SOME ASPECTS OF THE ENVIRONMENTAL SAFETY OF COMMERCIAL OIL SPILL BIOREMEDIATION AGENTS
Jeffrey L. Kavanaugha, C. Richard Cripeb, Carol B. Danielsb, Wallace T. Gilliamb, Rochelle Araujoc and Joe E. Lepod
aUniversity of West Florida, Center for Environmental Diagnostics and Bioremediation, 11000 University Parkway, Pensacola, FL, fax (904) 474-3130; bU.S. Environmental Protection Agency, Gulf Ecology Division, Gulf Breeze, FL; cU.S. Environmental Protection Agency, Ecosystems Research Division, Athens, GA; and dUniversity of West Florida, Center for Environmental Diagnostics and Bioremediation, Pensacola, FL
SUMMARY
We have developed protocols and test systems to evaluate the efficacy and environmental safety of commercial bioremediation agents (CBAs) intended for application to oil spills. The environmental safety protocols were tested with CBAs composed of a variety of ingredients including: nutrients, microorganisms, surfactants, enzymes, or a combination of these. CBAs were prepared as specified by the manufacturers for application to an oil spill. Toxicity of each CBA was evaluated using 7 d chronic estimator tests with a fish, the inland silverside, Menidia beryllina, and the mysid, Mysidopsis bahia (a shrimp-like crustacean). Both tests evaluated survival and growth, and the mysid test included a measurement of egg production. Additional data are reported on the toxicity of effluents collected from flow-through test systems that modeled interactions between a CBA and oil; 7 d mysid toxicity tests were used to determine the effects of CBAs in effluents from a simulated open water system, as well as from a simulated marsh system. Finally, we examined the use of a biological endpoint (effective mortality of a burrowing amphipod, Leptocheirus plumulosus), as an alternative to chemical analysis for measuring bioremediation effectiveness in a simulated beach test system.
The following findings are discussed:
Key words: Bioremediation, Mysidopsis bahia, Menidia beryllina, Leptocheirus plumulosus, toxicity, crude oil
INTRODUCTION
Bioremediation is a technology for removal of toxic chemicals from environmental media (water, soil, sediment, etc.) through the controlled use of biodegradation, either by adding microorganisms or stimulating the activity/populations of indigenous degraders. Following the Exxon Valdez oil spill in Prince William Sound, Alaska, on March 24, 1989, the U.S. EPA and Exxon tested the potential for bioremediation of spilled oil by applying fertilizers to oiled beaches on Knight Island in Prince William Sound (Pritchard and Costa, 1991). The results of studies in Prince William Sound suggested that this approach enhanced indigenous microorganism biodegradation rates by two- to three-fold. Due in part to the success of these field demonstrations, commercial bioremediation agents (CBAs) have proliferated, but little is known about the environmental risks associated with the use of these products.
Reducing the impact of oil spills through application of CBAs raises a number of questions concerning environmental safety, since any cleanup technology requiring the addition of chemicals or other substances into aquatic habitats could potentially harm an already affected environment. Although the presence of petroleum hydrocarbons may contribute to toxicity, the CBA, with its associated chemical constituents (e.g., nutrients, dispersants, enzymes, microorganisms, and inert ingredients), may add to this toxicity. Also, the interaction of the CBA with oil may cause adverse environmental effects either through an increase in the concentration of available petroleum hydrocarbons (e.g., through biosurfactant activity) or by the generation of toxic metabolites.
No standardized methods exist for evaluating these types of agents for effectiveness or environmental safety. Thus, the U.S. EPA, in cooperation with the National Environmental Technology Applications Center (NETAC) at the University of Pittsburgh, convened a panel of experts from business, government and academic sectors to develop standardized testing protocols for oil-spill CBAs. By establishing standard testing protocols, a defined procedure would be available for evaluating the efficacy and safety of CBAs. Data generated by the protocol could be placed on the EPA's National Oil and Hazardous Substances Pollution Contingency Plan Product Schedule (U.S. EPA, 1992a). The data could also be made available to Regional Response Teams (RRTs) which are responsible for planning and preparedness under the Comprehensive Environmental Response, Compensation, and Liability Act of 1980.
The current protocol for assessing the efficacy and environmental safety of CBAs consists of a series of levels, or tiers, which increases in environmental realism (NETAC, 1993). In the Base Tier, the vendor provides information on product safety including formulation and unacceptable chemical or biological components such as pathogenic microorganisms. Tier I is a feasibility assessment concerning the manufacturer s production capabilities, a description of how the product will be used, and information on previous usage. The Tier II safety evaluation consists of 7 d chronic estimator toxicity tests with the CBA (without oil) in natural seawater using as test organisms the mysid, Mysidopsis bahia (a shrimp-like crustacean), and a fish, the inland silverside, Menidia beryllina. These organisms are important estuarine species for which much published literature exists. Additional Tier II toxicity tests evaluate the potential for interaction between the product and the water-soluble-fraction of a weathered crude oil. The Tier II efficacy evaluation uses shake flasks to compare the degradation of artificially-weathered crude oil in natural seawater with and without a CBA. In Tier III, products are applied to flow-through biodegradation test systems that attempt to simulate an oil spill in the open ocean, on an intertidal beach, or in a saltwater marsh. Daily collections of effluents from the biodegradation test systems are used for test solution renewals in 7 d mysid chronic estimator toxicity tests. Efficacy of oil biodegradation in Tier III is measured in the test system at termination by comparing reductions of chemical endpoints (e.g., oil residue weight, n-alkanes and aromatic compounds) from oiled systems with and without a CBA added. Tier IV field studies would be used to further validate these protocols.
In addition to standard environmental risk applications, there is a potential to use toxicology (along with analytical chemistry) to assess bioremediation efficacy (Middaugh et al., 1994). For instance, information that contaminated material, such as sediment, becomes less toxic and thus more readily recolonized after bioremediation compared to untreated material, would aid RRTs and environmental scientists in addressing the ultimate utility of bioremediation, given its risk and expense.
The goal of our protocol research was to establish procedures and test systems to evaluate the efficacy and environmental safety of CBAs. A variety of CBAs encompassing different bioremediation technologies were used to evaluate the protocols and test systems, we did not attempt to generate efficacy or toxicity rankings for specific products. This manuscript summarizes the results from selected experiments conducted during the protocol development process, providing novel evidence of the relatively low intrinsic toxicity of a range of CBAs. We report here the utility of using model test systems to simulate biodegradation in aquatic habitats to enhance our knowledge of product toxicity and efficacy for an emerging technology. In addition, we provide information on the potential use of a biological endpoint as a tool for evaluating bioremediation effectiveness.
MATERIALS AND METHODS
NETAC supplied CBAs to the EPA laboratories without trade names to ensure their use as blind treatments in the protocol validation process. The products included one or more of these ingredients: nutrients, microorganisms, surfactants, enzymes, or various inert materials (Table 1). Due to the complex nature of the bioremediation products, quantification of CBA concentrations was not feasible. Nominal concentrations of CBAs are used in these tests and refer to the whole product prepared as recommended by the manufacturer for application to an oil spill. An artificially-weathered oil, ANS521, was chosen for use in these protocols because it seemed unlikely that a CBA would be applied to an oil spill immediately, i.e., a number of days would pass before bioremediation would be attempted.
Toxicity of CBAs alone. Toxicity tests to determine the intrinsic toxicity of the CBAs alone followed standard guidelines for conducting 7 d chronic estimator tests with M. bahia and M. beryllina(U.S. EPA, 1992b). Test conditions are summarized in Table 2.
Simulated Open-Water Test System
Efficacy Test System. Test systems consisted of glass jars (8.6 cm i.d.) containing approximately 100 ml of 30 to 36 natural seawater (the salinity of collected seawater varied slightly between tests) and a 2.5 ml slick of ANS521 (0.5 mm nominal thickness). Both the control (containing oil), and the treatment (CBA applied to the oil slick according to the manufacturer s recommendation), were prepared in triplicate. The water column was gently stirred without disruption of the oil layer during the 7 d test period. Pumps provided a continuous flow of 800 ml of seawater per day through the test system.
Toxicity Testing. Each day, 400 ml of effluent from each replicate was pooled within controls and CBA treatments. These pooled samples were used daily to replace test solutions in mysid toxicity tests. Table 2 shows the conditions for mysid toxicity tests (U.S. EPA, 1992b). Effluents collected from the efficacy test systems were diluted to 20 salinity, the salinity used for the mysid cultures. Thus, the maximum possible effluent concentration tested for toxicity ranged from 66 % to 55 %, depending on the seawater used for efficacy tests. The effects of effluent from the CBA plus oil treatment were compared to the oil-only control.
Simulated Marsh Test System
Efficacy Test System. Test systems consisted of glass columns (10 cm i.d.) containing 5 cm of natural marsh sediment collected from Sapelo Island, GA. Timer-controlled pumps simulated two daily 360 ml tidal cycles of 20 natural seawater during each 24 h period. A volume of 3.9 ml of ANS521 was layered on the water surface of each container, forming a slick with a 0.5 mm nominal thickness. Triplicate test systems consisting of a control (containing oil), and a treatment (CBA applied to the oil slick according to the manufacturer's recommendation), were operated continuously for 28 d.
Toxicity Testing. Each day, effluents from replicate test systems were pooled to be used for test solution replacement in mysid toxicity tests. Four consecutive 7 d toxicity tests were conducted during the 28 d period of test system operation. Table 2 describes the conditions for these tests. Effluent test concentrations from the treated test systems were: 100, 30, 9.0, 2.7, and 0.8 %. Because the simulated marsh test systems and toxicity testing were conducted at different locations, effluents could not be used for toxicity test solution renewals on the day of collection. Effluents were shipped at 4°C and used within 36 h, according to effluent toxicity testing guidelines (U.S. EPA, 1992b). The effects of effluent from the CBA plus oil treatment were compared to the oil-only control.
Statistical Analysis of M. bahia and M. beryllina Toxicity Tests.
Results were evaluated using TOXSTAT 3.3 (Gulley et al., 1991) to determine the lowest-observed-effect concentration (LOEC, the lowest test concentration significantly different ( = 0.05) from the control) and the highest concentration not significant different from the control. The no-observed-effect concentration (NOEC) is one treatment below the LOEC.
Simulated Beach Test System.
This study included several variations from the standard protocol to address the potential use of a biological endpoint as a tool for evaluating bioremediation effectiveness. Mysid toxicity tests were not conducted with effluents; instead, the burrowing amphipod, Leptocheirus plumulosus was used to assess toxicity of the bioremediated sediments. The amount of oil added to the test system was reduced from a level which would provide a 0.5 mm slick (1.9 ml), to levels less toxic to L. plumulosus. Instead of a CBA, a mixture of oil-degrading bacteria and nutrients was used as a surrogate. This mixture of three oil-degrading bacteria was selected for their ability to degrade certain aromatic and heterocyclic compounds (e.g., phenanthrenes, fluorenes, naphthalenes, dibenzothiophene, naphthobenzothiophene), as well as straight-chain alkanes and some isoprenoids (phytane, pristane).
Efficacy Test System. The test system consisted of 250 ml fluorocarbon beakers containing sand which were placed on an orbital shaker to simulate gentle wave action (70 rpm). Timer-controlled pumps produced two daily 190 ml tidal cycles of 30 unfiltered natural seawater for 28 d. Two treatments with either 0.05 ml (0.04 g) or 0.11 ml (0.10 g) of oil added at high tide were inoculated weekly at loe tide with oil-degrading bacteria and nutrients. Both an unoiled control (without oil, bacteria, or nutrients), and an oiled control with the two amounts of oil but no bacteria or nutrients were also tested. Each control and treatment consisted of eight replicates. Four replicates were used for analytical chemistry (one was analyzed for total weight of remaining oil, and three were used for GC/MS analysis); the remaining four replicates were used in a sediment toxicity test.
GC/MS analyses of oil remaining in the sand at the termination of the test are presented to contrast analytical chemistry results with a biological endpoint for measurement of bioremediation success. The oil remaining in four of the replicate test systems was extracted with methylene chloride. The extract from one was evaporated to dryness for gravimetric evaluation. For the other three test systems, asphaltenic and polar compounds were removed from the extract with an alumina column. The remaining oil components were analyzed by GC/MS for selected n-alkanes, isoprenoids, and aromatic compounds and quantified as a percentage of the components in undegraded ANS521. Oiled controls and inoculated treatments were statistically compared for remaining analytes using a one-tailed Student's t-test. Data were checked for equal variances with a chi-square test, and an unequal variance t-test was used when appropriate.
Toxicity Testing. At the end of the 28 d efficacy test, L. plumulosus was exposed for 10 d to sediment from four replicate beakers of each of the controls and treatments. Filtered, natural seawater adjusted to 20 salinity was used in the sediment test. Each beaker was aerated and contained 20 subadult (3-5 mm) amphipods. Effective mortality (ASTM, 1990) was measured at the end of the 10 d acute sediment toxicity test by counting dead animals plus the number still alive but unable to rebury in clean sediments after being transferred from the contaminated sediments. Measurement of effective mortality provides an environmentally-relevant evaluation of sublethal effects, since oil-covered animals unable to burrow probably cannot carry out normal activities such as feeding, escaping predators and reproduction. A Student s t-test was used to compare effective mortality in bioremediated test systems to that in the oiled controls. The unoiled control indicated test organism health as well as compatibility with test system design, e.g., sediment particle size.
RESULTS AND DISCUSSION
Toxicity of the CBAs Alone
The intrinsic toxicity of the CBAs tested was relatively low to M. bahia and M. beryllina in 7 d chronic estimator toxicity tests (Tier II): NOECs for the CBAs tested ranged from 75 to 20,000 mg/L for M. bahia and from 150 to 5000 mg/L for M. beryllina (Table 3). Terminology for the degree of concern (e.g., high, medium, low) associated with values from aquatic toxicity tests can be found in Smrchek et al.(1993) and U.S. Fish and Wildlife Service (1984). The most toxic CBAs tested were a nutrient (D) and a surfactant (I) with NOECs of 75 and 110 mg/L, respectively. For the two CBAs tested with both test organisms (products D and F), M. bahia was more sensitive than M. beryllina.
Effluent Toxicity in the Simulated Open-Water Test System
Application of five of the six CBAs to oil in the open-water test system resulted in effluents with no apparent toxicity to mysids (Table 4). If CBAs do not cause toxicity in these test systems, it seems unlikely they would be toxic if applied to an actual oil spill, since under field conditions there would probably be greater dilution. Thus, these test systems probably provide a conservative estimate of environmental safety.
CBA I was the only product to show effluent toxicity; it was also the only CBA applied at a level high enough to potentially cause significant acute mortality based on Tier II toxicity data of the product alone (Table 5). Only 18 % mortality was observed within 24 h in the highest effluent concentration tested with M. bahia. However, much higher mortality would have been predicted based on the Tier II toxicity of CBA I and its application rate in the open-water test system. Assuming all of CBA I washed out of the test system in 24 h, there would have been 1,562 mg/L of the product in the highest effluent concentration tested. This assumption is based on the volume of the mixed water column and flow rate for the test system in which eight volume additions per day provide a >99 % molecular replacement in 24 h. Tier II data presented in Table 5 show that 100 % mortality occurred within 24 h to mysids at concentrations of 1000 mg/L of CBA I. Either all of CBA I did not wash out, or an interaction occurred between the CBA and, for example, the oil or the oil water-soluble-fraction to decrease the expected effluent toxicity.
Effluent Toxicity in the Simulated Marsh Test System
CBA D. During the first week of operation of the marsh test system, we noted a significant decrease in mysid growth and fecundity, but not in survival, in the undiluted effluents from the test systems with CBA D. No other effects were measured during the remaining 3 wk of effluent testing. Since essentially all the water drains from the test system between the two daily tidal cycles, nearly all the product could be expected to flush out in 24 h. Assuming all the CBA washed out of the test system within 24 h of application, the maximum estimated concentration of the product in the effluent would have been 504 mg/L. Exposure of mysids to CBA D alone at 500 mg/L produced 90 % mortality within 24 h. Because there was no survival effect in the effluent test, we suggest that either all the CBA did not leave the system or toxicity of the effluent was decreased by, for example, nutrient depletion by microorganisms or product sorption to sediments or oil.
CBA J. As with CBA D above, for the first week of test system operation in which CBA J was applied, mortality of mysids was less than would be predicted based on toxicity data for the product alone. Effluent from the marsh test system with CBA J did not cause significant mortality to mysids during Week 1. Assuming all the product washed out of the test system within 24 h, the maximum estimated concentration in the effluent would have been 105,000 mg/L. Toxicity data for CBA J (Tier II) at 100,000 mg/L produced 13 % mortality of mysids within 24 h. Only 2.5 % mortality was observed after the first day of effluent testing during Week 1, which again suggests that either not all of CBA J was transported into the effluent in 24 h, or an interaction between the CBA and a component or components in the test system had decreased the expected toxicity of the effluent.
In contrast, during the fourth week of toxicity testing, we observed a significant mysid mortality (45%) that could not be explained solely by the concentration of CBA J (including its nutrients) in the system. If the product alone had been responsible for the mortality of M. bahia in Week 4, then survival should also have been affected during Week 1, when 10 times more nutrients were added (Day 0) than during any subsequent CBA application (Weeks 2, 3, or 4). Survival effects measured in the effluents during Week 4 could have resulted from toxic metabolites, an increase of petroleum hydrocarbons in the effluent due to surfactant-producing microorganisms, or a combination of these factors. Changes in the population dynamics and/or physiology of the microbial assemblage within the marsh system may also have contributed to the observed lag between product application and the observed biological response. The delayed biological response was reproducible (data not shown).
Sediment Toxicity in the Simulated Beach Test System
The addition of 40 mg and 100 mg of oil to the beach test systems increased effective mortality of L. plumulosus to 69 % and 79 % in the oiled controls, respectively (Figure 1). Because effective mortality in the oiled controls was considerably more than 0 %, there was a potential to significantly decrease effective mortality in oiled sediments by treatment with nutrients and oil-degrading bacteria. Chemical analysis, typically used for evaluating bioremediation effectiveness, indicated substantial losses of total oil and selected analytes from both the oiled control and the treated test systems. There was a 19.1 % and 26.3 % weight loss of oil in the 100 mg and 40 mg oiled controls, respectively; and a 28.5 % and 31.5 % weight loss of oil occurred in the 100 mg and 40 mg treatments (n = 1), respectively. Table 6 shows the results of GC/MS analysis for the oil remaining in the beach test system. There were extensive losses of selected n-alkane and aromatic analytes from the oiled controls and statistically significant (p 0.05) losses of these analytes in test systems inoculated with bacteria and nutrients, relative to the oiled controls.
As assessed by the conventional endpoint of analytical chemistry, these experiments depict an apparently successful bioremediation of oiled sediments. However, we found no difference in the effective mortality of L. plumulosus between oiled controls and treatments receiving weekly inoculations of nutrients and oil-degrading bacteria (Figure 1). Mean effective mortality of L. plumulosus in the unoiled, uninoculated sediments was only 7 %, indicating no adverse effect of the test system on the amphipods. Amphipod effective mortality may be controlled by physical effects of highly recalcitrant fractions, such as resins, which may remain in the bioremediated sediments. In addition, metabolites created through oil biodegradation may have contributed a similar amount of toxicity as did the components that were degraded. In any case, these results suggest that the exclusive use of analytical chemistry for efficacy determinations may need to be re-examined.
In conclusion, these studies indicated that most of the CBAs tested exhibited relatively low toxicity either alone or in simulated test systems interacting with a weathered oil when they were applied at the manufacturer s recommended application rate. Data from the simulated marsh system, however, suggested that some unidentified factor(s) increased effluent toxicity. We were unable to use a biological endpoint to demonstrate bioremediation effectiveness in a simulated beach test system. The implications of these last two findings warrant further investigations.
ACKNOWLEDGEMENTS
This work was supported by cooperative agreements (CR-818991-01 and CR-822236-01-0) between the University of West Florida Center for Environmental Diagnostics and Bioremediation and the U.S. EPA Gulf Ecology Division, Gulf Breeze, FL. P. Hap Pritchard and C. Richard Cripe were the respective EPA Project Officers for the cooperative agreements.
Others who contributed ideas or technical assistance: Wanda Boyd, Ahmet Bulbulkaya, Charles Bundrick, Katharine Ruopp-Edwards, Timothy Gibson, Patricia Hancock, Joanne Konstantopoulos, Anthony Mellone, Geneva Norton, Philip Turner, David Whiting, Vicki Whiting, and Shiying Zhang.
REFERENCES
ASTM (1990) Standard guide for conducting 10-day static sediment toxicity tests with marine and estuarine amphipods. In: Annual book of ASTM Standards, ASTM E 1367-90. American Society of Testing and Materials, Philadelphia, PA.
Gulley, DD, Boelter AM, Bergman HL (1991). TOXSTAT 3.3. Fish Physiology and Toxicology Laboratory, Dept. Zool. and Physiol., Univ. WY, Laramie, WY.
International Organization for Standardization (1989) Draft International Standard ISO/DIS 8708. Crude petroleum oil: determination of distillation characteristics using 15 theoretical plates columns.
Middaugh, DP, Lantz SE, Heard CS, Mueller JG (1994) Field-scale testing of a two-stage bioreactor for removal of creosote and pentachlorophenol from groundwater: chemical and biological assessment. Arch. Environ. Contam. Toxicol. 26: 320-328.
National Environmental Technology Application Center (NETAC) (1993) Evaluation methods manual for oil spill response bioremediation agents. University of Pittsburgh Applied Research Center, Pittsburgh, PA.
Pritchard PH, Costa CF (1991) EPA's Alaska oil spill bioremediation project. Environ. Sci. Technol. 25: 372-379.
Smrchek J, Clements R, Morcock R, Rabert W (1993) Assessing ecological hazard under TSCA: Methods and evaluation of data. Environmental Toxicology and Risk Assessment, ASTM STP 1179, Landis WG, Hughes JS, Lewis MA, Eds., American Society for Testing and Materials, Philadelphia, PA, pp. 22-39.
U.S. Environmental Protection Agency (1992a) National oil and hazardous substances pollution contingency plan. 40 CFR Part 300.
U.S. Environmental Protection Agency (1992b) Short-term methods for estimating the chronic toxicity of effluents and receiving waters to marine and estuarine organisms. EPA/600/4-91/021. Environmental Monitoring and Support Laboratory, Cincinnati, OH.
U.S. Fish and Wildlife Service (1984) Toxic chemicals: How do they rate? Research Information Bulletin No. 84-78. Columbia National Fisheries Research Laboratory, Columbia, MO.
Figure 1. Effective mortality (reburial) of Leptocheirus plumulosus after a 10 day sediment toxicity test using control, oiled control, and treated sediments following 28 days of test system operation.