Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 63
4 Evaluating In Situ Bioremediation Showing that a bioremediation project is working requires evidence not only that contaminant concentrations have decreased but also that microbes caused the decrease. Although other processes may contribute to site cleanup during a bioremediation, the microbes should be critical in meeting cleanup goals. Without evidence of microbial involvement, there is no way to verify that the contaminant did not simply volatilize, migrate off site, sorb to subsurface solids, or change form via abiotic chemical reactions. This chapter discusses a strategy for evaluating the effectiveness of in situ bioremediation projects, based on showing that microbes were responsible for declining contaminant concentrations. Regulators and buyers of bioremediation services can use the strategy to evaluate the soundness of a proposed or ongoing in situ bioremediation system. Researchers can apply the strategy to evaluate the results of field tests. A THREE-PART STRATEGY FOR ''PROVING" IN SITU BIOREMEDIATION To answer the question "What proves in situ bioremediation?", one must recognize that only under rare circumstances is proof of in situ bioremediation unequivocal. In the majority of cases the complexities of contaminant mixtures, their hydrogeochemical settings, and competing abiotic mechanisms of contaminant loss make it a
OCR for page 64
challenge to identify biodegradation processes. Unlike controlled laboratory experiments in which measurements can usually be interpreted easily, cause-and-effect relationships are often difficult to establish at field sites. Furthermore, certain data that may convince some authorities of in situ bioremediation may not convince others. Although proving microbial involvement in cleanup with complete certainty is seldom possible, the weight of the evidence should point to microbes as key actors in the cleanup. Because one measurement is seldom adequate, the evaluation strategy must build a consistent, logical case that relies on convergent lines of independent evidence taken from the field site itself. The general strategy for demonstrating that in situ bioremediation is working should include three types of evidence: documented loss of contaminants from the site, laboratory assays showing that microorganisms in site samples have the potential to transform the contaminants under the expected site conditions, and one or more pieces of evidence showing that the biodegradation potential is actually realized in the field. The strategy applies not only to bioremediation projects in the implementation phase but also to those in the testing phase. The strategy is not just for research purposes. Every well-designed bioremediation project should show evidence of meeting the strategy's three requirements. Thus, regulators and buyers of bioremediation services can use the strategy to evaluate whether a proposed or ongoing bioremediation project is sound. The first type of evidence in the strategy—showing decreasing contaminant concentrations—comes from standard sampling of the ground water and soil over time as cleanup progresses. The second type of evidence—showing the potential for microorganisms to degrade the contaminants—is also relatively simple to provide. In most cases it requires taking microbes from the field and showing that they can degrade the contaminant when grown in a well-controlled laboratory vessel. For some cases, lab studies may not be needed when a body of peer-reviewed published studies documents that the compounds are easily and commonly biodegraded. The most difficult evidence to gather is the third type—showing that the biodegradation potential demonstrated in the laboratory is being realized in the field. Evidence of field biodegradation is essential: data showing that organisms are capable of degrading the contaminant in the laboratory are not sufficient because the organisms may not perform the same tasks under the less hospitable field con-
OCR for page 65
ditions. A variety of techniques, explained below, exist for demonstrating field biodegradation. TECHNIQUES FOR DEMONSTRATING BIODEGRADATION IN THE FIELD The goal of the techniques for demonstrating field biodegradation is to show that characteristics of the site's chemistry or microbial population change in ways that one would predict if bioremediation were occurring. The environmental changes measured in these tests should correlate to documented contaminant loss over time. No one technique alone can show with complete certainty that biodegradation is the primary reason for declining contaminant concentrations in the field. The wider the variety of techniques used, the stronger the case for successful bioremediation. As an example, Box 4-1 describes a site where several tests were combined. There are two types of sample-based techniques for demonstrating field biodegradation: measurements of field samples and experiments run in the field. In most bioremediation scenarios a third technique, modeling experiments, provides an improved understanding of the fate of contaminants. Examples of field measurements, field experiments, and modeling experiments are described below. These examples provide general guidance about which types of tests are appropriate. Detailed experimental protocols for carrying out the tests need to be developed and will vary depending on the types of contaminants present, the geological characteristics of the site, and the level of rigor desired in the evaluation. Measurements of Field Samples A number of techniques for documenting in situ bioremediation involve removing samples of soil and water from the site and bringing them to the lab for chemical or microbiological analysis. Many of these techniques require comparing conditions at the site once bioremediation is under way with site conditions under baseline circumstances, when bioremediation is not occurring. Baseline conditions can be established in two ways. The first method is to analyze samples from a location that is hydrogeologically similar to the area being treated but is either uncontaminated or is outside the zone of influence of the bioremediation system. The second method is to gather samples before starting the bioremediation system and compare them with samples gathered at several time points after the system is operating. This second method applies only to engineered
OCR for page 66
BOX 4-1 PROVING ENGINEERED BIOREMEDIATION OF CHLORINATED SOLVENTS IN A FIELD TEST— MOFFETT NAVAL AIR STATION, CALIFORNIA Researchers at Stanford University conducted a field demonstration to evaluate the potential for using cometabolism for in situ bioremediation of chlorinated solvents. The work done in this field demonstration shows how a variety of tests can be combined to evaluate whether new bioremediation processes researched in the lab can be applied successfully in the field. The demonstration site was a highly instrumented and well-characterized confined sand-and-gravel aquifer at Moffett Naval Air Station in Mountain View, California. To this site the researchers purposely added chlorinated solvents in a carefully controlled way that ensured that the solvents would not migrate beyond the research plot. Chlorinated solvents cannot support microbial growth on their own, but, if supplied with methane, a special class of organisms can destroy the contaminants through cometabolism (see Chapter 2). Thus, at this site, researchers stimulated native organisms by adding oxygen and methane. When stimulated, the organisms destroyed significant quantities of the chlorinated solvents. The researchers evaluated the success of their work using tests that meet the three criteria discussed in this chapter: Documented loss of contaminants: The researchers documented that 95 percent of the vinyl chloride, 85 percent of the trans-1,2-dichloroethylene, 40 percent of the cis-1,2-dichloroethylene, and 20 percent of the trichloroethylene added to the site were transformed. Laboratory assays showing that microorganisms have the potential to degrade the contaminants: When cores of the aquifer removed to the lab were exposed to methane and oxygen, the methane and oxygen were used up, showing that the cores contained bacteria that thrive on methane (methanotrophs). Previous research had shown that methanotrophs can cometabolize chlorinated solvents. Evidence that biodegradation potential is realized in the field: The researchers used a variety of methods to demonstrate biodegradation in the field. First, they showed that before the methanotrophs were stimulated with methane and oxygen, destruction of trichloroethylene was minimal. Second, they performed conservative tracer tests with bromine to show that the methane and oxygen added to the site were not disappearing by physical transport but were being used by microorganisms. Third, they identified microbial breakdown products from the solvents in aquifer samples. Fourth, they used models to show that theoretical estimates of biodegradation rates could account for contaminant loss in the field.
OCR for page 67
References Roberts, P. V., G. C. Hopkins, D. M. Mackay, and L. Semprini. 1990. A field evaluation of in situ biodegradation of chlorinated ethenes: Part 1—Methodology and field site characterization. Groundwater 28:591-604. Semprini, L., P. V. Roberts, G. D. Hopkins, and P. L. McCarty. 1990. A field evaluation of in situ biodegradation of chlorinated ethenes: Part 2—The results of biostimulation and biotransformation experiments. Groundwater 28:715-727. bioremediation systems, because the "starting time" of intrinsic bioremediation occurs whenever the contaminant enters the subsurface and therefore cannot be controlled. Following are several types of analyses that may be performed on samples removed from the field. Number of Bacteria When microbes metabolize contaminants, they usually reproduce. (In general, the larger the number of active microbes, the more quickly the contaminants will be degraded.) Thus, samples correlating contaminant loss with an increase in the number of contaminant-degrading bacteria above the normal conditions provide one indicator that active bioremediation may be occurring in the field. When contaminant biodegradation rates are low, such as when contaminant levels are low or biodegradable components are inaccessible, increases in the number of bacteria may not be great enough to detect above background levels, given the error in sampling and measurement techniques. Thus, the absence of a large increase in bacterial numbers does not necessarily mean that bioremediation is unsuccessful. The first issue for determining the size of the bacterial population is what to sample. In principle, the best samples include the solid matrix (the soil and rock that hold the ground water) and the associated pore water. Because most microorganisms are attached to solid surfaces or are trapped in the intersticies between soil grains, sampling only the water normally underestimates the total number of bacteria, sometimes by as much as several orders of magnitude. In addition, water samples may misrepresent the distribution of microbial types, because a water sample may contain only those bacteria easily dislodged from surfaces or that can be transported in the moving ground water. While obtaining soil samples from the earth's surface is not diffi-
OCR for page 68
cult, subsurface sampling is expensive and time consuming. Subsurface samples are obtained by removing cylindrical cores from below ground. A major effort is required to prevent microbial contamination of the sample during the coring operation and while handling the sample. Whenever possible, sampling equipment should be sterilized before use. Contamination from extraneous material, including the air, soil, and human contact, should be prevented. Although ground water samples have important deficiencies, they have a role as semiquantitative indicators of microbial numbers. Major increases in the number of bacteria in the ground water usually correlate to large increases in the total number of bacteria in the subsurface. The main advantages of ground water samples are that they can be taken repeatedly from the same location and that they are relatively inexpensive. The second issue for determining bacterial numbers is how to assay for the bacteria. Several standard and emerging techniques, each of which has advantages and disadvantages, are available: Direct microscopic counting is a traditional technique that involves using a microscope to view the sample and count the bacteria, which are distinguished from solid debris based on their size and shape. Microscopic counting is greatly aided by the use of the acridine orange stain and epifluorescence microscopy, which make intact bacteria stand out from other particles. Microscopic enumeration can be tedious and requires technician experience, particularly when the sample contains solids. The technique provides data on total bacterial counts but does not give information on cell types or metabolic activity. The INT activity test can enhance direct microscopic counting by identifying only those bacteria active in electron transport, the main force behind all metabolism. If the sample (or bacteria harvested from the sample) is incubated with a tetrazolium salt under controlled conditions, actively respiring bacteria transfer electrons to the tetrazolium salt, forming purple INT crystals that can be observed microscopically inside the metabolically active bacteria. Plate counts, another standard technique, can be used to quantify the number of bacteria able to grow on a prescribed set of nutrients and substrates immobilized in a solid medium. The solid medium is created from a liquid solution with the appropriate nutrients and substrates, solidified with agar to form a gel. A sample containing the bacteria of interest is spread thinly over the surface of the gel. After the plate is incubated, visible bacterial colonies form, and the colonies can be counted to indicate the number of metabolically ac-
OCR for page 69
tive bacteria in the original sample. Because plate counting requires significant growth to form visible colonies, the method often underestimates the number and diversity of bacteria. On the other hand, considerable information on the metabolic capability of bacteria can be obtained by using a range of growth substrates to prepare the media. The most-probable-number (MPN) technique also relies on significant growth in prescribed media, but enumeration is carried out through a statistical analysis. Instead of counting colonies for a few incubations on solid media, the MPN technique uses a large number of incubations from portions of the sample diluted to prescribed levels in a nutrient solution. Based on simple statistics and the number of diluted liquid samples that show evidence of bacterial growth, the number of bacteria in the original undiluted sample can be calculated. Although details of the MPN and plate-counting methods differ, both techniques have the same general advantages and disadvantages. Modern tools of biochemistry and molecular biology are becoming available to provide more precise ways of identifying and enumerating bacteria in site samples. The tools exploit the growing understanding of genetically determined characteristics of particular cell components: Oligonucleotide probes are small pieces of deoxyribonucleic acid (DNA) that can identify bacteria by the unique sequence of molecules coded in their genes. The small DNA probe bonds with a complementary region of the target cell's genetic material, and the amount of bound probe can be quantified and correlated with the number of cells. Advanced probing techniques to count the cells in intact samples are under development. Probing is a very powerful technique for identifying which types of bacteria are present, as long as the genetic sequences are known for the target bacteria. This knowledge is available for some common types of bacteria and often is known for genetically engineered microorganisms or other specialized microorganisms that might be used as part of a bioaugmentation strategy. Probing also can be used to determine whether the gene for a particular biodegradation reaction is present. The drawbacks of probing are that it requires considerable prior knowledge of the cells' genetic sequences, it is only semiquantitative in its current state, and it requires specialized equipment and knowledge. Fatty acid analysis is a second new bacterial identification technique; it uses the characteristic "signature" of fatty acids present in the membranes of all cells. The distribution of fatty acids is unique
OCR for page 70
and stable among different bacteria and therefore can be used as an identifying signature. Like gene probing, fatty acid analysis requires specialized knowledge and equipment. It is limited in its quantification ability and may have sensitivity limitations for small populations. With gene probing and fatty acid analysis, it is unnecessary to grow the bacteria in the laboratory to detect what kinds and how many are present in a sample. While holding great promise, these new methods are still in the development and testing stages. Number of Protozoans Because protozoans prey on bacteria, an increase in the number of protozoans suggests a major increase in the number of bacteria. Therefore, samples correlating contaminant loss with growth in the protozoan population can provide further evidence of active bioremediation. The protozoan population can be counted using a statistical MPN technique similar to that used for bacteria. The MPN technique for counting protozoans requires growing various dilutions of the soil or water sample in cultures containing a large number of bacterial prey. Whether protozoans grow to feed on these prey can be determined by viewing the diluted samples through a microscope. Rates of Bacterial Activity While an increase in bacterial numbers usually is a key sign that bioremediation is working, the stronger measure of success is that potential biotransformation rates are great enough to remove the contaminant rapidly or to prevent contaminant migration. Therefore, measurements demonstrating that the bacteria are capable of performing the desired reactions at significant rates help to provide evidence of successful bioremediation. The most direct means for estimating biodegradation rates is to construct laboratory microcosms with environmental conditions as close as possible to those from which the sample was taken. (See Box 4-2 for an example of the use of microcosms.) To these microcosms, field samples and the accompanying microbes (or other microbes that could be released into the field) are added. Microcosms are useful because substrate concentrations and environmental conditions can be controlled and the loss of the contaminant or other markers of biodegradation can be measured relatively easily. For many pollutants (including BTEX and PCBs), versions labeled with carbon-14 are
OCR for page 71
BOX 4-2 PROVING ENGINEERED BIOREMEDIATION OF AN OIL AND FUEL SPILL—DENVER, COLORADO In Denver, Colorado, a temporary holding tank under a garage used to service vehicles leaked crankcase oil, diesel fuel, and gasoline. The leak contaminated the surrounding soil and created a plume of benzene, toluene, ethylbenzene, and xylene (BTEX) in the ground water below. An engineered bioremediation system was installed at the site in 1989. The soil was cleaned by removing the remaining pools of leaked contaminants and by venting to supply oxygen and promote biodegradation. The ground water was treated by circulating oxygen (in the form of hydrogen peroxide), phosphorus (in the form of phosphate), and nitrogen (in the form of ammonium chloride) to promote bioremediation. By March 1992, after three years of treatment, the dissolved plume of contaminants had been nearly eliminated from the ground water. However, tests revealed that the aquifer contained a small layer that had trapped considerable quantities of BTEX. This layer is relatively impermeable and therefore had been bypassed by the fluids circulated to promote bioremediation. When the bioremediation system was shut down in 1992, long-term monitoring began to ensure that the native microbial community could act quickly enough to degrade any contamination that might leak from this layer. Although the engineered bioremediation system at this site was unable to eliminate all of the contamination, it succeeded in reducing the amount and risk of the contamination to acceptable levels. Furthermore, it is likely that microbes at the periphery of the remaining contamination will provide effective intrinsic bioremediation that will prevent the reemergence of a contaminant plume. The cleanup using the engineered bioremediation met the three criteria set forth in this report: Documented loss of contaminants: At the monitoring well closest to the gallery used to deliver oxygen and nutrients to the site, the BTEX concentration dropped from 2030 µg/l before bioremediation to 6 µg/l after bioremediation. At other monitoring wells, the concentration dropped more than an order of magnitude, to less than 46 µg/l. Laboratory assays showing that microorganisms have the potential to degrade the contaminants: Studies showed that microorganisms in the transmissive layers adjacent to the trapped contaminants could consume as much as 7 mg/l of oxygen per day, resulting in the potential destruction of as much as 2 mg/l of hydrocarbons per
OCR for page 72
day. This oxygen consumption rate was determined by placing a dewatered core from the site in a sealed glass mason jar and measuring the amount of oxygen the microbes in the core consumed in 24 hours. No direct tests—other than measuring the oxygen consumption rate—of the native microbes' ability to degrade BTEX were performed. However, the ability of subsurface microorganisms to degrade BTEX is well established (see Table 2-1), so direct lab tests were not as important for this site as for sites with contaminants for which bioremediation techniques are still emerging. Evidence that biodegradation potential is realized in the field: At this site, two types of tests provided evidence of biodegradation in the field. First, the oxygen consumption rate in microcosms constructed with cores from the site was highest when the cores came from near the layer of trapped contaminants. Thus, microbes with access to the largest supply of contaminants consumed oxygen most rapidly, supporting the expectation that bacterial growth on the hydrocarbons had been stimulated. Second, the ratio of BTEX to total petroleum hydrocarbons (TPH) was lower in the bioremediated area than in the contaminant source. Research has shown that microorganisms prefer BTEX to other components of TPH, leaving a TPH residual that is relatively low in BTEX after a successful remediation. Reference Nelson, C., R. J. Hicks, and S. D. Andrews. In press. In situ bioremediation: an integrated system approach. In Bioremediation: Field Experiences, P. E. Flathman, D. E. Jerger, and J. H. Exner, eds. Chelsea, Mich.: Lewis Publishers. available and can be used in microcosm tests to trace the pollutant's fate very precisely. Comparing the microcosm-generated biodegradation rates under a variety of conditions can provide valuable information concerning whether environmental conditions in the field are conductive for high degradation rates. The careful control and monitoring possible in microcosms make rate determinations much less ambiguous than rates measured in the field. Methods that rely on laboratory microcosms have uncertainties associated with directly extrapolating the laboratory results to the field. The delicate balance of chemical, physical, and biological relationships that influence bioremediation can change rapidly with environmental disturbances, such as to oxygen concentration, pH, and nutrient concentration. Research has shown that microbes removed to the laboratory may behave differently from those in the field—
OCR for page 73
quantitatively and qualitatively. Thus, laboratory experiments may impose artifacts that distort the interpretation of field conditions. Bacterial Adaptation Over time, bacteria at a contaminated site may develop the capability to metabolize contaminants that they were unable to transform—or that they transformed very slowly—when the contaminant was first spilled. Thus, metabolic adaptation provides evidence of bioremediation in the field. Adaptation can result from an increase in the number of bacteria able to metabolize the contaminant or from genetic or physiological changes within the individual bacteria. Microcosm studies are well suited for assessing adaptation. An increase in the rate at which microorganisms in the sample transform the contaminant in microcosm tests provides evidence that adaptation has occurred and bioremediation is working. The rate increase can be determined by comparing samples from the bioremediation zone with samples from an adjacent location or by comparing rates before and after bioremediation. Developments based on tools used in molecular biology may provide new methods for tracking whether bacteria have adapted to degrade certain contaminants. Gene probes specifically targeting degradative genes can be constructed and can, at least in principle, determine if that gene is present in a mixed population. Using probes in this manner requires knowledge of the DNA sequence in the degradative gene. For the special case when a genetically engineered microorganism is applied to a site for bioaugmentation, the engineered organism can be fitted with a reporter gene that is expressed only when a degradative gene of interest also is expressed. Thus, the protein product of the reporter gene signals (for example, by emitting light) that the degradative gene is present and is being expressed in the in situ population. Inorganic Carbon Concentration In addition to more microbes, bacteria produce inorganic carbon—usually present as gaseous CO2, dissolved CO2 or HCO3-—when they degrade organic contaminants. Therefore, samples showing enrichment of the water and gas phases with inorganic carbon indicate active biodegradation. Gas chromatography is the method of choice for determining gaseous CO2 concentrations; inorganic carbon analysis is appropriate for water samples.
OCR for page 80
dissolved O2) is the electron acceptor, bromide can be used as a conservative tracer. In this approach the bromide is added to the water circulated through the ground to supply the electron acceptor. Although conservative tracers that mimic contaminant behavior often are added to the site, they also may be fortuitously present in the contaminant. As discussed under ''Measurements of Field Samples," some contaminants contain mixtures of degradable and nonbiodegradable compounds that move through the subsurface in similar ways. When the concentration of degradable compounds drops faster than the concentration of conservative tracers, the difference can be attributed to microbial activity in the field. Labeling Contaminants A fourth type of field experiment involves monitoring the fate of "labeled" contaminants. Contaminants can be labeled by synthesizing versions in which the contaminant molecules contain a known amount of a stable isotope, usually 13C or deuterium (a hydrogen isotope). If the expected metabolic byproducts, such as inorganic carbon and intermediary metabolites, carry the same relative amounts of 13C and deuterium as the labeled contaminants, bioremediation is occurring. This technique is useful primarily for field research and not commercial bioremediation because it involves synthesizing a special version of the contaminant, which is costly, and adding it to the site, which temporarily increases the level of contamination. In addition, contaminant labeling is useful only for situations in which the contaminant source can be located. Adding the labeled compound to the wrong location may result in a false negative. Modeling Experiments A final type of technique for evaluating whether bioremediation is occurring in the field uses models—sets of mathematical equations that quantify the contaminant's fate. Models keep track of all the contaminant mass that enters the subsurface, describing how much dissolves, how much sorbs to solids, how much reacts with other chemicals, how much flushes out in the water, and how much biodegrades. The goal of using models is to see whether predictions of contaminant fate based on interpretation of the phenomena taking place during the bioremediation, as described by the model, match what is happening in the field, as determined by field sampling. Contaminated field sites can be efficiently managed with the aid of models because models provide a means for synthesizing all rel-
OCR for page 81
evant information. Furthermore, because models quantitatively link many types of measurements, they assist in evaluating the significance of a limited number of field observations. When models are sufficiently accurate, they may be powerful tools for assessing bioremediation. Depending on the type and amount of data, the stage of process understanding, and the types of questions being asked, models can vary from very simple to highly complex. For example, a conceptual model, which does not yet have mathematical equations, may be appropriate when limited data are available during initial site characterization. On the other hand, complex mathematical models, solved on high-speed computers, become possible and more appropriate as understanding of the site expands during design and operation of a bioremediation project. Types of Models Because so many complex processes interact in the subsurface, four different types of models have been developed: saturated flow, multiphase flow, geochemical, and reaction rate models. Each model describes a different suite of subsurface processes and is used in particular ways to evaluate bioremediation. Ultimately, researchers often combine two or more types of models to do a complete evaluation. Saturated Flow Models. Saturated flow models start by describing where and how fast the water flows through the saturated zone (the region below the water table). These models are derived from basic principles of conservation of fluid mass. Saturated flow is reasonably well understood, and the basic forms of these models for water flow are relatively simple, accurate, and accepted. Once the direction and velocity of water flow are known, saturated flow models can be extended to describe the movement of dissolved contaminants. These contaminant transport models are based on principles of conservation of chemical mass. When the model contains no terms for reactions, it describes the fate of a conservative tracer. The conservative material basically moves with the water flow, although it is subject to processes that disperse, or mix, the contaminants. Sorption of contaminants to the solid matrix slows the movement of the dissolved contaminants, compared to the water. Sorption effects often can be modeled simply by incorporating "retardation factors" that reflect the slower rate of transport of the contaminant rela-
OCR for page 82
tive to the water. In other cases, sorption phenomena are more complex than can be captured by a simple retardation factor and must be modeled using equations that consider sorption and desorption rates. In special cases, biodegradation reactions can be described by very simple expressions (for example, first-order decay) that are easily incorporated into the transport part of a saturated flow model. However, many biodegradation phenomena are too complex to be incorporated so simply into a saturated flow model. Special modeling tools are needed and are discussed in the section below on biological reaction rate models. Multiphase Flow Models. Whereas saturated flow models describe the flow of only one fluid, the ground water, through a porous medium, multiphase flow models describe the situation in which two or more fluids exist together in the porous medium. The fluids can be liquids or gases. The most common multiphase flow models predict the movement of water and contaminants above the water table, where a gas phase is present. This situation is called unsaturated flow. Addition of a light nonaqueous-phase liquid contaminant such as gasoline, which resides at or near the top of the water table, is a further complication that may be considered in models of unsaturated flow. Multiphase flow models also can describe the flow of dense nonaqueous-phase liquids such as chlorinated solvents, which move in a distinct mass separate from the ground water. The phenomena controlling multiphase flow are not as well understood and are much more difficult to represent mathematically than are those for water flow in the saturated zone because they involve complex interactions among solids, water, air, and nonaqueous phases. The accuracy of multiphase flow models for water direction and velocity is limited by the large number of required transport parameters. Furthermore, the modeling community has not yet reached a consensus as to which modeling approach is most valid. Despite these limitations, multiphase flow modeling provides a framework for conceptualizing the movement of fluids in the subsurface and for making order-of-magnitude estimates of fluid movement. If the direction and velocity of fluid flows can be predicted, modeling contaminant transport with multiphase flow models is similar to that for saturated flow. However, contaminant transport is complicated by the multiple phases, which introduce heterogeneities that affect dispersion and sorption. Geochemical Models. At many contaminated sites, the contaminants are subject to a significant number of different chemical reactions.
OCR for page 83
Geochemical models describe how a contaminant's chemical speciation is controlled by the thermodynamics of the many types of chemical reactions that may occur in the subsurface. Today, geochemical models are used primarily to understand the fate of inorganic compounds. For example, these models can be used to analyze the series of reactions that influence whether a particular metal will precipitate. Geochemical models also can aid in determining the availability (solubility) of nutrients and trace metals required for microbial metabolism. Although they are valuable, geochemical models have had limited use for assessing bioremediation. There are three reasons for the relatively low level of use. First, existing commercial applications of bioremediation have focused on aerobic biodegradation of petroleum hydrocarbons, a situation for which inorganic geochemistry usually is not a crucial factor. As bioremediation is applied to more complex sites, especially those containing contamination by heavy metals, the need for geochemical modeling will increase. Second, traditional geochemical models are founded on the principle of equilibrium conditions—in other words, all possible reactions are assumed to occur to their maximum possible extent. The equilibrium assumption typically is not valid for bioremediation because the key reactions are almost always controlled by kinetics—the rate at which a reaction moves toward equilibrium. Third, traditional geochemical models are very complicated and expensive to use, even when they are not connected to transport modeling. Therefore, their use has been limited to evaluating possible changes in subsurface chemicals. Biological Reaction Rate Models. Biological reaction rate models represent how quickly the microorganisms transform contaminants. They are useful for evaluating bioremediation systems because the rate at which the microbes work is the key factor influencing how much time the cleanup will take. The rate of biodegradation depends on the amount of active biomass present; the concentrations of contaminants, electron acceptors, and other "food" sources for the bacteria; and certain parameters that describe transport rates of key chemicals to the bacteria and rates of enzyme-catalyzed reactions. All of this information can be packaged into a rate expression of the form: in which qmax describes the reaction rate per unit amount of biomass for optimal conditions, X is the amount of active biomass, f(S1, S2 , .…) is a mathematical function that describes how substrate transport
OCR for page 84
and concentration reduce the rate from the optimal rate, and S1, S2,. … represent different substrates that participate in the reaction. The value of X is not necessarily constant; it can change with time and location. Keeping track of X is part of the model. The f(S1, S2,. …) function can range from very simple, such as the concentration of just one substrate, to complex sets of equations involving several substrates and rate parameters. Determination of appropriate rate expressions and parameter values for those expressions is an active research area. Combining Models. In many cases, evaluating bioremediation involves combining two or more of the model types. For example, in situ bioremediation of a chlorinated solvent may require a multiphase flow model coupled to a sophisticated biodegradation rate model. The multiphase flow model tracks the movements of the water and the solvent; once the flows are known, a transport model uses a biodegradation rate model as a sink term. Biodegradation rate models are most easily combined with flow models when one rate-limiting material can be identified. The rate-limiting material often is the primary electron donor or electron acceptor. For example, the biodegradation rate of petroleum hydrocarbons often can be modeled with dissolved oxygen as the rate-limiting substance. In several successful modeling studies, overall biodegradation rates could be modeled by the rate at which oxygen entered the bioremediation zone. The major simplification achieved by assuming rate limitation solely by oxygen should not be considered a general rule. It can be appropriate for biodegradation of petroleum hydrocarbons (a process that is especially sensitive to low oxygen concentrations) when the input rate for dissolved oxygen is low compared to the amount of hydrocarbon present and the site is large. Because these conditions are not true in many other situations, biodegradation rate modeling may require different and more sophisticated approaches. Except when the biodegradation or geochemical models are very simple, coupling them with flow models requires more than an extension of the existing contaminant transport models used for conservative tracers. Considerable attention must be given to proper model formulation and to efficient and accurate solution techniques. Otherwise, costs and computer time will be excessive. How to Use Models Models provide a framework for organizing information about
OCR for page 85
contaminated sites. They increase understanding of contaminant behavior by requiring the model user to confront details such as the mass of contaminants, their chemical properties, and their dynamic interactions with site hydrogeochemical characteristics. When this required information is available and integrated into the proper model, modeling predictions become useful tools for managing field sites and evaluating bioremediation. Models can be useful for evaluating in situ bioremediation in two ways. One approach is to see if a model representing only abiotic mechanisms can or cannot account for all of the contaminant loss. A second approach goes a step further and evaluates if "reasonable" estimates of microbial processes, quantified through the model, can explain contaminant losses (see Box 4-4). This second approach requires detailed knowledge of rate coefficients describing how quickly the microbes degrade the contaminant, in addition to parameters describing transport and other abiotic phenomena. Mass Losses. The first modeling approach requires analyzing whether abiotic mechanisms (for example, dilution, transport, and volatilization) can explain all of the losses of the contaminant mass. The approach recognizes that biodegradation rate models often have greater uncertainty that do models of abiotic processes. The uncertainty can be caused by poor understanding of the biochemical reactions, difficulty estimating parameters, and inadequate site characterization. Eliminating biological reactions from the model avoids this uncertainty. When the model of abiotic mass losses shows that some contaminant mass remains after all the abiotic sinks are considered, there are two possible explanations: (1) biodegradation processes are implicated as the sink for the "missing" mass, or (2) the modeling parameters were improperly selected, have led to inaccurate predictions, and are therefore misleading the modeler. Because judgments about microbiological involvement in contaminant loss may be contingent upon the selection of parameters used to describe abiotic losses, a modeler must be vigilant—constantly scrutinizing the validity of decisions and parameters that affect the modeling results. Adjustments in modeling parameters can lead to vastly different predictions; therefore, it is prudent to give credence to evidence for bioremediation only when the modelers have a high degree of confidence in their results and when discrepancies between actual and modeled contaminant behavior are unambiguous. Conclusions about effective bioremediation should only be drawn when concentrations of contaminants found in field sites are not simply lower but significantly lower than would be
OCR for page 86
BOX 4-4 PROVING INTRINSIC BIOREMEDIATION OF A SPILL AT A NATURAL GAS MANUFACTURING PLANT— NORTHERN MICHIGAN At a plant in northern Michigan, waste products from natural gas manufacturing leaked from a disposal pit into the surrounding ground water. Having installed wells around the plant to prevent off-site migration of contaminated water, the company in charge of the facility chose intrinsic bioremediation to cleanup the contaminants (primarily benzene, toluene, and xylene, or BTX). In demonstrating the effectiveness of bioremediation, the company provided evidence that meets the three criteria discussed in this report: Documented loss of contaminants: The company began its extensive site-monitoring program to follow the effectiveness of intrinsic bioremediation in 1987. Since that time the benzene concentration has dropped by approximately 90 percent and the contaminant plume has shrunk considerably. Laboratory assays showing that microorganisms have the potential to degrade the contaminants: The company performed a series of lab tests with soil cores retrieved from the field showing that the site's native microbes could degrade BTX at a high rate—5 to 10 percent per day—if supplied with adequate oxygen (1 to 2 ppm or more). Evidence showing that biodegradation potential is realized in the field: The company used a computer-based model, BIOPLUME II, to demonstrate that the rate of contaminant loss that one would predict if bioremediation were occurring closely matched the actual contaminant loss rate in the field. In 1987 the company measured the BTX and dissolved oxygen levels at various points in the plume. These values were input into BIOPLUME II to predict how they should change with time if bioremediation were occurring. The field measurements of both the contaminant concentrations and the dissolved oxygen levels taken since 1987 closely match the model's predictions. In addition, the biodegradation rate predicted by the model closely matches the rate measured in the field. Monitoring at this site is still ongoing to demonstrate the long-term effectiveness of intrinsic bioremediation. Reference 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. Groundwater 27(6):823-834.
OCR for page 87
expected from predictions based on abiotic processes. Thus, this approach works well when biodegradation is the dominant sink and when the abiotic processes are well characterized. Uncertainty in modeling the abiotic processes makes this approach unreliable when biodegradation is not the dominant removal mechanism. Direct Modeling. When reasonable estimates of biological processes and parameters are available, directly modeling the biodegradation process is the superior modeling strategy. These estimates can be obtained from the scientific literature, past experience with similar circumstances, laboratory experiments, or field-scale pilot studies, depending on the site conditions and biodegradation reactions. One approach is to use the model to answer the question, "Does our best representation of the biodegradation rates, when combined with the simultaneously occurring abiotic rates, support the conclusion that biological reactions are responsible for observed changes in contaminant levels or other relevant observations?" If the answer is "yes," modeling provides a much greater measure of confidence that observations supporting biodegradation are not artifacts. A second approach is to use direct modeling to predict the contaminant's concentration at unsampled locations or to predict the future concentration. The model then identifies sample locations and times that should yield particularly definitive measurements. Subsequent sampling, if consistent with model predictions, confirms the analyst's understanding of what is occurring in the subsurface. Lack of agreement between model predictions and actual developments forces a reevaluation of the model and improves understanding of the site and the parameter values. In some cases, direct modeling must involve highly sophisticated computer codes that take into account the three-dimensional nature of the site, heterogeneities, and complex reactions. These models are expensive to formulate and run, but they are essential tools for investigators who require a detailed description of what is happening at a site. Currently, these types of models are viewed primarily as research tools appropriate for highly monitored research, demonstration, or pilot sites. In many practical applications, direct modeling can be greatly simplified by eliminating all but the most essential phenomena. A good strategy is to compare expected rates of all phenomena that might affect the bioremediation. For example, the expected rate of contaminant loss due to biodegradation can be compared with the expected contaminant loss rate due to volatilization. Normally, a few of the possible phenomena will have expected rates much greater
OCR for page 88
than those of the other phenomena, and the model can consider only the phenomena having relatively high rates. If the biodegradation rate is high enough that it should remain in the model, the model provides prima facie evidence that bioremediation is working. Solution of the complete model can verify the evidence. Limitations of Models Although a powerful tool, modeling has its shortcomings. One shortcoming is that a model's validity must be established on a site-by-site basis, because no "off-the-shelf" models are available for evaluating bioremediation on a routine basis. Although a drawback in terms of time and cost, model validation may be a net advantage because it results in a more complete understanding of the site. Another limitation is that determining each of the many modeling parameters (such as hydraulic conductivity, retardation factors, and biodegradation rate parameters) may be as demanding and expensive as making the measurements for other types of verification criteria. Thus, a trade-off may exist between better modeling and more field measurements. Despite its limitations, modeling should be a routinely used tool for understanding the dynamic changes that occur in field sites during bioremediation. Although the complexity and type of model can vary, modeling is a valuable tool for linking conceptual understanding of the bioremediation process with field observations and for giving weight to a limited set of data. Even if site complexities preclude assembling a model that provides valid quantitative predictions, models are valuable management tools because they integrate many types of information relevant to the fate of contaminants. LIMITATIONS INHERENT IN EVALUATING IN SITU BIOREMEDIATION Because the subsurface is complex and incompletely accessible, knowledge of the fate of ground water contaminants always will be limited. This situation is intensified for in situ remediation technologies of any type, because frequently the amount, location, and type of contamination are unknown. Without knowing the starting point for a remediation, defining the finishing point is difficult. Errors in measurements, artifacts imposed by extrapolating lab results to the field, and an inherent shortage of data further complicate the evaluation and create uncertainty about the performance of a remediation process. For example, in analyzing chemical concentrations in ground water, a large number of samples from spatially different locations
OCR for page 89
may be gathered. Even assuming the laboratory results are completely error free, uncertainty arises from extrapolating these point samples in an attempt to portray a complete picture of how the water's chemical composition varies in space. Because evaluation of bioremediation requires integrating concepts and tools from very different disciplines, efforts to synthesize information from these different disciplines can create problems. For example, microbiologists and hydrogeologists use space and temporal scales that seldom match. The seconds and micrometers characteristic of microbial processes are very much smaller than the months and kilometers typical of hydrogeological descriptions of landscape processes. Thus, the hydrologic data describing large-scale water flow do not always meet a microbiologist's needs for understanding the small-scale mechanisms that control microbial activity. For instance, models efficient for the typical space scale of water movement (i.e., meters to kilometers) obliterate all of the details of microbial reactions, which occur in distances of micrometers to centimeters. A prime example of the problem of trying to synthesize different scales is illustrated by the problems encountered when trying to document major increases in biomass during in situ bioremediation. Microorganisms often are highly localized near their food sources. This localization makes it difficult to "find" the organisms when only a few samples can be taken. Microbial numbers, biodegradation rate estimates, or biodegradation potentials can vary tremendously, depending on whether the sample was from a location of high microbial activity or from a nearby location with low activity. Microbiological variability occurs on a small scale compared to the scale represented by field samples. Consequently, uncertainty in microbiological parameters always is a risk. Three strategies can help minimize uncertainty and should play important roles in evaluating bioremediation: (1) increasing the number of samples, (2) using models so that important variables are properly weighted and variables with little influence are eliminated, and (3) compensating for uncertainties by building safety factors into the design of engineering systems. Investigators can trade-off these three strategies. For example, if gathering a large number of samples or using sophisticated models is not possible, larger safety factors can cover the resulting uncertainty. At small field research sites designed to investigate bioremediation of contaminants not yet treated on a commercial scale, a large number of samples and complex models may be possible—and necessary—to draw detailed conclusions from the research results. On the other hand, at large commercial sites, a similarly high density of samples may be cost prohibitive, and it may
OCR for page 90
be more appropriate to rely on larger safety factors to account for the greater uncertainties. Uncertainties in evaluating bioremediation can be minimized but not eliminated. Investigators cannot fully understand the details of whether and how bioremediation is occurring at a site. The goal in evaluating in situ bioremediation is to assess whether the weight of evidence from tests such as those described above documents a convincing case for successful bioremediation.
Representative terms from entire chapter: