\ UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON D.C. 20460
August 11, 1992
OfHCEOF
THE-ADMINISTRATOR
SCIENCE ADVISORY BOARD
EPA-SAB-EPIC-LTR-92-015
Mr. William K. Reilly
Administrator
U.S. Environmental Protection Agency
401 M street, s.W.
Washington, D-C- 20460
RE: SAB REVIEW OF THE ALASKAN BtOREMEDIATION OIL SPILL PROJECT
\.
Dear Mr. Reilly,
The Alaskan Bioremediation Task Croup of the Science Advisory Board (SAB)
has completed its review of the final report on the results of the Alaskan oil
Spill Bioremediatioft Project. This group met on June 1-2, 1992 to conduct its
review. Dr. John Skinner, Deputy Assistant Administrator, Office of Research and
Development asked the SAB to review this report and to address several points as
part of its review. The charge to the SAB is attached.
The SAB reviewed the preliminary plan for this research in 19S§ (EPA-SAB-
EETFC-89-023). 'Since that time, the SAB has received interim updates of progress,
so the task group was already aware of many of the results. In addition to the
questions that QRD asked the SAB to address, the Task Group addressed issues
relevant to how EPA will apply its experience to future oil spills and massive
chemical releases that may be cleaned up by bioremediatioja. Many of the lessons
learned from this investigation ara of a generic character and can be translated
to, apply to other types of field studies of deliberately stimulated biological
processes.
The Task Group commends the Agency's efforts to rapidly address a
significant problem under adverse environmental conditions within a highly complex
political and legal framework. The project represents a significant
accomplishment which should lay the foundation to improve research and planning
for emergency responses in the future.
' ' ADEQUACY OF CONCLUSIONS
The data collected from the test sites in Prince William sound snow that the
application of fertiliser solutions to oil-contaminated beaches enhanced oil
removal at some locations. It ia likely that this removal is attributable to a
combination of biodegradation and physical and chemical removal associated with
this technique. This biorewediation teehni
-------
Harbor, and did not have an effect at Disk Island. The reasons for these
differences are not totally clear, but the research does demonstrate the
importance of site characteristics in determining the effects of the treatment
used for bioremediation. The data suggest that a number of factors affect the
outcome of bioreicediation efforts, it is likely that with increasing information
from more locations and with more microcosm and laboratory research, the
controlling factors will be defined as a set of parameters (e.g., porosity of
beach materials, beach slope, bioavailability of oil constituents, or fertilizer
nutrient ratios), some of which can be modified to allow a greater success rate
for bioremediation. The simple comparison below, shows that the sites are not
comparable instead they represent individual teats at distinct locations using
different fertilizers And application techniques. Thus each site contributes
additional information on the effectiveness of the bioremediation technique, -but
the conclusions from one site may not apply to other locations in a direct manner.
Comparison of Bioremediation Sites
Sites
Disk Island
1990
Snug Harbor
1989
Fertilizer Applications/Beach Conditions Oil Loss vs. Control
Sfaw release (B) fertilizer, low slope,
sand-gravel beach material
Oleophilic (O) and (B) fertilizers, sites with
cobble over gravel and mixed sand and gravel
Passage Cove (O)fQranule (G) and dotty fertilizer solution
1989-90 from sprinkler (S), prewetshing of cobble and
subsurface may haw spread oil, increasing
exposure
Elrington Island (S) treatment, focus on subsurface oil,
1990 over mixed gravel, high energy beach
No significant difference
Mixed results, faster hss on cobble,
KO difference with sand/gravel
Marked increase, sprinkler judged
best delivery system, changes in both
cobble and subsurface sand and gravel
Marked increase with (S), some loss
occurred with (O)/(G) mixture
Exxon Corporation, the State of Alaska, and other federal agencies
(particularly the National Oeeanographic and Atmospheric Adminiptrmtion and the
U.S. Fish and Wildlife service} have all collected extensive data sets relevant to
bioremediation. We recommend that EPA organize a meeting of the company and
agencies to review the findings and reconcile the data sets relative to the
critical questions regarding the mechanisms of oil removal at the sites where
bioremediation was conducted.
EXT1APOLATIQN
The results of the Alaskan Project confirm that microbial biodegradation can
be stimulated to bring about the destruction of complex organic constituents of
oil, providing the specific decomposers are present among the indigenous microbial
populations. The project demonstrated that the use of fertilizers to enhance
decomposition of petroleum residuals is a sound approach, providing that treatment
designs take into consideration differences in local conditions and the variables
-------
that may have an effect on the biodegradation process. One can anticipate, based
on the results of the Alaska Project, considerable variation in the degree of
bioremediation. A challenge to the Agency lies in taking these results from
Prince William Sound and establishing a basis for dealing with future oil spills
wherein bioremediation using fertilizers would be the method of choice.
REDUCTION OF CLEANUP TIME
Bioremediation efforts reduced the cleanup time {relative to natural
degradation) at sites in Prince William Sound, but the effect of bioremediation
was dependent, on tha particular beach and on the depth of oil penetration below
the beach surface. Bioremediation was effective at the Passage Cove and Elrington
Island sites, with cleanup reduction times for surface material of the order of 60
to 120 days. Biodegradation of oil occurred at Snug Barbor, but a reduction of
cleanup time through biorewediation was not clearly demonstrated. Bioremediation
was not effective and a reduction of cleanup time was not shown at Disk Island.
The reductions in cleanup time were strongly dependent on location, and the
greatest enhancements in bioremediation were observed for subsurface beach
material. However, the definition of cleanup tiae in the report is somewhat
confusing. Several indicators of cleanup can be used and these may be based on
changes in chemical composition of the oil. The analysis used in the report to
demonstrate the reduction of cleanup time is based solely on the loss of total
oil.
The conclusion that bioremediation reduced cleanup time must be qualified in
view of the high variability in oil chemistry at -the sites, the fact that some
beaches were prewashed and the fact that the oil was continuously aging and
weathering during the bioremediation period. Moreover, the specific estimates of
cleanup tune given in this report have considerable statistical uncertainty.
Quantification of the effect of bioremediation is difficult because of the limited
number of sites that received different treatments and the fact that the sites had
different geological characteristics.
ADEQUACY OF SUPPORTING RESEARCH
Considerable supporting research was performed that yielded data important
for making operational decisions necessary for the field operations. However,
much of the supporting research was insufficient to interpret the field results,
since — if for no other reason the field results were often not fully
available at the time that the supporting research was designed and conducted. To
maximize the effectiveness of future bioremediation efforts, laboratory and
microcosm research should be conducted with a view to providing the needed
explanations.
SELECTION AND TESTING OF FERTILIZERS
Major constraints existed at the time that selections had to be made on the
specific fertilizers to be used. These constraints included, the absence of a body
of directly relevant information for circumstances at the Alaskan sites and the
lack of availability of sufficient quantities of'fertilizers that might be of
possible utility. Similarly, testing procedures specifically designed for
evaluating fertilizers to be used on beaches had not been devised and validated.
-------
Further research should be conducted to determine which components in the
fertilizer stimulate removal of oil. Given the limitations, we believe that EPA
made reasonable choices in selecting and evaluating the fertilizers. However, in
light of the apparent effects of fertilizers in enhancing biodegradation,' it is
essential that a research arid development program be implemented to 'determine the
types of fertilizer materials, formulations and composition needed to optimize the
rate and extent of bioremediation, to devise testing procedures that will permit
evaluation of the fertilizer materials for the likely types of spill sites and to
develop fertilizer application methods moat useful for various' sites and types of
oil spills.
SAFETY OF BIORZMEDIATION
The Agency conducted an assessment of the safety of bioresnediation and
evaluated possible ecological effects. The chief ecological issue of concern was
the potential impacts of nitrate- and phosphate fertilizers applied during the
bioremediation activity on the structure and/or function of the near-shore marine
community. The potential mechanism* of impact include acute toxicity of ammonium,
eutrophication resulting in low dissolved oxygen, nutrient enrichment of the
waters resulting in blooms of algae and the bioaecumulation in marine benthic food
chains of intermediate compounds formed during biodegradation. A second issue
involved the potential redistribution of the oil residue back into the offshore
aquatic environment.
Assessments were performed of acute toxicity and potentials for
eutrophication. The maximum ammonium concentration observed in the water
immediately adjacent to the test plots was Or035ppm, which is well below the
estimated standard for chronic toxicity of 8 ppm. The nutrients released from the
bioremediation test sites did not appear to significantly enhance the available
nutrients in near shore waters. The issue of redistribution of the oil residue
was addressed by placing mussels in cages on the bottom in areas adjacent to the
test plots. Mussels filter fine particles and are good bioaecumulators for
adsorbed residues. The distribution of solubilized oil residues were not
monitored. Major changes in phytoplankton abundance and productivity were not
observed. Changes in benthie algal abundance and algal species were not
monitored. Such monitoring would have been useful for nutrient loading or the
redistribution of oil residues to adversely affect food availability for filter
feeders or to stimulate toxic dinoflagellateo. The bioaccumulation pathways were
identified through the use of stable isotopes of carbon and nitrogen. Direct
measurements of oil residues at various levels in this food chain were not made
because EPA reasoned that since the oil residues had been weathered for at least &
months before the bioremediation was initiated, there was little need to analyze
for bioaccumulation. Given the site-specific conditions of this Alaskan
ecosystem, the timing of the onset of bioremediation, the limited areas of
fertilizer application and the limited application rates, adequate field
information was gathered to conclude that the bioremediation effort would not
negatively impact the Prince William sound ecosystem. Furthermore there was no
demonstrable evidence-of adverse impacts. The potential for impact is site
specific, ' '
The methods developed for assessing environmental safety of the program can
provide a foundation for future assessments but should not be considered
-------
sufficient by themsei'ves for all situations.1'' The environment'of Prince William
Sound is unique in many ways, and additional testing may be required on a site-
specific basis. Factors to be considered should include: type of beach in terms
of particle size and slope; energy of the System, both from tides and wind;
sensitive species, habitats and communitiesj timing of a cleanup event relative to
ecological utilization of the site; trophic structure and potential for food chain
transport of metabolites through food chainsj potential for eutrophication,
particularly long-term adverse effects; potential for induction of anaerobic
conditions; and potential for human exposure.
These issues can best be addressed by performing initial site-specific risk
assessments (both human health and ecological), as was done by EPA in this
project, to estimate the relative importance of exposure pathways and the relative
sensitivity of various biological endpoints. These data can be used to determine
if the suite of tests employed are adequate or if additional site-specific tests
should be added. These types of assessments should ba carried out soon as part of
a research program for oil spills so that ecosystem specific suites of tests can
be established. This approach would be particularly important for sensitive
environmental habitats such as coral reefs, mangrove swamps, and salt marshes
which would be particularly sensitive to oil spills*
BIOREMEDIATION STRATEGIES FOR FOTTOE OIL SPILLS
As would be expected from research implemented during an emergency response,
the data gathered in the Alaska Bioremediation Project are highly variable in both
quantity and quality depending on location, site, and experimental protocol
employed. However, the research conducted during the course of this project
represents an important first step in developing a scientific basis for strategies
to deal more effectively with future Oil Spills. Nevertheless, the data from the
Alaska Bioremediation Project need further analysis to carefully differentiate
between conclusions drawn from data sets that differ greatly in quality.
The lessons learned from Alaska are manifold, but many are probably unique
to the location, temperature, seasonal cycles, type(s) of shore, etc. Transfer of
the technology to other locations may be thus limited to generic issues only, and
caution should be exercised in applying the methods and approaches used in Alaska
to other locations. For example, fertilizer addition clearly enhanced the rate of
oil removal in some instances and not in others, but the data do not provide an
adequate basis for deciding when to apply fertilizers to oil spills (i.e., shortly
after the spill or some time later) in more temperate climates or to beaches with
higher contents of sand and organic matter or in less energetic areas with longer
hydraulic detention times, etc. Banco, considerable research will be necessary to
identify, the causes of variability in different spill scenarios. Systematic
analysis of spill events will both indicate where the information gained in Alaska
can be best applied and provide guidance for prioritizing future research needs
for wide scale use of bloremediation for oil spills.
"The measurements performed in the bioremediation project were a good initial
approach to quantifying the effectiveness of the treatments. Nevertheless these
measurements have limitations. With further.development some offer great promise.
Chemical analyses of total oil remaining (gravimetric), hydrocarbon composition (
GC and GC-MS) and nutrients ate essential to understanding the fate of the oil and
-------
need for and effectiveness of nutrient additions. Measurements of microfaial
activity suitable for use in the field need further development. The Task Group
is not convinced of the usefulness of measurements of microbial biomass in the
field, and also cautions EPA that use of predation models to explain observed
biomaes is likely to be unproductive. The measurements should be developed and
evaluated in the context of specific hypotheses to be tested. Research will need
to be undertaken to adequately develop and assess the efficacy of mil measures
used to quantify the success of bioremediation.
PATA INTERPRETATION
A large amount of useful data, was collected by the Alaska Oil Spill
Bioremediation Project. If these data are to be used to their fullest extent,
rigorous interpretation is essential. Only in some- of the field studies was
convincing evidence of bioremediatlon obtained, yet many of the summaries and
conclusions read the same. The document should be revised to clearly distinguish
the data that show bioremedlation and those that do not. Much could be learned
from careful studies in which bioreroediation was not successful. The role of
chemical and physical factors in removal and degradation of oil in addition to
and/or in concert with biodegradation also needs to be more clearly discussed»
OVERVIEW CHAPTER
The report of the Alaska Oil Spill Bioremediation Project contains an
immense amount of data. Most readers, however, will be unable to assess the
results of the project because of the enormous amount of detailed information.
Thus, a chapter is needed that gives an overview of the issues, the purpose of the
program, the methods used, the major findings, the problems of variability of the
data from field assessmentB, the ecological hazards that may have been minimized
or prevented, and the conclusions that are particularly pertinent for future
bioremediation., This chapter will not only be useful to the general reader but
will give guidance to the specialist who indeed will read the full report. That
chapter probably would be best placed before the detailed presentation of methods
and results.
TESTS AND MEASUREMENTS
Many measurements have been made, and several types of tests have been
conducted as a prelude to field activities or to permit interpretation of those
field activities. Undoubtedly, some of these measurements and tests were
essential. It is also likely that hindsight will show that acme of the chemical,
microbiological or ecological, measurements or testa that were conducted were not
really necessary and that other measurements or tests should have been performed.
EPA should make such determinations and propose a revised suite of measurements
and tests for future oil spills and for the supporting research. Particular
attention should be given to the tests and analyses that should be conducted in
the short time period between the time of * spill and the time when field
bioremediation should be initiated.
-------
STATISTicar; :DESIGN"AND ANALYSIS"*''"'' ':"'
Large-scale applications of bioremediation technologies in heterogeneous .
environments require detailed consideration of experimental design, data
management systems and statistical analyses. The physical heterogeneity of the
Sites in Prince William sound is likely to be encountered often in the future.
Alternative statistical designs range from stratified analysis of variance with
suitable replications to multiple regression, which maximizes the range of
environmental characteristics*
Parametric analyses should be utilized that provide confidence intervals on
rates and endpoints of bioremediation. The process and endpoint parameters must
be identified before the appropriate experimental design can be chosen. The
number of replicate samples, the range of stratified conditions, the frequency of
monitoring samples.and the number of treatments being tested at any point in time
are all dependent on tha experimental design of choice. These decisions all must
be explicitly made before any field activities are initiated.
Since oil spills are episodic events that can not be predicted ahead of
time, the Task Group recommends that specific alternative design and analysis
strategies be considered "now for various classes of ecosystems as part of EPA'B
remedial-response planning process.
MODELING
Mechanistic models of bioremediation technologies based on fertilizer
applications must be developed that reflect state-of-the-art capabilities (e.g.,
fate and transport processes, critical analytical data, and metabolic rate
processes). These models will be- essential for the engineering design required
for scaling the activity from test plots to full-scale remediation. The
mechanisms required in these design models must involve transport, fate and
storage phenomena. The models can be parameterized and validated using laboratory
and field microcosms, stable isotope analyses of field experiments, and mass
transfer/balance measurements obtained during remediation projects.
The attempt to develop predator/prey models of protozoa and bacterial
bioma-SB during bioremediation is inappropriate at this time, because the effect of
protozoa is unknown, iacterial biomass is not the appropriate" state variable, and
the reported modeling activities are. not the most productive alternatives.
ORGANIZATION AND MANAGEMENT
we highly commend the EPA staff that organized and managed the project.
They planned, initiated, and conducted a research program of which the Agency may
be proud. Nevertheless, considering that the Exxon Valdez spill will not be the
last oil spill, EPA should evaluate the effectiveness of the organisational
structure used for this project for its relative success, the problems it
encountered and eas* of operation so that future initiatives*can be the moat
productive. EPA should consider the value of having a rapid response group
'(including aquatic physical scientists, microbiologists, environmental
toxicologists, chemists, and atatiaticans) to undertake research that could be
initiated under emergency conditions such as those experienced in Prince William
-------
Sound. This might provide opportunities to rapidly improve on approaches and
management strategies that could increase the efficiency of bioremediation under
emergency conditions.
We appreciate the opportunity to review this completed project and look
forward to receiving the final documents. In particular, we mra interested in how
SPA applies this experience as guidance for selecting appropriate responses to oil
spillB in the future and for planning further research to improve field
monitoring, including its statistical design.
sincerely,
Rayfeond Loehr,
Executive Committee
Science Advisory Board
\je**~*ejsr*
Kaaneth L* Dickson, Chair
Ecological Processes and
Effects Committee
Martin Alexander, Chair
Alaskan
Task Group
Attachment
-------
US ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD
ECOLOGICAL PROCESSES AND EFFECTS COMMITTEE
ALASKAN BIOREMEDIATION TASK GROUP
ROSTER
CHAIR
DR. MARTIN ALEXANDER, Professor, Soil, Crops and Atmospheric Sciences,
Cornell University, Ithaca, New York
MEMBERS
DR, STANLEY I. AUERBACH, Director, Environmental Sciences Division, Oak
Ridge National Laboratory, Oak Ridge, Tennessee
DR. WILLIAM E. COOPER, Professor, zoology Department, Michigan 'State,
University, East Lansing, Michigan
DR, RODNEY FUJITA, Environmental Defense Fund, New York, New York
DR. KENNETH JENKINS, Director, Molecular Ecology Institute, California State
University, Long Beach, California
DR. ANNE MCELROY, SUNY at stoney Brook, Stoney Brook, New York
' DR. WILLIAM MelNTYRE, Professor, VA Institute of Marine Science, College of
William and Mary, Gloucester, Virginia
DR. BETTY H. OLSON, Professor, University of California, Irvine, Irvine,
California ,
DR. C. HERBERT WARD, Professor, Rice University, Houston, Texas
SCIENCE ADVISORY BOARD STAFF
DR. EDWARD 5. BINDIRj Designated Federal Official, US EPA* Science Advisory
Board, 401 M Street, S.W. (A-101F), Washington, D.C. 20460
MRS. MARCIA K. JOLLY (MARCY), Secretary to the Designated Federal Official
-------
Distribution List
Deputy Administrator
Assistant Administrators
EPA Regional Administrators
EPA Laboratory Directors
Director, office of Emergency and Remedial Response (Superfund),
Director, Office of Solid Waste, QSWER
Director, Office of Environmental Processes and Effects Research
Director, Environmental Research Laboratory-Gulf Breeze
EPA Headquarters Library
EPA Regional Libraries
EPA Laboratory Libraries
-------
ATTACHMENT 1
MEMORANDUM . MW 1 5 199?
SUBJECT: Charge to the Science Advisory Board for Review of the
Alaska Bioremediation Project
FROM: John M. Skinner
Deputy Assistant Administrator
for Research and Development (RD-672)
TO: Donald G. Barnes, Director
Science Advisory Board
I offer the following questions as a charge to the panel
that will be reviewing the results and conclusions generated in
the Alaska Bioremediation Project;
o Do the data and interpretations adequately support the
conclusions that bioremediation was directly
responsible for the enhanced disappearance of oil from
the beaches and that under similar conditions,
disappearance of oil can be enhanced on other types of
contaminated beaches?
o Based an disappearance rates for oil residues, coupled
with large changes in oil composition, did
bioremediation substantially reduce cleanup time of
oil-contaminated beaches, thereby justifying its large
scale use on the Prince William Sound shoreline?
o Were adequate supporting research studies performed to
allow proper interpretation of field results?
o Were the selection and testing procedures for fertilizers
appropriate for the demonstration of bioremediation on
Prince William Sound beaches? Does SAB agree that these
procedure have future application?
o Was sufficient research performed and field information
collected to conclude that bioremediation in Prince
William Sound was safe and did not result in any
adverse ecological effects? Are the methods developed
for assessing the environmental safety of bioremediation
on oil-contaminated beaches appropriate as a foundation
for future assessments?
-------
Based on the response of the spilled oil to
bioremediation, should future responses be planned and
contingencies developed? Is the data also generated from
the Alaska Bioremediation Project sufficient to use as
the basis for the development of bioremediation
strategies that can be applied to future oil spills? Is
it appropriate to develop measures of effectiveness for
future oil spill bioremediation efforts using combination
of gravimetric, chemical (oil and nutrients), and
microbial (biomass and activity) measurements?
------- |