. Proceedings of the U.S.
        Geological  Survey Toxic Substances Hydrology  Program, Phoenix, AZ (September
        26-30, 1988).  Water Res. Invest. Rep. 88-4220. pp. 351-359.

 5.      Imbrigiotta, T.E., T.A. Ehlke, and M. Martin.  1991. Chemical evidence of processes
        affecting the fate and transport of  chlorinated solvents in ground water at  Picatinny
        Arsenal, New Jersey.  In: Mallard, G.E., and D.A. Aronson,  eds. Proceedings of the U.S.
        Geological Survey Toxic Substances Hydrology Program, Monterey, CA (March 1 1 -1 5).
        Water Res. Invest. Rep.  91-4034. pp. 681-688.

 6.      Imbrigiotta, T.E., and M.Martin. 1 991. Overview of research activities on the movement
        and fate of chlorinated solvents in ground water at Picatinny Arsenal, New Jersey. In:
        Mallard, G.E., and D.A. Aronson, eds. Proceedings of the U.S. Geological Survey Toxic
        Substances Hydrology Program, Monterey, CA (March 1 1 -1 5). Water Res. Invest. Rep.
        91-4034. pp. 673-680.

 7.      Smith, J.A., C.T. Chiou, J.A. Kammer, and D.E. Kile. 1 990. Effect of soil moisture on
        sorption of trichloroethene vapor to vadose-zone soil at Picatinny Arsenal, New Jersey.
        Environ. Sci. Technol. 24(5):676-683.

 8.      Koller,  D.,  T.E.  Imbrigiotta,  A.L. Baer,  and  J.A.  Smith.  1994.  Desorption  of
       trichloroethylene  from  aquifer sediments at  Picatinny  Arsenal,  New Jersey.  In:
       Morganwalp, D.W., and D.A. Aronson, eds. Proceedings of the U.S. Geological Survey
       Toxic Substances Hydrology Program, Colorado Springs, CO (September 20-24,1 993).
       Water Res. Invest. Rep. 94-4014. In press.
152

-------
9.
Wilson, B.H., T.A.  Ehlke,  I.E.
dechlorination of trichloroethyiene in
New Jersey.  In:  Mallard,  G.E.,  and
Geological Survey Toxic Substances H
Water Res. Invest. Rep. 91-4034. pp.
10.
Imbrigiotta,  and  J.T. Wilson.  1991.  Reductive
   anoxic aquifer material from  Picatinny Arsenal,
    D.A. Aronson, eds.  Proceedings  of the  U.S.
   'drology Program, Monterey, CA (March 1 1-15).
   704-707.
Ehlke,  T.A.,  B.H.  Wilson, J.T.  Wilson,  and  T.E.  Imbrigiotta.  1994.  In  situ
biotransformation of trichloroethyiene and cis-1,2-dichloroethylene at Picatinny Arsenal,
New Jersey. In: Morganwalp, D.W., and D.A. Aronson, eds. Proceedings of the U.S.
       Geological Survey Toxic Substances
       (September 20-24, 1 993). Water Res
11.
                                    Hydrology Program, Colorado Springs,  CO
                                  .  Invest. Rep. 94-4014. In press.
Martin,  M. 1994.  Simulation of transport, desorption, volatilization, and  microbial
degradation of trichloroethyiene in ground water at Picatinny Arsenal, New Jersey. In:
Morganwalp, D.W., and D.A. Aronson|, eds. Proceedings of the U.S. Geological Survey
Toxic Substances Hydrology Program, Colorado Springs, CO (September20-24,1 993).
Water Res. Invest. Rep. 94-4014. In press.
                                                                                  153

-------
Intrinsic Bioremediation of TCE in Ground Water at an NPL Site in
St. Joseph, Michigan

John T. Wilson, James W. Weaver, and Don H. Kampbell
U.S. Environmental Protection Agency, Robert S. Kerr Environmental Research Laboratory,
Ada, OK
introduction

The ground water at the St. Joseph, Michigan, National Priority List (NPL) site is contaminated
with  chlorinated aliphatic compounds (CACs) at concentrations  in the  range of 10 mg/L to
100 mg/L. The chemicals are thought to have entered the shallow sandy aquifer either through
waste lagoons that were used from 1968 to 1976 or through disposal of trichloroethene (TCE)
into dry wells at the site (1). The contamination was determined to be divided into eastern and
western plumes, as the suspected sources were situated over a ground-water divide. Both plumes
were  found to contain TCE,  cis-  and trans-1,2-dichoroethene  (c-DCE and t-DCE), 1,1 -
dichloroethene (1,1-DCE), and vinyl chloride (VC).

Previous investigation  of the site indicated that natural anaerobic  degradation of the TCE was
occurring, because transformation products and significant levels of ethene and methane (2, 3)
were present. The purpose of this presentation is to present the results of later sampling  of the
western plume near Lake Michigan, to estimate the contaminant mass flux, and to estimate
apparent degradation constants. The  estimates  are based  on visualization of the data that
represent each measured concentration by  a zone of influence that is  based  on the sample
spacing. The presentation of the data is free from artifacts of interpolation, and extrapolation
of the data beyond the measurement locations is controlled.
Data Summary

In 1991, three transects (1, 2, and 3 on Figure 1) were completed near the source of the
western plume (2). The three transects consisted of 1 7 borings with a slotted auger. In 1 992,
two additional transects (4 and 5 on Figure 1) were completed, consisting of 9 additional slotted
auger borings. In each boring, water samples were taken at roughly 1.5 m (5 ft) depth intervals.
Onsite gas chromatography was performed to determine the width of the plume and find the
point of highest concentration. Three of the transects (2, 4, and 5) are roughly perpendicular
to the contaminant plume. Of the remaining transects, transect 1 crosses the plume at an angle
and transects lies along the length of the plume. The perpendiculartransects form logical units
for study of TCE biotransformation.
154
Symposium on Intrinsic Bioremediation of Ground Weter

-------
                                       Approx. Extent of TCE Plume

                                                  II
                                                             St Joseph, Michigan
                                                                NPL  Site
measurement point. The blocks are defined
Figure 1. Site plan, St. Joseph, Michigan, NPL site.
The site data from the transects are visualized as sets of blocks that are centered around the
so that the influence  of a particular measured
concentration extends halfway to the next measurement location both horizontally and vertically.
Thus, presentation of the data is simple and direct. The visualization of the data is performed
on a Silicon Graphics Indigo workstation using a two-dimensional version  of the fully three-
dimensional field-data analysis program calle<
S. Kerr Environmental Research Laboratory.

The mass of each chemical per unit thickness
calculated by summing over the blocks. By f
concentrations are not extrapolated into the c
d SITE-3Dthat is under development at the Robert
and the advective mass flux of each chemical are
 llowing this procedure, the measured chemical
ay layer under the site, nor are they extrapolated
beyond a short distance from the measurement locations (5 ft vertically and 50 ft to TOO ft
horizontally). Other interpolation schemes such as inverse distance weighting or kriging could
also be used to estimate the concentration fie
3 show the distributions of the VC and TCE a
"color" scale. Notably, the maximum VC con
transect 5 was 205,Mg/L. The maximum TCE
d and perform the mass estimates. Figures 2 and
 transect 5 using a logarithmic, black-and-white
centration at transect 4 was  1,660 jug/L and at
concentration at transect 4 was 8,720/ig/L and
at transect 5 was  163 ^g/L. As  noted  previously for other portions of the site (2, 4), the
contamination is found near the bottom of the aquifer. The highest concentrations of VC and
TCE do  not  appear to be  co-located. In
perpendicular transects ordered from farthesn
(transect 5). The data in Table 1  represent tre mass in a volume of aquifer that has an area
equal to the cross-sectional area of the transect and is 1.0-m thick in the direction of ground-
water flow.
able  1,  mass estimates are  presented  for the
upgradient (transect 2) to farthest downgradient
                                                                                   155

-------
                            St. Joseph, Michigan
                            Vinyl Chloride
                            transect: 5
                            mass:   0.4811 E-01 Kg/m
                                      151     tS2
                                                        tS4    tS3  tSS
                                   Ground surfica
                         10faet,
                                    approx. N
                                                     100 feet
Concentration
    ug/L

  250000.1

  25000.



   2500.



   250.0



   25.00
                                                                                2.500
                                                                               0.2500
                                                                               0.0250
Figure 2. VC distribution at transect 5.
                             St. Joseph, Michigan
                             Trichloroethene
                             transect: 5
                             mass:   0.2821 E-01 Kg/m
                                       tS1     152
                                   Ground surface
                         10 feet
                                    approx. N
                                                        154     153  tSS
                                                      100 feet
Concentration
    ug/L

  250000.

   25000.


    2500.


    250.0


    25.00
                                                                                2.500
                                                                               0.2500
                                                                               0.0250
Figure 3. TCE distribution at transect 5.
156

-------
Table 1. Mass per Unit Thickness (kg/m) at St. Joseph, Michigan
Chemical
Vinyl chloride
1,1 -DCE
t-DCE
c-DCE
TCE
Methane
Ethene
Ethane
TOC
Chloride
Sulfate
NO3-Nitrogen
NH4-Nitrogen
TKN-Nitrogen

2
1.523
0.2377
0.566
12.32
10.67
5.855
0.6847
No data
No data
129.9
37.05
2.904
1.835
2.987

1
Transect
1
8969
o|()816
0
5
5
5
5059
1127
5804
4826
OJ8925
N|D data
N
148
34
2
> data
8
376
471
2J5609
3
8357
4
0.4868
0.01451
0.03628
1 .890
1.397
4.620
0.1747
0.2085
12.63
213.1
95.78
4.421
0.4562
0.6353
5
0.04811
0.001047
0.007041
0.2832
0.02821
1.373
0.004901
0.001689
8.314
156.2
66.19
8.247
0.2256
0.3646
Advective Mass Flux Estimates

Results from the calibrated MODFLOW mocel of Tiedeman and Gorelick (4) were used to
estimate the ground-water flow velocity at each transect.  The estimate is an upper bound
because the modeled vertical component of flow was neglected in the present analysis. The
head drop from  one location to the next was assumed to generate horizontal flow  only.
Tiedeman and Gorelick (4) also represented the aquifer by single values of hydraulic conductivity
and porosity. They gave, however, 95 percent confidence limits for the hydraulic conductivity.
Well yields estimated for each sample location indicate declining hydraulic conductivity toward
the west (i.e., towards Lake Michigan and tramsects 4 and 5). Thus, using the single parameter
values from the MODFLOW simulations may overestimate the flux of water into the lake.

As would be expected, the advective mass fluxes decline toward the downgradient edge of the
plume (Table 2). There the concentrations are lower, due to either transient flow or degradation
of the TCE. Notably the mass fluxes using the average hydraulic conductivity result in a total flux
of 1 3 kg/y of TCE, c-DCE, t-DCE, 1,1 -DCE, and VC at transect 5. This value contrasts with the
total flux of these CACs of  310 kg/y at transec 12 near the source of contamination. Thus, there
 is a 24.4-fold decrease in mass flux of CACs
across the site. Using the 95 percent confidence
                                                                                   157

-------
 limits on the hydraulic conductivity determined by Tiedeman and Gorelick (4), the range total
 of mass flux of these five chemicals ranges from 205 kg/y to 420 kg/y at transect 2 and from
 8.4 kg/y to 1 7 kg/y at transect 5. The range of fluxes at transect 5  is an upper bound on, and
 best estimate of, the flux into Lake Michigan.
Table 2. Mass Flux (kg/y) at St Joseph, Michigan
Chemical
Vinyl chloride
1,1 -DCE
t-DCE
c-DCE
TCE
Methane
Ethene
Ethane
TOC
Chloride
Sulfate
NO3-Nitrogen
NH4-Nitrogen
TKN-Nitrogen
Transect
2
18.81
2.934
6.995
152.1
131.7
72.29
8.453
No data
No data
1604
457.4
35.85
22.66
36.88
1
36.03
1.551
9.609
97.11
106.0
104.1
16.95
No data
No data
2826
652.9
46.93
48.64
72.85
4
10.69
0.3185
0.7963
41.48
30.67
101.4
3.836
4.577
277.2
4678
2102
97.05
10.01
13.95
5
1.676
0.03648
0.2453
9.868
0.9829
47.86
0.1708
0.05885
289.7
5444 •
2306
287.4
7.861
12.70
Apparent Degradation  Constants

The mass per unit thickness of TCE at transects 2, 4, and 5 was used to estimate apparent first-
order degradation constants.  The constants are estimated by applying the first-order rate
equation
                                In

=   XAt
(1)
to the site data, where cf is the average concentration in the transect j, c)+1 is the average
concentration in the downgradient transect j+1, At is the advective travel time for TCE to move
158

-------
between the transects, and A is the apparent cegradation constant. The mass per unit thickness
data forTCE and the cross-sectional area were used to determine the average concentrations
G| and ci+1 in the up- and downgradient transects. The porosity, bulk density, fraction organic
carbon, organic carbon partition coefficient (5), ground-water gradient, and distance between
the transects were used to determine the advective travel times.  The values used in equation 1
are given in Table 3.  From these quantities, [the apparent degradation constant for TCE was
determined to  be -0.0076/wk from transect 2 to 4 and -0.024/wk from transect 4 to 5.
Table 3. Chemical and Hydraulic Values Us« d in Estimating Apparent Degradation Rates



Tran-
sect
2

4

5

Area with
nonzero
TCE
concen-
tration
(m2)
1592

2774

1943



Mass per
unit thickness
from SITE-3D
(kg/m)
10.67

1.397

0.0282

Average TCI
concentratio
in the transe
(kg/m3)
Cj and Cj+1 in
equation 1
6.70e-3

5.04e-4

1 .44e-5

i
ct









Distance
between
transects
(m)

260

160


Gradient
estimated
from
Tiedeman
and
Gorelick
(1993)

7.3e-3

l.le-2



Retarded
seepage
velocity
for TCE°
(m/d)

0.11

0.156

Estimated
travel time
between
transects
(weeks)
At in
equation 1

340

145

°Constants used in seepage velocity calculation:
Hydraulic conductivity: 7.5 m/d
Retardation factor for TCE: 1.78 = 1 + K.xiocpb/e
Porosity, 0: 0.30
Bulk density pb:  1.86 g/cm3
K^: 126 mL/g, foc: 0.001
References
i.
2.
Engineering Science, Inc. 1 990. Reme
Ml, phase 1 technical memorandum.
dial investigation and feasibility study, St. Joseph,
.iverpool, NY.
Kitanidis, P.K., L Semprini, D.H. Kampbell, and J.T. Wilson. 1993. Natural anaerobic
bioremediation of TCE  at the St. Joseph, Michigan, Superfund site. In: U.S. EPA.
Symposium on bioremediation of hazardous wastes:  Research, development, and field
evaluations. EPA/600/R-93/054. Washington, DC (May), pp. 57-60.
3.
McCarty, P.L., and J.T. Wilson. 1 992.
Joseph, Michigan,  NPL site. In:  U.
EPA/600/R-92/126. pp. 47-50.
Natural anaerobic treatment of a TCE plume, St.
S.  EPA. Bioremediation of hazardous wastes.
                                                                                   159

-------
4.     Tiedeman, C., and S. Gorelick. 1 993. Analysis of uncertainty in optimal ground-water
       contaminant capture design. Water Resour. Res. 29(7):2,139-2,1 53.

5.     U.S. EPA. 1 990. Subsurface remediation guidance table 3.  EPA/540/2-90/01 1 b.
160

-------
Poster
Session

-------

-------
Technical Protocol for Implementing the
                      Intrinsic Remediation  With Long-Term
Monitoring Option  for  Natural Attenuation  of  Fuel-Hydrocarbon Contamination in
Ground  Water
Todd H. Wiedemeier
Engineering-Science, Inc., Denver, CO

John T. Wilson and Donald H. Kampbell
U.S. Environmental Protection Agency, Rober
Ada, OK

Ross N. Miller and Jerry E. Hansen
U.S. Air Force Center for Environmental Exce
Brooks AFB, TX
                      S. Kerr Environmental Research Laboratory,
                      lence, Technology Transfer Division,
This paper presents a brief overview of the technical protocol, currently under development by
the U.S. Air Force Center for Environmental Excellence (AFCEE), Technology Transfer Division,
for data collection, ground-water modeling,
                     and exposure assessment in support of intrinsic
contaminated ground water (1). The material
effort of AFCEE,  the
Agency's  Robert  S.
Engineering-Science,  Inc.,  to  facilitate
remediation (natural attenuation) with long-term monitoring for restoration of fuel-hydrocarbon
                     presented herein was prepared through the joint
Bioremediation Research Team at the U.S.  Environmental  Protection
Kerr Environmental  Research Laboratory in Ada, Oklahoma, and
of dissolved-phase fuel hydrocarbons having
The  intended  audience for this document
consultants, regulatory  personnel, and  othe
contaminated ground water at USAF facilitie:
                     mplementation  of  intrinsic  remediation  at
fuel-hydrocarbon contaminated sites. Specifically, this protocol is designed to evaluate the fate
                     regulatory maximum contaminant levels (MCLs).
                     s U.S. Air Force (USAF)  personnel, scientists,
                     rs charged with remediating fuel-hydrocarbon
Intrinsic remediation is achieved when
biodegradation (aerobic and  anaerobic),
contaminantdissolved in ground water. During
transformed to innocuous byproducts (e.g.,
another phase or location  within the
integration of several subsurface attenuation
or  nondestructive.  Destructive processes
hydrolysis. Nondestructive  attenuation
dispersion and infiltration), and volatilization
                    b ing
In some cases, intrinsic remediation reduces
below MCLs before the contaminant plume
naturally occurring attenuation mechanisms, such as
         about a  reduction in the total  mass  of a
     intrinsic remediation, contaminants are ultimately
    carbon dioxide and water), not just transferred to
environment. Intrinsic remediation  results  from the
    mechanisms that are classified as either destructive
    include biodegradation, abiotic  oxidation,  and
             include  sorption, dilution (caused  by
                 mechanisms
                      dissolved-phase contaminant concentrations to
                      reaches potential receptors, even if little or no
source removal or reduction takes place. In situations where intrinsic remediation will not reduce
contaminant concentrations to below regulatory MCLs, in an acceptable time frame, less
stringent cleanup goals may be implemented.
that intrinsic remediation will result in a continual reduction in contaminant concentrations over
time such that calculated risk values are reduced.
                      This is especially likely if it can be demonstrated
Symposium on Intrinsic Bioremediation of Ground Water
                                                               163

-------
Intrinsic remediation is gaining regulatory acceptance and has been implemented at several sites
over the past few  years  (2-5). In addition  to  bringing about complete mineralization  of
contaminants, intrinsic remediation is nonintrusive and allows continuing use of infrastructure
during remediation. The main limitation is that intrinsic remediation is subject to natural and
institutionally  induced  changes  in  local  hydrogeologic  conditions.  In  addition,  aquifer
heterogeneity may complicate site characterization as it will with any remedial technology.

Evaluating the effectiveness of intrinsic remediation requires the quantification  of ground-water
flow and solute transport and transformation processes, including rates of natural attenuation.
Quantification of contaminant migration and attenuation rates, and successful implementation
of the intrinsic remediation option, requires completion of the following steps,  each  of which is
discussed in the following sections and outlined in  Figure 1:

        1.      Review existing site data.

        2.      Develop a preliminary conceptual model for the site, and assess the potential
               significance of intrinsic remediation.

        3.      Perform site characterization in support of intrinsic remediation.

        4.      Refine the conceptual model based  on  site characterization data, complete
               premodeling calculations, and document indicators of intrinsic remediation.

        5.      Model intrinsic remediation using numerical fate and transport models that allow
               incorporation of a  biodegradation term (e.g.,  Bioplume II or Bioplume 111).

        6.      Conduct an exposure assessment.

        7.      Prepare a long-term monitoring plan, long-term monitoring wells atthe site, and
               point-of compliance wells.

        8.      Present findings to regulatory agencies, and obtain approval for the  intrinsic
               remediation with long-term monitoring option.

Collection of an adequate database  during  the iterative site characterization process is an
important step  in  the  documentation  of  intrinsic  remediation. At  a minimum,  the site
characterization phase should provide data on the location and extent of contaminant sources;
data on the location, extent, and concentration of dissolved-phase contamination; ground-water
geochemical data;  geologic  data on the type and distribution  of subsurface materials; and
hydrogeologic  parameters such as hydraulic conductivity, hydraulic gradients, and potential
contaminant migration pathwaysto human orecological receptors. Contaminant sources include
nonaqueous-phase liquid (NAPL) hydrocarbons present as  mobile NAPL (NAPL occurring  at
sufficiently high saturations to drain, under the influence of gravity, to a well) or residual NAPL
(NAPL occurring at immobile  residual saturations that are unable to drain to a well  by,gravity).
164

-------




Review Available >
Site Data
\
f


Develop Preliminary
Conceptual Model
Jr
Make Preliminary
Assessment of Potential
For Intrinsic Remediation
Based on Existing Site
Characterization Data
- Contaminant Type
and Distribution
- Hydrogaology
- Location of Receptors
'^
f
Perform Site Characterization
in Support of Intrinsic Remediation
>
f
Refine Conceptual Model and
Complete Pre-Modeling
Calculations
\
r
Document Occurrence of
Intrinsic Remediation and
Model Intrinsic Remediation
Using Numerical Models
>
f
Use Results of Modeling and
Site-Specific Information in
an Exposure Assessment
^'"unacceptable Risk Tn^YES





Ev<
(
S Con
Intrins

Free Product
Recovery j
/
Pump
&
Treat

iuat
erR
)ptic
unc
cR
' h

[Bioslurpingj 	 '
t 15-°! Assess Potential For
emfdial ^ Intrinsic RfimfiHiatinn
;i'3n"',. ' With Remediation
£«&, s^m lnstalled
f^V >
r
I \ | Bioventing | Refine Conceptual Model and
1 \ Complete Pre-Modeling
1 \ Calculations
I 1 Sparging!
I1- 	 ' >
1 Bairiei-6 I Model Intrins
bprSlL .OptionSe,




>
r
c Remediation
irith Remedial
acted Above
3rical Models
r
Use Results of Modeling and
Site-Specific Information in
an Exposure Assessment



^<
YES ^x^UnacceDt
^xsPotential
NO
r
rhere"*-^
able Risk To""^.
Receptors'x^^

Site Point-pf-Compliance
Monitoring Wells and
Prepare, Long-Term
Monitorinq Plan


u,


Present Findings
and Lbng-Term FIGURE 2. 1
Monitoring Plan To
Regulatory Agencies . . ,. .
and RearJh Agreement Intrinsic Remediation
on Monitdring Strategy Flow Chart




Figure 1. Intrinsic remediation flow chart.
                                                                                       165

-------
The following analytical protocol should be used for analysis of soil and ground-water samples.
This  analytical protocol includes  all  of  the  parameters necessary to document  intrinsic
remediation of fuel hydrocarbons, including the effects of sorption and biodegradation (aerobic
and anaerobic) of fuel hydrocarbons. Soil samples should be analyzed for total volatile and
extractable hydrocarbons, aromatic hydrocarbons, and total organic carbon. Ground-water
samples  should  be  analyzed for  dissolved oxygen,  oxidation-reduction  potential,  pH,
temperature, conductivity, alkalinity, nitrate,  sulfate, sulfide, ferrous  iron,  carbon dioxide,
methane, chloride, total petroleum hydrocarbons, and aromatic hydrocarbons. The extent and
distribution (vertical and horizontal) of contamination  and electron  acceptor and  metabolic
byproduct concentrations and  distributions are of paramount importance in documenting the
occurrence of biodegradation  of fuel hydrocarbons and in numerical model implementation.
Dissolved oxygen concentrations  below  background in an area  with  fuel-hydrocarbon
contamination are indicative of aerobic hydrocarbon  biodegradation. Similarly, nitrate and
sulfate concentrations below background in an area with fuel-hydrocarbon contamination are
indicative of anaerobic biodegradation through denitrification and sulfanogenesis. Contour maps
can be used to provide visible evidence  of these relationships.  Elevated concentrations of
metabolic  byproducts  in areas  with fuel-hydrocarbon contamination  are  indicative  of
hydrocarbon biodegradation. As iron II and methane concentrations increase during iron (III)
reduction and methanogenesis (anaerobic processes), BTEX concentrations should be seen to
decrease. Contour maps can be used to provide visible evidence of these relationships.

To support  implementation  of intrinsic remediation, the property  owner must scientifically
demonstrate that degradation of site contaminants is occurring at rates sufficientto be protective
of human health and the environment. Three lines of evidence can be used to support intrinsic
remediation: 1) documented loss of contaminants at the  field scale, 2) the use of chemical
analytical data in mass  balance calculations of  microbial  metabolism, and 3) laboratory
microcosm studies using aquifer samples collected from the site.

The first line of evidence involves using measured dissolved-phase concentrations of biologically
recalcitrant tracers found in fuels in conjunction with aquifer hydrogeologic parameters such as
seepage velocity  and dilution  to show that a  reduction in the total mass of contaminants is
occurring at the site. The second line of evidence involves the use of chemical analytical data
in mass balance calculations to show that a decrease in contaminant and electron acceptor
concentrations can be directly  correlated to increases in metabolic byproduct concentrations.
This evidence can be used  to show that  electron acceptor concentrations are sufficient to
degrade dissolved-phase contaminants. Numerical models can be used to aid mass-balance
calculations and to collate information  on  degradation. The third  line  of evidence,  the
microcosm study, involves studying  site aquifer materials under controlled conditions in the
laboratory to show that indigenous biota are capable of degrading site contaminants and to
confirm rates of contaminant degradation  measured at the field scale.

The primary objective of the  intrinsic remediation investigation is to determine  if natural
processes of degradation will  reduce contaminant concentrations in  ground water to below
regulatory standards before  potential exposure pathways are completed. This requires that a
projection of the potential extentand concentration  of the contaminant plume in time and space
is made  based on  governing  physical, chemical, and biological processes. This projection
should be based on historic variations in—and the current extent and concentration of—the
contaminant plume, as well as on the measured rates of contaminant attenuation.
166

-------
modeling the combined effects of advection,
assess the possible risk to potential downgrad
The data collected during site characterization can be used to model the fate and transport of
contaminants in the subsurface. Such mode ing  allows an estimate of the future extent and
concentration of the dissolved-phase plume to be made. Several models, including Bioplume
II (6),  have been used  successfully to model dissolved-phase contaminant transport and
attenuation. Additionally, a new version o  the Bioplume  model,  Bioplume  III,  is  under
development by AFCEE. The intrinsic remediation modeling effort has three primary objectives:
1) to estimate the future extent and concentre tion of a dissolved-phase contaminant plume by
                                          dispersion, sorption, and biodegradation; 2) to
                                          ent receptors; and 3) to provide technical support
for the natural attenuation remedial option at postmodeling regulatory negotiations.

Microorganisms generally utilize dissolved oxygen and nitrate in areas with dissolved-phase
fuel-hydrocarbon contamination at rates that a re instantaneous relative to the average advective
transport velocity of ground water. This results in the consumption of these compounds at a rate
approximately equal to the rate at which they are replenished by advective flow processes. For
this reason, the use of these compounds as electron acceptors in the biodegradation of
dissolved-phase fuel hydrocarbons is  a  mass-transport-limited process (7, 8). The  use of
dissolved oxygen and nitrate in the biodegradation of dissolved-phase fuel  hydrocarbons can
be modeled using Bioplume II. Microorganisms generally utilize sulfate, iron III, and carbon
dioxide (used  during  methanogenesis)  in  areas  with  dissolved-phase  fuel-hydrocarbon
contamination at rates that are slow relative to the advective transport velocity of ground water.
This results in the consumption of these comppunds at a rate slower than the rate at which they
are replenished by advective flow processes. Therefore, the use of these compounds as electron
acceptors in the biodegradation of dissolved-phase fuel  hydrocarbons is  a reaction-limited
process that can be approximated by first-order kinetics. The  Bioplume II model  utilizes a
first-order rate constant to model  such biocegradation.  First-order  decay constants can be
determined by simple calculations based  on ground-water chemistry or through the use of
laboratory microcosm studies. In addition, tre use of radiolabeled materials in a microcosm
study  can be used to provide evidence of the ultimate fate of the contaminants.

The results of the modeling effort are not in themselves sufficient proof that intrinsic remediation
is occurring at a given site. The results of the
data input into the model and the model itse
                                          modeling effort are only as good as the original
                                          f. Because of the  inherent uncertainty associated
with such predictions, it is the responsibility cf the proponent to provide sufficient evidence to
demonstrate thatthe mechanisms of intrinsic remediation will reduce contaminant concentrations
to acceptable levels before potential receptors
input parameters and numerous sensitivity anc
contaminant migration scenarios. When poss
used to provide information that collectively a
                                          are reached. This requires the use of conservative
                                          lyses so that consideration is given to all plausible
                                          ble, both historical data and modeling should be
                                          id consistently supports the natural reduction and
        of the dissolved-phase contaminant plume. In some cases, simple calculations of
contaminant  attenuation rates are all  thai  are  required  to successfully  support intrinsic
remediation.
removal
Upon completion of the fate and transport rrodeling effort, model predictions can be used in
an exposure assessment. If intrinsic remediation is sufficiently active to mitigate risks to potential
receptors, the proponent of intrinsic remediation has a reasonable basis for negotiating this
option with regulators. The exposure assessrient allows the proponent to show that potential
exposure pathways will not be completed.
                                                                                    167

-------
The  long-term  monitoring  plan consists of locating  ground-water monitoring  wells and
developing a ground-water sampling and analysis strategy. This plan is used to monitor plume
migration over time and to verify that intrinsic remediation  is occurring at rates sufficient to
protect potential downgradient receptors. The long-term monitoring plan should be developed
based on the results of a numerical model such as Bioplume II.

Point-of-compliance  (POC)  monitoring wells  are  wells  that are  installed  at  locations
downgradient of the contaminant plume and upgradient of potential receptors. POC monitoring
wells are generally installed along a  property boundary or  at a location approximately 5 yr
downgradient of the current plume at the seepage velocity of the ground water of 1 yr to 2 yr
upgradient of the nearest downgradient receptor, whichever is more protective. The final number
and location of POC monitoring wells depends on regulatory considerations.

Long-term monitoring wells are  wells that are placed upgradient of,  within, and immediately
downgradient of the contaminant plume. These wells are  used to monitor the effectiveness of
intrinsic remediation  in reducing the total mass  of contaminant within the plume. The final
numberand location of long-term monitoring wells depends on regulatory considerations. Figure
2 shows a hypothetical long-term monitoring scenario. The results of  a numerical model such
as Bioplume II can be used to help site both the long-term and POC monitoring wells.
                                            Anaerobic Treatment Zone
                          Plume Migration
                                                         Extent of Dissolved-
                                                         Phase BTEX Plume
                                                        Aerobic Treatment
                                                        Zone
            LEGEND

  8 Point-of-Compliance Monitoring Well
  O Long-Term Monitoring Well
Not To Scale
     FIGURE 2
Hypothetical Long-Term
  Monitoring Strategy
Figure 2. Hypothetical long-term monitoring strategy.
168

-------
References
i.
2.
3.
4.
5.
6.
7.
8.
                                      "OCQ
Wiedemeier, T.H., D.C. Downey, J.T
Hansen. 1994. Draft technical prot<
long-term  monitoring  option  for
contamination in ground water. U.S.
Technology Transfer Division.
         Wilson,  D.H.  Kampbell,  R.N. Miller, and J.E.
         for implementing the intrinsic remediation with
        natural attenuation   of  dissolved-phase  fuel
        Air Force Center for Environmental Excellence,
Klecka, G.M., J.W. Davis, D.R. Gray, and S.S. Madsen. 1 990. Natural bioremediation
of organic contaminants in ground  water—Cliffs-Dow Superfund site. Ground Water
28(4):534-543.
Downey, D.C., and  M.J.  Gier.  1991
hydrocarbon  spill  site.  In:  Proceec
Technology Symposium, San Antonio
                                        dt
Wiedemeier, T.H., P.R. Guest, R.L. H«
to support regulatory negotiations
Proceedings of the Petroleum
Prevention, Detection, and Restoratio i
        nry, and C.B. Keith. 1 993. The use of Bioplume
         a  fuel spill site  near  Denver,  Colorado. In:
Hydroccrbons and Organic Chemicals in Ground Water:
         Conference. NWWA/API. pp. 445-459.
Wiedemeier,  T.H.,  B.  Blicker,  and
bioremediation of fuel hydrocarbons
Federal Environmental  Restoration
Exhibition, New Orleans, LA.
60.
        P.R. Guest.  1994.  Risk-based  approach  to
       at a major airport. In: Proceedings of the 1994
           and Waste Minimization  Conference and
Hazardous Materials Control Resources Institute, pp. 51 -
Rifai,  H.S.,  P.B.  Bedient, J.T. Wils
Biodegradation modeling at aviatior
1,0029.

Borden, R.C., and P.B. Bedient. 1 986
by  oxygen limited  biodegradation
22(13):1973-1 982.

Wilson, J.T., J.F. McNabb, J.W. Coch
1 985. Influence of microbial adapta
water. Environ. Toxicol. Chem. 4:72
           Supporting the no-action  alternative at a
        ings  of  the  USAF Environmental Restoration
        , TX (May 7-8). Section U. pp. 1 -1 1.
        >n, K.M.  Miller,  and J.M.  Armstrong.  1988.
         fuel  spill site.  J. Environ. Eng. 1 14(5):1,007-
       . Transport of dissolved hydrocarbons influenced
       •theoretical  development. Water  Resour.  Res.
        ran, T.H. Want, M.B. Tomson, and P.B. Bedient.
        ion on the fate of organic pollutants in ground
        -726.
                                                                                  169

-------
 Wisconsin's Guidance on  Naturally Occurring Biodegradation as a
 Remedial Action Option

 Michael J. Barden
 Wisconsin Department of Natural Resources, Emergency and Remedial Response Section,
 Madison, Wl
In February 1 993, the Wisconsin Department of Natural Resources issued an interim guidance
on naturally occurring biodegradation as a remedial action option for contaminated sites. The
focus of this guidance was primarily on soil contamination by petroleum hydrocarbons and on
the requirements for site characterization and monitoring  necessary to use this approach.
Subsequent implementation of the interim guidance has resulted in further refinement to make
this an effective approach for both soil and ground-water contamination.

The application of naturally occurring biodegradation as a remedial action requires that the site
be adequately characterized and that an adequate monitoring program be developed and
implemented. This is a long-term remediation option, likely requiring years or decades to effect
adequate cleanup. From a regulatory  perspective, the primary concerns  are that  the site
conditions are amenable to naturally occurring biodegradation and that the process is effective
in reducing contaminant concentrations to acceptable levels within  a reasonable period with
respect to potential contaminant migration or impacts to receptors.

Adequate site characterization is required during the site investigation  so that naturally occurring
biodegradation can be evaluated with other possible remedial action options. This also provides
baseline information for the potential application of enhanced bioremediation as well, because
the basic site characterization requirements are essentially the same.  Characterization involves
identification of 1) the contaminants present and their concentrations and biodegradability, 2)
physical and chemical  parameters  affecting availability of oxygen  and alternative electron
acceptors, 3) nutrients, and 4) microbiological parameters indicating the presence and  viability
of appropriate microbial populations. A sufficient  number  of samples should be used to
representthe extent of contamination and site heterogeneity. The guidance provides a framework
for interpretation and evaluation of the results.

If site conditions  are favorable, a  monitoring plan must  be developed  and implemented.
Monitoring indicates that contaminant concentrations are decreasing  overtime, ensures that no
unexpected contaminant migration is occurring, and provides information regarding the nature
and rate of biodegradation at the site. A variety of monitoring approaches and techniques are
available  for soil  and  ground water.  In  general,  monitoring changes in contaminant
concentrations  and/or concentrations of co-reactants are appropriate.

Experience with implementation indicates that many responsible parties are unlikely to select this
option  due  to the long  time frame involved.  This suggests  that   naturally  occurring
biodegradation is more viable as an option for stable entities where time is not an issue. The
availability of the  guidance, however, has encouraged consideration of bioremediation  in
general as a viable remedy due to perceived  regulatory acceptance of the technology and
because the required  consideration of biodegradation  potential provides baseline site
information that can be used in  the evaluation and design of enhanced bioremediation systems.
170
                                                    Symposium on Intrinsic Bioremediation of Ground Water

-------
Assessing the Efficiency of Intrinsic Bioremediation
Francis H. Chapelle
U.S. Geological Survey, Columbia, SC
of biodegradation. If transport rates are fasf i
can migrate freely with ground-water flow anc
wildlife populations. Conversely, if transport
Abstract

The efficiency of intrinsic bioremediation to contain contaminant migration in ground-water
systems can be quantitatively assessed by conparing rates of contaminant transport with rates
elative to rates of biodegradation, contaminants
 possibly reach a point of contact with human or
rates are slow relative to biodegradation rates,
contaminant migration will be more confined and less likely to reach a point of contact. In either
case, the efficiency of intrinsic bioremediation can be assessed by evaluating the presence or
absence of contaminant transport to predetermined points of contact. Thus, this assessment
includes hydrologic (rates of ground-water flow), microbiologic (rates of biodegradation), and
sociopolitical (points of contact) considerations.
The U.S. Geological Survey, in cooperation wi
h the Naval Facilities Engineering Command, has
developed a framework for assessing the efficiency of intrinsic bioremediation that is based on
these three considerations.  In this  framework, hydrologic and microbiologic information is
synthesized using a solute-transport code (SUJTRA) and used to estimate rates of contaminant
transport to predefined points of contact (adjacent water supply wells or surface water bodies).
This framework is applied to two sites, in Beaufort and Hanahan, South Carolina, contaminated
with aviation fuel. At the Beaufort  site, rates  of  biodegradation are slow due to anaerobic
conditions (Kbio  ~0.01  d-1), but because rates of ground-water flow are low  (~0.02  ft/d),
soluble contaminants are effectively contained and are not transported to adjacent points of
contact. At the  Hanahan site, biodegradation rates are similarly slow  under the ambient
anaerobic conditions (Kbio ~0.01 d'1), but beqause rates of ground-water flow are relatively high
(~ 1.0 ft/d), contaminants are transported to multiple points of contact with humans. These
examples  illustrate the complex interplay that develops between hydrologic, microbiologic, and
sociopolitical considerations, and show that
be assessed on a site-by-site basis.
he efficiency of intrinsic bioremediation can only
 Symposium on Intrinsic Bioremediation of Ground Water
                                           171

-------
A Practical Approach  to  Evaluating Natural Attenuation of Contaminants in
Ground Water
 Paul M. McAllister and Chen Y. Chiang
 Shell  Development Company, Houston, TX
The extent of natural attenuation is an important consideration  in determining the most
appropriate corrective action at sites where ground-water quality has been affected by releases
of petroleum hydrocarbons or other chemicals. The objective of this presentation is to provide
guidelines for evaluating natural attenuation based on easily obtainable field and laboratory
data.

The primary indicators that can be used to evaluate natural attenuation include dissolved oxygen
(DO) levels  in ground water and plume characteristics. Background DO levels greater than 1
to 2 mg/L and inverse correlation between DO and soluble hydrocarbon concentration have
been identified through laboratory and field studies as key indicators of aerobic biodegradation.
Several unique plume characteristics include  1) plumes migrate more slowly than expected; 2)
plumes reach a steady state; and 3) plumes decrease in extent and concentration, which may
indicate the effects of natural attenuation.

When DO is depleted in  an aquifer, anaerobic conditions prevail. For biodegradation to occur,
an  alternative electron  acceptor such as  nitrate, carbonate,  or iron III must be available.
Between aerobic and anaerobic conditions  (i.e.,  0.1  ppm to 2  ppm), there  is a  region
sometimes labeled  the  hypoxic  zone.  Studies  in the  hypoxic zone  have indicated  that
biodegradation of benzene, toluene, ethylbenzene, and the xylenes (BTEX) may occur at relatively
low DO levels provided a secondary electron acceptor is available.

Other secondary indicators  (e.g., geochemical data)  and  more  intensive  methods  (e.g.,
contaminant mass balances, laboratory microcosm studies, and detailed ground-water modeling)
can be applied to demonstrate natural attenuation  as well. The  recommended approach for
evaluating natural attenuation is to design site assessment activities so that required data such
as DO  levels and historical plume flow path  concentrations are obtained. With the necessary
data, the primary indicators  should be applied to evaluate natural attenuation.
172
Symposium on Intrinsic Bioremediation of Ground Water

-------
The Use of Low Level  Activities
To Assist
Intrinsic Bioremediation
Robert D. Morris
Eckenfelder Inc., Nashville, TN

Jeffrey C. Dey
Resource Control Corporation, Rancogas, N.

Daniel P. Shine
Sun Company, Inc., Aston, PA
Intrinsic bioremediation as discussed in the recent report of the National Academy of Sciences
Committee on Bioremediation (1) can reduce the concentrations of some common contaminants
to levels generally considered protective of human health. Since the observations of the role of
biodegradation in limiting the extent of the ground-water plume at the Conroe, Texas, wood-
preserving site (2), many other sites have been
         observed to have undergone natural attenuation
at sufficient rates to limit the size of contaminart plumes; in several instances, contaminant levels
decreased to below cleanup levels.

The  use of intrinsic bioremediation, while generally attractive from a  cost perspective, may
actually be desirable environmentally because secondary effects  of active remediation are
avoided. Intrinsic bioremediation, however,  las several costs associated with its use. These
include some level of documentation that bio ogical degradation of constituents of concern is
taking place and  costs associated with monitoring and management of the site. Managing a
small plume, such as is frequently found at service stations, includes sampling and reporting to
the  responsible environmental agency. At one typical site (3), the annual  environmental
management costs have been approximately $1 1,900 per year, consisting of site visits and
sampling ($1,000), chemical analysis ($7,200), reporting ($2,400), and consulting ($1,300).
At other sites, costs have exceeded these values by 50 percent or more. Documenting that
biodegradation is occurring adds substantially  to these costs.

Intrinsic bioremediation will be effective where the electron acceptor requirements are relatively
small. While oxygen may reach the affected zones at rates sufficient  to prevent and shrink
contaminant  plumes, and thus eventually achieve remedial goals, the time  frame may be
unacceptably long from the site owner's perspective because of long-term monitoring costs and
management burdens.

Addition of appropriate electron acceptors would accelerate reduction  in constituents such as
monoaromatic hydrocarbons. In some  cases  where  intrinsic  bioremediation is technically
feasible, it may not be the most cost-effectivd approach. To evaluate the concept of applying
limited engineering solutions  at sites where  intrinsic  bioremediation  appears to be slowly
reducing the contaminant mass, we are testing the use of air sparging wells (3). The air sparging
wells are placed  immediately outside the plumes and operated intermittently at low flow. The
cost of installing three shallow  air sparging wells, routine maintenance,  limited additional
sampling, and reporting was budgeted at $8,500 per site for each of the three sites. If the time
to reach closure is shortened by 1  yr or more,
cost that would have been incurred by only nonitoring and managing the site.
         the cost of treatment will have been less than the
Symposium on Intrinsic Bioremediation of Ground Water
                                                   173

-------
Another approach to providing electron acceptors is to add aqueous solutions of hydrogen
peroxide or nitrate to wells located within or immediately upgradient of the plume; however,
nitrate addition has not been shown to be effective for degradation of benzene, and introduction
of hydrogen peroxide is  more labor intensive  and thus more costly than air sparging.
Alternatively, a slow-release oxygen compound such as magnesium peroxide can be placed in
wells and oxygen allowed  to diffuse into the formation.

Regulatory acceptance for natural attenuation may be more easily attained if a migration barrier
is created along the downgradientedge of the plume. Migration barriers can be created through
a series of low-flow air-sparging wells. Alternatively, a row of wells containing a slow-release
oxygen compound  can be placed perpendicular to the  ground-water  gradient near the
downgradientedge of the plume. As demonstrated by Bob Borden (4) of the University of North
Carolina at a commercial site in North Carolina and by Doug Mackay (5) of Stanford University
in tests conducted at the Borden Landfill in Canada, this approach can successfully prevent
migration of monoaromatic hydrocarbons.

Intrinsic bioremediation  should be.viewed as one approach  in a continuum of methods of
utilizing biodegradation processes to remediate soil and ground water. It should be used alone,
in combination with other approaches, or as a polishing step  based on evaluation of the site
conditions, regulatory issues, technical feasibility,  implementability, and cost.

This  poster session will present cost and analytical data from three New Jersey service stations
during the monitoring-only phase, site maps, diagrams of the air sparging systems that have
been installed, costs  of installation and operation, and analytical data available at the time of
the meeting. The cost of adding  hydrogen peroxide at two similar  sites will  be discussed.
Additionally, the costs and relative advantages of  the use of slow-release oxygen compounds
and  air sparging will be presented.

References
i.


2.


3.


4.


5.
National Research  Council.  1993. In situ  bioremediation: When  does it  work?
Washington, DC: National Academy Press.

Wilson, J.T., J.F. McNabb, J.  Cochran, T.H. Wang, M.B. Tomson, and  P.B. Bedient.
1 985. Influence of microbial  adaptation on the fate of organic pollutants in ground
water. Environ. Toxicol. Chem. 4:721-726.

Norris,  R.D., J.C.  Dey, and D.P. Shine.  1993. The advantages  of concerted
bioremediation  of lightly contaminated sites compared to intrinsic  bioremediation.
Presented at the American Chemical Society I&E Special Symposium, Atlanta, GA.

Kao, C.-M., and R.C. Borden. 1 994. Enhanced aerobic bioremediation of a gasoline
contaminated aquifer by oxygen-releasing barriers.  In: Hydrocarbon  bioremediation.
Boca  Raton, FL: Lewis Publishers.
Bianchi-Mosquera,  G.C.,  R.M. Allen-King,  and  D.D.  Mackay.
degradation  of  dissolved  benzene and toluene using a  solid
compound. Ground Water Monitor. Rev. pp.  120-128.
1994.  Enhanced
 oxygen-releasing
174

-------
Natural Attenuation of Jet Fuel in Ground  Wafer
Greg Doyle, Dwayne Graves, and Kandi Brovn
International Technologies Corporation, San
                                         Bernardino, CA
Natural attenuation is a minimum action remedial strategy that permits the biodegradation of
organic contaminants under  natural,  in  si u conditions. Mechanisms that  act to affect
contaminant biodegradation include aerobic b
                                          odegradation atthe plume boundary and various
anaerobic processes within the plume. Nitra e, iron, manganese, and sulfate reduction  and
                                          linant biodegradation. Naturally occurring levels
                                      mariganese,  and  carbonates  support  anaerobic
biodegradation.
methanogenesis are thought to support contami
of nitrate,  sulfate,  oxidized  iron  and  ma ic
Natural attenuation was proposed as a feasible remedial alternative for the dissolved jet fuel
plume at George Air Force Base in Victorvillel California. The plume, covering approximately
1.1 million square feet, was located in a poorly  yielding  perched aquifer, 120 ft below the
ground surface. Ground-water flow rate was approximately 20 tt/yr, and the edge of the plume
was about 2,000 ft from the property line. The ground water in the affected aquifer was not
being used. Because of subsurface conditions,
natural attenuation represented the most effic
dissolved jet fuel at this site.

A separate-phase layer of jet fuel floated on
aggressive skimming operation was employed
from the subsurface. Bioventing will be used to
in areas where jet fuel was  detected and
                                         slow migration, poor water yield, and lack of use,
                                         ent and cost-effective approach for remediating
                                          he water table near the center of the plume. An
                                          to remove all recoverable separate-phase jet fuel
                                          further remediate contaminated vadose zone soil
                                            d.
                                      remove^
Natural attenuation  for  aerobic and  aero Die/anaerobic conditions  was  modeled using
BIOPLUME II. The BIOPLUME II model simulates the transport of dissolved hydrocarbons under
the influence of oxygen-limited biodegradation. Using BIOPLUME II with site specific parameters,
various treatment scenarios were evaluated. Assuming that 60 percent of the separate-phase
product was removed, both anaerobic and aerobic biodegradation occurred, and residual jet
fuel diffused into the water based on Pick's Law of Diffusion, a remediation time of 42 yr was
determined. This treatment time was adequate to remediate the  plume before it migrated off site.
Based on this prediction and the cost savings
performance period was established prior to tr
efforts by the Air Force, the U.S. Environmental
Research Laboratory, and IT will provide data
                                          realized by applying natural attenuation, a 5-yr
                                         ie issuance of a finalized record of decision. Joint
                                         Protection Agency's Robert S. Kerr Environmental
                                         verifying the accuracy of the BIOPLUME II model
predictions and demonstrating the level of natural biological activity occurring in the ground
water. These efforts are expected to lead to regulatory acceptance of natural attenuation for the
full-scale remediation of the jet fuel plume on site.
Symposium on Intrinsic Bioremediation of Ground Water
                                                                                   175

-------
Evaluation of Intrinsic Bioremediation  at an Underground Storage Tank Site in
Northern  Utah
R. Ryan Dupont, Darwin L Sorensen, and Marion Kemblowski
Utah Water Research Laboratory, Utah State University, Logan, UT
A 2-yr field study was initiated  by the Utah  Water  Research Laboratory for the U.S.
Environmental  Protection  Agency's Office  of Underground Storage Tanks  at  two former
underground storage tank (UST) sites in northern Utah: a U.S. Air Force site at Hill Air Force
Base (AFB) near Ogden, Utah, and a private site in Layton, Utah. The study sought to evaluate
rapid site assessment techniques and data collection and summary methods that could be used
to provide a comprehensive description of the potential for intrinsic bioremediation at UST sites.
This poster presentation focuses on the Hill AFB site, where tank and line leaks from an 1 8,000
gal UST, removed in 1 989, were the probable sources of observed ground-water contamination.
Total dissolved petroleum hydrocarbon contaminant mass at the  beginning of the study was
estimated  to be 950,000  mg, while approximately 6,000 mg of benzene and 8,000  mg of
toluene were quantified in the plume underlying the site. The site was covered  with a permeable
gravel  surface  layer during the study, and dissolved  oxygen concentrations at a number of
locations in unconta mi noted areas surrounding the plume were over 2 mg/L.

Initial site assessment activities utilized cone penetrometry for rapid collection  of subsurface soil
stratigraphicdata and forthe placement of more than 60small-diameterground~watersarnpling
points at the two shallow field sites. Ambient temperature headspace analyses were conducted
using a polyethylene bag method developed by the Utah Water Research Laboratory in addition
to a commercial Lag-in-a-Bag apparatus  to provide rapid field-determined  measurements of
ground-watertotal hydrocarbons. Field total petroleum hydrocarbon (TPH) measurements were
collected on a near real-time basis to guide the initial placement of ground-water monitoring
points along and  perpendicular to the axis of the contaminant plume at each site. These field
methods proved that the original conceptual model of the nature and extent  of contamination
and of the potential contaminant migration pathways at the Hill AFB site, based on conventional
site assessment techniques  (soil  gas survey, collection  of  limited soil core samples, and
placement of more limited numbers of large-diameter ground-water  monitoring wejls), was
greatly  in error. The  model  could  be  improved  significantly with these cost-effective field
techniques, which are now widely available to the  consulting community.

Seven field sampling events, beginning in April 1 992, were conducted overthe 2-yr study period
as part of a proposed intrinsic bioremediation strategy developed in the project. Ground-water
quality data were collected from the small-diameter sampling points and existing ground-water
monitoring wells  to assess the distribution  and transport of contaminants, along with the
predominant microbial reactions taking place within the contaminant plume. Ground-water
quality data collected included field measurements of pH, dissolved oxygen, temperature, and
ambient headspace total  hydrocarbon concentrations; laboratory determinations included
nitrate-N,  sulfate,  dissolved  iron and manganese, total hydrocarbons (purge and trap and
semivolatile constituents), specific organic contaminants (C-6 to C-15 alkanes and benzene,
toluene, ethylbenzene,p-xylene, naphthalene, and methylnaphthalene), and contaminant boiling
point split concentrations.
176
Symposium on Intrinsic Bioremediation of Ground Water

-------
Ground-water data were used to generate tote11 and specific compound dissolved contaminant
and dissolved electron acceptor mass data for each sampling period. The location of the center
of the mass of contaminants and electron acceptors were evaluated, and changes  in these
parameters overtime, with respect to bulkgro
magnitude of natural degradation processes
jnd-waterflow, were used to assess the rate and
aking place at each site.
constituents displayed zero-order decay rates.
mg (>99.9 percent removal) by the end of
Over the 2-year study period, the mass of TPH showed exponential decay, while all specific
Dissolved TPH mass declined to less than 1,000
the 630-d monitoring period, while dissolved
benzene and toluene mass remaining in the contaminant plume declined to less than 200 mg
(97 percent removal). TPH, benzene, and toluene mass decay rates were found to be -0.01 3/d
(p = 0.03, r2 = 0.933), -11.2 mg benzene/d jp = 0.01, r2 = 0.976), and -14.2 mg toluene/d
(p = 0.01, r2  = 0.998), respectively. Mass center data indicate that  while  ground-water
velocities  at the site through  the study period averaged 0.45  ft/d, the  net  movement of
contaminants was attenuated significantly, with measured center of mass  velocities of TPH,
benzene, and toluene being 0.03 ft/d, 0.05 ft/d, and 0.07 ft/d,  respectively. Corresponding
utilization  of oxygen and  other terminal electrDn acceptors occurred across the  plume.
Intrinsic bioremediation of the dissolved plum
removed the residual hydrocarbon mass exis
 at the Hill AFB site successfully attenuated and
ing at the site at the beginning of the study. In
January 1 994, only 1 of 34 monitoring well/piezometer ground-water samples contained a
benzene concentration (20.7 /Jg/L) above regulatory concern, and the site is expected to be
eligible for closure at the next routine, semiannual sampling event.
This postersession will detail the physical/chemical characteristics of the field sites and the rapid
site assessment techniques and  typical results collected, as  well  as highlight the  data
collection/reduction/interpretation methodology developed in this study. Finally, more complete
results demonstrating natural degradation o
presented, along with  a summary of a natu
 hydrocarbon contaminants  at this site will be
ral attenuation decision support system to  aid
investigators in assessing the viability of intrnsic bioremediation for the selection of  a  "no
action'/natural attenuation monitoring alternctive at their sites.
                                                                                   177

-------
Case Studies of Field Sites To Demonstrate Natural Attenuation of BTEX
Compounds in Ground Water

Chen Y. Chiang and Paul M. McAllister
Shell Development Company, Houston, TX
The most definitive indicptors of natural attenuation such as plume characteristics and dissolved
oxygen  (DO) concentrations  are based on actual concentrations obtained during periodic
monitoring events. Based on appropriate data from monitoring wells, the following parameters
can be used to indicate and demonstrate that natural attenuation is occurring: 1) the mass of
benzene, toluene, ethylbenzene,  and the xylenes (BTEX) present and 2) the extent and  rate of
migration and distribution of  BTEX concentrations. Data collected from two field sites  will be
used to demonstrate natural attenuation mechanisms.

The first site is characterized by 42 monitoring wells to show the relationship between soluble
BTEX  and DO  plumes. Results from 10 sampling periods over 3 years show a significant
reduction in total BTEX mass with time in ground water. These reduction and leakage rates from
sources are determined from material balance and nonlinear least-squares analyses. The natural
attenuation rate is calculated to be 0.95 percent/d. Spatial relationships between  DO and total
BTEX are shown to be strongly correlated by statistical analyses and solute transport modeling.
In addition, laboratory  microcosm biodegradation experiments are performed  to determine
possible threshold limits for aromatic hydrocarbon oxidation under varying levels of DO. The
results are remarkably consistent with field data on the presence of high or low levels of BTEEX
and DO in several monitoring well-water samples.

The second site data will be used to demonstrate natural attenuation from a cost-effectiveness
perspective through evaluation of plume characteristics overtime. The benzene concentrations
along the primary flow path at this site are observed to decrease from 2,600 ppb at the  source
to 2.7 ppb at a distance 1,425 ft downgradient. The decrease in concentrations  with distance
from the source is a direct indication that some degree of natural attenuation is occurring. If no
natural attenuation was  occurring, then concentrations would  remain relatively constant out to
the leading edge, where a sharp front would be observed. It is emphasized that natural
attenuation  also   includes  other mechanisms than  biodegradation: dispersion, sorption,
volatilization, and  chemical transformation.
178
Symposium on Intrinsic Bioremediation of Ground Water

-------
Demonstrating Intrinsic Bioremediation of BTEX at a Natural  Gas Plant
Keith Piontek and Tom Sale
CH2M Hill, St. Louis, MO, and Denver, CO

Steve de Albuquerque and John Cruze
Phillips Petroleum Company, Bellaire, TX, an
J Bartlesville, OK
Intrinsic bioremediation is being characterizqd
plant operations  resulted in the release  of
deposits  beneath  the site.  Ground  water
ethylbenzene,  and xylene (BTEX) constituen
biodegradation that occurs under natural con
realistic assessment  of potential risks posed
management decision-making.
  at the site of a former natural gas plant. Gas
nonaqueous-phase liquid  (NAPL) to the eolian
 beneath the site  contains  benzene, toluene,
•s of the NAPL.  The  rate and  extent of BTEX
Jitions is being characterized to provide for more
  by  the  site,  and to support  long-term  site
The characterization of intrinsic bioremediaticn is being conducted in general accordance with
the protocol proposed by Dr. John Wilson of the U.S. Environmental Protection Agency's Robert
S. Kerr Environmental Research Laboratory. Key parameters being assessed include seasonal
variations in ground-water flow direction and velocity, changes in hydrocarbon concentrations,
and changes in the concentrations of electror
of the stoichiometry of hydrocarbon biodegrcdation under various redox conditions, are used
to confirm that BTEX  biodegradation  is occurring at the site and to estimate  the BTEX
biodegradation rate.
The characterization of intrinsic bioremediation
are collected and analyzed on a quarterly bas
performed, and  two additional sampling
following:
 acceptors. These data, together with knowledge
 at this site is underway. Ground-water samples
s. Two ground-water sampling events have been
    are planned. To date, findings  include the
events
              The extent of the dissolved  hyd.rocarbon plume is smaller than would be
              expected if the hydrocarbons were not being biodegraded. BTEX concentrations
              decrease by five orders of magnitude within a lateral, downgradient distance of
              300 ft from the NAPL zone.

              While some oxygen flux into th 3 plume occurs, the majority of contaminant mass
              removal occurs underanoxic conditions. Naturally high concentrations ofsulfate
              in the ground water provide ar
              biodegradation undersulfate-
 essentially infinite supply of electron acceptor for
"educing conditions. Site data suggest that sulfate
              is the most significant electron acceptor in terms of hydrocarbon mass removal,
              with over 90 percent of the h/drocarbon mass removal attributable to sulfate
              reduction.

              Additional hydrocarbon mass removal is attributable to reduction of nitrate, iron,
              carbon dioxide, and oxygen.
Symposium on Intrinsic Bioremediation of Ground Water
                                         179

-------
       •       Elevated concentrations of bicarbonate downgradient of the plume, associated
               with hydrocarbon mineralization, confirm the role of intrinsic bioremediation in
               limiting plume migration and provide evidence that the plume has reached its
               steady-state extent.

The poster paper will present information on the site setting, monitoring and data evaluation
methodology, evidence of intrinsic bioremediation mechanisms and rates, and  impact of
observed biodegradation on plume migration.
180

-------
Demonstrating the Feasibility of Intrinsic
Manufactured  Gas Plant
 Bioremediation at a Former
Ian D. MacFarlane
EA Engineering, Science, and Technology, Sparks, MD

Edward J. Bouwer
The Johns Hopkins University, Baltimore, MD

Patricia J.S.  Colberg
University of Wyoming, Laramie, WY
A  former manufactured gas  plant (MGP)
investigation to assess natural, in situ biodeg
decisions on  intrinsic  bioremediation  and
in  Baltimore, Maryland,  is the  subject of an
adation for the purposes of basing remediation
sngineered,  enhanced  bioremediation.  Tar, a
byproduct of the former gas manufacturing operation, is found in the site's subsurface to depths
as great as  100  ft. The tar  is a dense  nonaqueous-phase liquid  (DNAP.L) that contains
monocyclic  aromatic hydrocarbons (MAHs),
aromatic hydrocarbons (PAHs), such as naph
as a long-term source of aromatic hydrocarbo ns to ground water, and concentrations in ground
water over much of the 70-acre site are clos

Laboratory and  field data were collected to
 such as benzene and toluene, and polycyclic
halene and benzo(a)pyrene. The tar DNAPL acts
5 to theoretical effective solubilities.

evaluate  biodegradation. The first phase of the
laboratory investigation, consisting of tests c n 1 6 soil samples from various depths, was to
discern whether or not microbes existed in the subsurface, if they could  use MGP-tar as a
carbon source, and if relationships could be established between  hydrogeology, contaminant
distribution, and microbiological characteristics.  Aerobic  and anaerobic enumerations were
performed, followed by fatty acid analyses to identify the microbes. Viable aerobic bacteria were
detected in all subsurface samples. Bacteria wjere grown in the samples plated under anaerobic
conditions, but at counts of 10 percent to 50 percent less than  the corresponding aerobic
counts. Tar-degrading bacteria were detected in 7 of the  16 samples.
 The second laboratory phase consisted of more detailed microcosm studies performed by the
 Johns Hopkins University (JHU) and the Unive -sity of Wyoming (UW). JHU used 49 soil samples
 from  five boreholes with site ground-water  and individual  radiolabeled target  substrates
 (benzene, naphthalene, phenanthrene, and acetic acid) in sealed vials to make microcosms that
 mimicked in situ redox conditions (i.e., oxygenated or unoxygenated). Viable bacteria  were
 enumerated, and total cell counts were perforjmed by the acridine orange direct-count method.
 Similar microcosm studies are being performed by UW under sulfidogenic, iron-reducing, and
 methanogenic conditions using phenol, benzsne, toluene, naphthalene, and phenanthrene as
 the targeted radiolabeled compounds. Benzere, naphthalene, and phenanthrene were observed
 to mineralize 6 percent to 24 percent, 8 percent to 43 percent, and 3  percent to 31 percent,
 respectively, in JHU aerobic  microcosms
 naphthalene mineralization (7 percent to 13
over a  4-week incubation  period.  Anaerobic
oercent) was observed in two JHU samples in the
 presence of NO3. Half-lives calculated from f
 tens to hundreds of days, with a lower half-li
rst-order degradation rates typically ranged from
e in the initial stages of incubation followed by a
 slower rate presumably  indicative of oxygen- or nutrient-limiting conditions.  Under sulfate-
 Symposium on Intrinsic Bioremediation of Ground Water
                                                                                    181

-------
 reducihg conditions, phenol mineralized 1 3  percent to 1 8 percent in 200 d, but benzene,
 naphthalene, and phenanthrene showed less than 1 percent mineralization in the same period,
 and toluene showed  less than  1  percent mineralization in 165 d.  The apparent limited
 transformation may, in  fact, be  due to the small inoculum  sizes used, a  phenomenon
 documented by others.

 Field investigations for natural in situ biodegradation included aqueous phase redox conditions,
 biogenic product analyses, and apparent attenuation of model contaminants. Ground-water
 quality data generally showed reduced conditions with little or no measurable oxygen, low redox
 potential (-70 mV average), high  biochemical oxygen demand in source zones (>200 mg/L),
 elevated sulfate (2,200 mg/L average), and elevated iron (570 mg/L average). Biogenic gases
 (CO2, H2S, and CH4) were detected at levels  greater than atmospheric in 1 1  of the 16 wells
 measured.

 Apparent degradation rates were calculated using the first-order model by regressing the natural
 log of constituent concentrations (adjusted for dilution) by estimated in situ  travel time. Because
 no nonreactive, conservative tracers are unique to the tar sources, dilution was estimated by
 considering dissolved carbon (organic plus inorganic) as a tracer. While the use of carbon as
 a tracer is prone to error due to its reactive  nature, its  use is conservative in that observed
 carbon concentrations would tend to  be  less  than actual anthropogenic  carbon, resulting  in
 overestimates of  dilution  and underestimates of degradation rates. Half-lives for benzene,
 toluene, ethylbenzene, xylenes, and naphthalene were calculated to be 729, 660, 877, 855,
 and 2,1 66 d, respectively. Half-lives forthe three-ring and greater PAHs were not calculated due
 to the poor regression correlation coefficients. Degradation estimates showed thattoluene is the
 most preferred aromatic substrate studied and that naphthalene appears to degrade the slowest
 as predictable from the literature.  Surprisingly, the benzene rate was only slightly less than the
 toluene rate.

 Laboratory investigations have shown that 1) microbes exist in the subsurface, 2) microbes are
 capable of using  tar as a carbon source,  3) various redox conditions can be established with
 site consortia, and 4) site  bacteria can degrade selected aromatic constituents  under aerobic
 and anaerobic conditions. Field  evidence showing various  redox conditions  and biogenic
 products of organic degradation gives  clues to the possible fate of aromatic hydrocarbons, but
 this indirect evidence can  only be used to support more definitive evidence in demonstrating
 intrinsic bioremediation. Although estimates of in situ constituent decay are  based on numerous
 assumptions and are fraught with uncertainty, this evidence is needed to show real attenuation
 of aromatic hydrocarbons. In this case, enough geochemical and hydrogeologic data were
 availableto segregate dilution (an important "attenuation" process) from degradation processes.
 The task of estimating apparent degradation was facilitated by the relatively simple hydraulics
 and the aged system to allow assumption for negligible sorption. The attenuation assessment
 demonstrated that contaminant loss was observed overtime along the aqueous plume travel
 path (i.e., travel time, rather than atone point over time) and degradation, probably biotic, was
 measured for target contaminants.

The combined laboratory and field evidence point to natural in situ biodegradation as an active
 process in the site's low oxygen, subsurface environment. Intrinsic bioremediation, albeit slow
due to mass transfer limitations from the tar to the aqueous phase/may be technically viable
for controlling aqueous aromatic hydrocarbon contamination emanating from MGP-tar sources.
 Laboratory studies are continuing to explore compound-specific biodegradation under various
conditions, and plans are  being formulated now for an in situ biodegradation  pilot study.
182

-------
Natural  and  Enhanced Bioremediation o
California:  Laboratory and Field Investigations
 Aromatic. Hydrocarbons at Seal Beach,
Harold A. Ball, Gary D. Hopkins, Eva Orwin,
Western Region Hazardous Substance Resear
 and Martin Reinhard
ch Center, Stanford, CA
Introduction

The objective of this study was to develop our
important for intrinsic anaerobic biodegrada
ground-water aquifers, and to determine me
 understanding of environmental factors that are
 ion of aromatic  hydrocarbons in contaminated
 hods to enhance this process. The focus of the
investigation was a site at the Seal Beach Naval Weapons Station in southern California, where
a significant gasoline spill resulted in contamiration of the ground-water aquifer (1). Inthe field,
nitrate was present at about 5 mg/L in background wells and approached detection levels (<0.1
mg/L) in the contaminated wells. There was a
ligh natural background sulfate concentration of
about 85 mg/L in the ground water, and methane was  detected in the contaminated well
headspace. The distribution of aromatics present in the contaminated ground water differs from
that expected from dissolution of pure gasoline (2). This suggests that natural biotransformation
of several organic species is occurring at the site. The project was divided into laboratory and
field components, which were  interrelated.  The goals  of both the laboratory  and field
experiments were to determine the capability of the native aquifer microbial community to
transform aromatic hydrocarbon compounds under anaerobic conditions and to understand the
effect of alternate environmental conditions on the transformation processes. Field experiments
were  carried  out on  site at Seal  Beach.
biotransformation under natural (presumed si
have been carried out.
At the field  site,  experimental  monitoring  of
Ifate-reducing) and nitrate-enhanced conditions
Approach  and Results

Laboratory Study

In a laboratory microcosm experiment (3), we evaluated several factors which were hypothesized
to influence in situ biotransformation processes.  Individual monoaromatic compounds (e.g.,
benzene, toluene, ethylbenzene, and m-, p-J and o- xylene) were the primary substrates. In
replicate bottles during the first 52 d of the study, toluene and m+p-xylene (here, m-xylene and
p-xylene were measured as a summed pare meter)  were biotransformed in the  unamended
ground-water samples under presumed sulfa e-reducing  conditions. Addition of nitrate to the
ground water increased rates of toluene  biotransformation coupled to nitrate reduction,
stimulated  biotransformation of ethylbenzene, and inhibited the complete loss of m+p-xylene
that was observed when nitrate was  not added and sulfate-reducing conditions prevailed.
Addition of the nutrients ammonia and phosphate had no effect on either the rate  of aromatics
transformation or the distribution of aromatics transformed. When Seal Beach sediment was
placed  into nitrate-reducing media, ethylbenzene was transformed first, followed by toluene.
When the  sediment was placed into sulfatejreducing media, lag times were increased, but
toluene and m-xylene were ultimately transformed just as  in the microcosms with ground water
Symposium on Intrinsic Bioremediation of Ground Water
                                          183

-------
                                                                                    no
 alone.  Although  methane  had  been  detected in the  field, there  appeared  to  be  .,~
 transformation of aromatic compounds in the methanogenic microcosms during the period of
 the experiment.

 Bioreactor Study

 A pilot-scale facility consisting of 90-L reactors was constructed at the Seal Beach site (4-6). The
 facility was designed for the operation of  three anaerobic  in situ bioreactors. The reactors
 consisted of aquifer-sediment-filled, stainless steel cylindrical vessels with the capability to control
 and monitor both hydrodynamic flow and supplements to the composition of the native ground-
 water influent.  Initial operation of the three anoxic/anaerobic reactors focused  on evaluating
 anaerobic bioremediation strategies foraromatic hydrocarbons under natural (presumed sulfate-
 reducing) and enhanced denitrifying conditions.

 Bioreactor results were consistent with the laboratory microcosm experiments. Toluene arid m+p-
 xylene were degraded in both the unamended and nitrate-amended bioreactors. Degradation
 of ethylbenzene was stimulated by nitrate addition. There was no evidence that  benzene or o-
 xylene was transformed in either reactor. The final percentage removal efficiency appeared to
 be higher in the unamended bioreactor, where  flow was slower.

 Field Study

 Field experiments  have been conducted to assess aromatic  biotransformation  in  a test zone
 within the contaminated aquifer at the Seal  Beach site. Initial work focused  on evaluation of
 intrinsic bioremediation as evidenced by the distribution of aromatic species in background wells.
 Subsequent experiments to determine our ability to enhance this biotransformation have been
 conducted using a slug test experimental design in which a single well was used forthe injection
 of the "slug" or test pulse and the same well was used to extract the test pulse. Since the native
 ground  water contained a variety of electron acceptors and the water  used for the injected
 pulses was water that had previously been extracted  from the test zone, the ground water was
treated to control the concentration of all electron acceptors and organics during the injection
 of the test pulse. Before injection, the desired  salts  were  added  back  to the  deoxygenated
 injection stream, and the stream was metered into the injection well. Sodium bromide was added
 as a conservative tracer. Under this scenario, the'different electron acceptors investigated (e.g.,
 nitrate and sulfate) could be added as desired. During initial tracer studies, the injection water
was  organics free, and thus the source of the organics was desorption from the  in  situ aquifer
solids. In subsequent and ongoing bioremediation studies, benzene, toluene,  ethylbenzene, m-
xylene, and  o-xylene were added with the injection pulse at a concentration  of approximately
200 fj.g/1 each.

The initial bromide tracer data showed stable tracer concentrations and indicated  no substantial
encroachment of native ground water detected in the first 0.4 pore volumes.  There was a very
small hydraulic gradient at the site, hence  recovery of the bromide mass from the test wells
 ranged from 93 percent to 99 percent with the extraction of three pore volumes  over a 1 03-d
 period.  During the tracer test, the  equilibrium  desorption concentrations  for  the aromatic
 hydrocarbons when the electron acceptors nitrate and sulfate were absent from the ground water
were evaluated. Benzene, ethylbenzene, and o-xylene concentrations remained relatively stable
and thus appeared to be at an equilibrium. The toluene and  m+p-xylene concentrations had
a downward trend  relative to benzene once  the  native ground  water encroached after
approximately 0.4 pore volumes, suggesting thatthe nitrate and sulfate concentrations available
184

-------
in the native ground water supported some i
experiment for toluene and m+p-xylene remc
In a nitrate augmentation experiment, nitrate c
       irtrinsic biological activity in the latter part of the
         val.
         nd aromatics were added to the injection pulse,
resulting in complete consumption of toluene cind m-xylene followed by ethylbenzene within the
first 2 wk. o-Xylene was slowly degraded, and its concentration approached zero by day 60.
There was no apparent loss of benzene when! compared with the inert tracer. The addition of
nitrate to the test region appeared to enhance the natural anaerobic denitrifying population. This
would confirm that there was already an active
activity was enhanced by the addition of nitrate. With the exception of o-xylene transformation,
these results were comparable with those from
         nitrate-reducing population in the aquifer whose
         the nitrate-amended microcosm and bioreactor
experiments, wherein toluene, ethylbenzene, and m-xylene were transformed under denitrifying
conditions.

During the tracer study, methane was detected in the test wells. With the encroachment of the
native ground water and associated increase in nitrate and sulfate concentrations, the methane
concentration decreased to values close to zero, suggesting that  nitrate and sulfate inhibit
methanogenesis at this site.
Acknowledgment

Funding for this study was provided by the U.S.  Environmental Protection Agency's Office of
Research and Development,  under agreement  R-81 5738-01 through the Western  Region
Hazardous Substance Research Center. The content of this study does not necessarily represent
the views of the Agency. Additional funding
Technology Company, Richmond, California.
         was obtained from the Chevron  Research and
References
1.     Schroeder, R.A.  1991. Delineation o
       shallow deposits at the U.S. Naval
       Invest. Rep. 89-4203.
          a hydrocarbon (weathered gasoline) plume in
     Weapons Station, Seal Beach, California. Water Res.
2.     Cline, P.V.,J.J. Delfino, and P.S.C. Rao. 1 991. Partitioning of aromatic constituents into
       water  from gasoline and
       25(5):914-920.
other complex  solvent mixtures. Environ.  Sci. Technol.
3.      Ball, H.A., and M. Reinhard. 1 994. laboratory study of monoaromatic hydrocarbon
        degradation under anaerobic conditions at Seal Beach, California. In preparation.
4.      Ball, H.A., and M. Reinhard, M. 1994
        degradation under anaerobic conditi
         , Pilot-scale study of monoaromatic hydrocarbon
          ns at Seal Beach, California. In preparation.
5.      Huxley, M.P., C. Lebron, M. Reinhard, H. Ball, H.F; Ridgway, and D. Phipps. 1 992.
        Anaerobic and aerobic degradation of aromatic hydrocarbons using in situ bioreactors
        at an unleaded gasoline spill site. Presented at the 18th Environmental Symposium of
        the American Defense Preparedness /Association, Alexandria, VA.
                                                                                   185

-------
 6.      Reinhard, M., LE. Wills, H.A. Ball, T. Harmon, D.W. Phipps, H.F. Ridgway, and M.P.
        Eisman.  1991. A field experiment for the anaerobic biotransformation of aromatic
        hydrocarbon compounds at Seal Beach, California. In:  Hinchee, R.E., and  R.F.
        Olfenbuttel, eds. In situ bioreclamation: Applications and investigations for hydrocarbon
        and contaminated site remediation. Boston, MA: Butterworth-Heinemann.
186

-------
The Complete Dechiorination of Trichloroefhene to Ethene Under Natural
Conditions in a Shallow Bedrock Aquifer) Located  in New York State
David Major, Evan Cox, and Elizabeth Edwards
Beak Consultants Limited, Guelph, Ontario

Paul W. Hare
General Electric Company, Corporate Envircnmental Programs, Albany, NY
Introduction

In anaerobic environments, chlorinated ethenes can act as electron acceptors in a process called
reductive dehalogenation (specifically, reduct
however, has been shown to vary in anaerobi
:ve dechlorination). The extent of dechlorination,
: dechlorination studies depending upon the flow
and  availability of electrons  within the  anaerobic microbial  community.  For example,
dechlorinated intermediates such as dichloro<
ithene (DCE) and vinyl chloride (VC) were found
to accumulate during the dechlorination of tetrachloroethene (PCE) and trichloroethene (TCE)
(1 -3). Freedman and Gossett (4), however, were the first to observe the complete dechlorination
of PCE to ethene in a methanogenic enrichme it culture, and DeBruin et al. (5) showed that PCE
could be reduced to ethane. The first field observation of the complete dechlorination of PCE
to ethene was documented by Major et al. (5). Their laboratory and field study showed that
anaerobic microorganisms in a low-permeabil
tyaquiferwere capable of naturally dechlorinating
PCE in the presence of methanol. This paper documents that microorganisms in a bedrock
aquifer unit are also capable of completely cechlorinating TCE to ethene.
Study Site Conditions

The study site is located in the Finger Lakes region of central New York. The property was used
                                          for a variety of electrical components, including
high-voltage semiconductors. In the early to mid-1 960s waste solvents were disposed of in an
unlined evaporation pit. TCE, which was often mixed with acetone or methanol, was among the
solvents disposed of in the unlined evaporation pit. As a result, these chemicals are now found
                                        Lll X
in the overburden and bedrock units beneath the study site.
Results

Our study involved collecting representative
ground-water samples from 21 existing ground-
water monitoring wells, mostly in the shallow bedrock unit, for geochemical and microbiological
analyses.  1,2-DCE and VC were detected in around-water samples, which indicated that TCE
was being biodegraded in the subsurface at the site. These TCE degradation products were not
used or produced at the site, and thus their presence can only be attributed to the dechlorination
of TCE. In addition, the detection of ethene provides evidence that VC is being dechlorinated
at the site. Three observations of the relative distribution of TCE and its dechlorination products
suggested that the migration of the volatile organic compounds (VOCs) in the shallow bedrock
Symposium on Intrinsic Bioremediation of Ground Water
                                         187

-------
 unit is being controlled by biodegradation. First, the distribution of TCE is much less extensive
 than the observed distributions for 1,2-DCE, VC, and ethene. Second, the distribution of VC,
 which should be greater than that of 1,2-DCE as predicted by its mobility in ground water, was
 essentially the same as the distribution of 1,2-DCE. Third, VC and  ethene migrate at similar
 rates relative  to  ground-water velocity and should have had similar distributions, but  the
 distribution of VC was less than ethene. The distribution of acetone and methanol is considerably
 limited in comparison to the distribution of the  chlorinated VOCs  and  ethene. The mobility of
 acetone  and methanol should  be approximately the same as the average linear ground-water
 flow; however, their distribution was less than the distribution of the VOCs. This suggests that
 acetone  and methanol are also being biodegraded.

 In addition to the distributions of VOCs, the distributions of inorganic onions,  methane and
 methane isotopes, and volatile fatty acids (e.g., acetate) were used as indirect measures of the
 activity of functional groups of  microorganisms  in the bedrock. The depletion of sulfate and the
 production  of methane  and  acetate  indicated  that  sulfate-reducing,  methanogenic,  and
 acetogenic bacteria were active in the bedrock  aquifer. Isotopic analysis of methane indicated
 thatthe methane was produced biotically. Furthermore, the distribution of methane and methane
 isotopes  clearly showed that the  microorganisms were  active in the  bedrock.

 Microbial biomass, composition, and nutritional  status were assessed by extracting and analyzing
 phospholipid fatty acids (PLFAs) and respiratory quinones from microorganisms thatwere trapped
 onto 0.22-^m membranes. Analysis  of respiratory quinones indicated that the  microbial
 populations at the site are strictly anaerobic. The microbial biomass in the ground water ranged
 from 1.6x 102 cells/mLto4.2 x TO4 cell/ml. The total biomass appeared generally to correlate
 with the  presence of VOCs and other nutrients, and  was found to be higher in the areas
 containing acetone and methanol. The total biomass was orders of magnitude higher near VOC
 source areas, as well as in areas with measurable concentrations of acetone and methanol, than
 at downgradient  or background (upgradient)  locations. The microbial  biomass distribution
 suggested that a  biologically active zone (BAZ)  has developed in response to the presence of
 VOCs,  acetone,  and methanol.  The microbial  populations  in the  samples  generally
 demonstrated  nutritional  or environmental stress, as indicated by the  ratio of specific  PLFAs.
 Stress may be due to an inadequate supply of nitrogen and phosphorous to support ideal
 growth. Cluster analysis of the  PLFA data showed that three population groups exist at the site.
 The population groups appearto coincide with observed changes in the concentration and types
 of VOCs and other geochemical parameters.
Conclusions

This study provides evidence that microbial  populations can exist and function  in bedrock.
Furthermore, these populations possess an intrinsic capability to anaerobically dechlorinate TCE
to ethene when suitable substrates are present to support their growth. At this study site, an
active and diverse anaerobic microbial community, consisting of sulfate-reducing, methanogenic,
and acetogenic  bacteria, has been established and is being  maintained by acetone and
methanol. This anaerobic microbial community is affecting the distribution and migration of TCE,
TCE biodegradation products, and other chemicals at the site.                    :
188

-------
References
i.
2.
3.
Bouwer, E.J., and P.L McCarty. 1 983.
aliphatic organic compounds
45(4):1,286-1,294.
        Transformation of 1 - and 2-carbon halogenated
under methanogenic conditions. Appl. Environ. Microbiol.
Parsons, F., P.R. Wood, and J. DeMarco
and trichloroethylene in microcosms cm-
Wilson, B.H., G.B.  Smith,  and J.J.
alkylbenzenes and  halogenated alif
material: A microcosm study. Environ
4.     Freedman, D.L,  and J.M. Gossett.
         .  1984. Transformations of tetrachloroethylene
        id ground water. J. Am. Water. Assoc. 76:56-59.
         Rees. 1986. Biotransformations of  selected
        hatic  hydrocarbons  in  methanogenic  aquifer
         Sci. Technol. 20(10):997-1,002.

         1989. Biological  reductive dechlorination of
       tetrachloroethylene and trichloroethylene to ethylene under methanogenic conditions.
       Appl. Environ. Microbiol. 55(9):2,14-

5.     De Bruin, W.P., M.J.J.  Kotterman, M,
                                   -2,151.

                                  A. Posthumus, G. Schraa, and A.J.B. Zehnder.
       1 992. Complete biological reductive transformation of tetrachloroethene to ethane.
       Appl. Environ. Microbiol. 58(6):1,996-2,000.

6.     Major, D.M., E.W. Hodgins, and B.J. Butler. 1 991. Field and laboratory evidence of in
       situ biotransformation of tetrachloroet
       facility in North  Toronto.  In:  Hind
       bioremediation, pp. 147-171. Bostor
                                  nene to ethene and ethane at a chemical transfer
                                   ee,  R.E., and R.F.  Olfenbuttel,  eds.  On-site
                                   , MA: Butterworth-Heinemann.
                                                                                   189

-------

-------