INTRODUCTION
Since the mid-1980s, the environmental industry has mounted a tremendous effort to mitigate groundwater contamination, reaching some $10 billion per year in the 1990's. Critical technical review of this very costly groundwater restoration effort by the National Research Council (1994) indicates few successes have been achieved to date, for several reasons including:

• Stringent drinking water standards (maximum contaminant levels or MCLs), which were originally intended for public water supplies, were selected as cleanup goals with the naïve expectation that existing technology could meet these goals
• Pump-and-treat was the primary technology applied to quickly cleaning up groundwater to the MCLs
• Pump-and-treat did not perform as projected due to:

1. Subsurface physical heterogeneity and complex contaminant migration pathways, which are extremely difficult if not impossible to evaluate adequately
2. Non-aqueous phase liquids (NAPLs) present in the saturated zone are not efficiently removed by pumping groundwater
3. Contaminant recovery generally controlled not by the rate of groundwater flow, but by the rate of diffusion from inaccessible regions of the saturated zone to which contaminants have migrated
4. Sorption of contaminants to subsurface materials has generally been unaccounted for, which has resulted in an underestimation of the total contaminant mass that must be removed and time needed for restoration
5. Difficulties involved with characterizing subsurface heterogeneity has resulted in increased uncertainty in engineered remedial actions.

The best available technology for restoration of chlorinated solvent impacted soil and groundwater seems to include a handful of remedial alternatives or combinations thereof: soil vapor extraction, bioventing, in situ bioremediation, surfactants, two-phase extraction, passive barriers, and engineered microorganisms. While application of these technologies help to reduce the contaminant mass, all of these technologies are limited in their ability to perform. With the exception of very few "simple sites," there is currently no existing technology that can be applied to restore chlorinated solvent impacted groundwater to drinking water standards. However, over the past 10 to 15 years it has been observed that in some cases, the chlorinated solvent plume has not continued to migrate, but in fact has reached equilibrium in a relatively short distance. A closer look at many of these sites indicates that remediation is occurring by natural attenuation; in other words, that natural biogeochemical processes have caused the contaminants to break down to non-hazardous components.

Natural attenuation of chlorinated organics in groundwater was the subject of the September 1996 symposium that the two authors of this article were fortunate enough to attend. The symposium took place over three days in Dallas, Texas, and was sponsored by the U.S. EPA, the Air Force Center for Environmental Excellence (AFCEE), and US Air Force Armstrong Laboratory. This article is our attempt to provide HYDROVISIONS readers with a brief overview of some of the information provided at the symposium.

WHAT IS NATURAL ATTENUATION ?
"The biodegradation, dispersion, dilution, sorption, volatilization, and/or chemical and biochemical stabilization of contaminants to effectively reduce contaminant toxicity, mobility, or volume to levels that are protective of human health and the ecosystem" (U.S. EPA Office of Research and Development and Office of Solid Waste and Emergency Response). Also referred to as intrinsic remediation, intrinsic bioremediation, or passive remediation, the current "trend" for consistency is to call it remediation by natural attenuation (RNA). In groundwater, natural attenuation typically occurs through destructive (for example biodegradation) or non-destructive mechanisms (e.g., sorption, dispersion, dilution, volatilization). Biodegradation of fuel hydrocarbons, especially the aromatics (BTEX), is mainly limited by electron acceptors and with few exceptions will proceed to complete destruction. However, biodegradation of chlorinated solvents (e.g., PCE, TCE) most commonly occurs via reductive dechlorination, a process that requires both electron acceptors (chlorinated aliphatic hydrocarbons) and an adequate supply of electron donors (natural organic carbon, fuel hydrocarbons, landfill leachate) in order to proceed to complete destruction.

BIODEGRADATION OF CHLORINATED SOLVENTS
Chlorinated solvents undergo biodegradation through three different pathways: (1) use as an electron acceptor (reductive dechlorination); (2) use as an electron donor (primary substrate); (3) co-metabolism where degradation of the chlorinated solvent provides no benefit to the microorganism but is simply fortuitous. In general, biodegradation of chlorinated solvents is an electron-donor-limited process.

The generalized process of biodegradation of chlorinated solvents begins in the saturated subsurface where native/anthropogenic carbon is utilized as an electron donor, and dissolved oxygen is utilized first for the prime electron acceptor. Once the dissolved oxygen is depleted, anaerobic microorganisms typically utilize additional available electron acceptors in the following order: nitrate, ferric iron hydroxide, sulfate, and carbon dioxide. In the absence of nitrate and dissolved oxygen, chlorinated solvents compete with other electron acceptors and donors, especially sulfate and carbon dioxide. By looking at the spatial distribution and concentrations of electron acceptors and donors, the mechanism(s) and rates of biodegradation can be assessed.

The most important process for the natural biodegradation of chlorinated solvents is reductive dechlorination. The chlorinated solvent is utilized as an electron acceptor, and a chlorine atom is removed and replaced with a hydrogen atom. Because chlorinated solvents are utilized as electron acceptors during reductive dechlorination, an appropriate carbon source is required for microbial growth to occur. Reductive dechlorination has been demonstrated under nitrate- and sulfate-reducing conditions, but the highest rates of biodegradation occur during methanogenic conditions.

A less consequential natural biodegradation process for chlorinated solvents is as a primary substrate or electron donor reaction. Under aerobic and in some cases anaerobic conditions, less-oxidized chlorinated solvents (1,2-DCA and vinyl chloride) can be utilized as a primary substrate in biologically mediated redox reactions. The facilitating microorganisms obtain energy and carbon from the degraded chlorinated solvent in this process.

Co-metabolism is another natural degradation process for chlorinated solvents in the saturated subsurface, albeit much less consequential. The process of co-metabolism of chlorinated solvents has been best documented in cases where fuel hydrocarbons including BTEX are present with the chlorinated solvents. The chlorinated solvents are indirectly transformed by bacteria as they use the BTEX or another substrate to meet their energy requirements.

CHLORINATED SOLVENT PLUME BEHAVIOR
The behavior of chlorinated solvent plumes is dependent upon several elements of the biogeochemistry of the aqueous subsurface environment: (a) mass and concentration of solvent present, (b) amount of biologically available organic carbon, (c) distribution and concentration of natural electron acceptors, (d) types of electron acceptors being used. The behavior of solvent plumes has been categorized into three types: (Type 1) primary substrate is anthropogenic carbon (e.g., fuel hydrocarbons, landfill leachate) and anthropogenic carbon drives the reductive dehalogenation; (Type 2) primary substrate is native organic carbon and native carbon drives the reductive dehalogenation; (Type 3) primary substrate of carbon essentially absent (no natural or anthropogenic source available), dissolved oxygen concentrations greater than 1 milligram per liter. Conditions conducive to RNA are represented by rapid and extensive reductive dehalogenation with Type 1 plumes, and slow to rapid reductive dehalogenation with Type 2 plumes. In Type 3 plumes, reductive dehalogenation is absent and natural attenuation occurs predominantly by advection, dispersion, and sorption; however, vinyl chloride can be rapidly oxidized.

Generally what occurs is a mixture of complex plume behaviors. For example, Type 1 behavior may be exhibited in the near-source area, with fuel hydrocarbon degradation accompanied by breakdown of PCE and TCE to DCE. Downgradient, with the disappearance of fuel hydrocarbons and primary substrate, behavior is represented by the presence of DCE and no other daughter products. A preferred plume behavior combination is to have a Type 1 in the source area with the majority of reductive dechlorination occurring there, and the degradation of the chlorinated solvents to vinyl chloride and ethene. Downgradient of the Type 1 zone, a Type 3 oxidizing zone would provide the mechanism to degrade the vinyl chloride to carbon dioxide and chloride.

CHARACTERIZING FOR RNA
The characterization of natural attenuation of chlorinated solvents sites requires very detailed and calculated monitoring. In the opinion of Dr. John Cherry of the University of Waterloo, what by appearance is suggested to be natural attenuation of chlorinated solvents may in fact be a result of inadequate site characterization or poor monitoring. Cherry's work characterizing chlorinated solvent plumes in the Borden aquifer, Canada, indicates that even in a relatively homogeneous sandy aquifer, the core of the plume may occur in a very discrete zone on the order of a foot or so thick. For this reason, identification of the plume core cannot necessarily be expected to be accomplished with conventional monitoring well networks and construction. Detailed investigations involving direct push technology at vertical intervals as small as one foot to assemble three-dimensional transects are necessary to elucidate contaminant distribution patterns. Cherry suggests that the emphasis of site characterization data collection should be on detailed spatial rather than temporal data, and that the most economic way to collect the finer grid spatial data is with direct push technology, collecting of soil cores and discrete groundwater samples during vertical advancement of the probe.

For RNA investigations, it is necessary to determine which reactions and processes are driving the subsurface biogeochemistry. General physical parameters include oxidation-reduction potential, temperature, pH, field-titrated alkalinity, and conductivity. Additional chemical parameters include the dissolved oxygen, nitrate, nitrite, ferric iron, ferrous iron, sulfate, sulfide, chloride, total organic carbon (TOC), fuel hydrocarbons, aromatic and chlorinated solvents. In the RNA context, TOC is a measure of the total electron donor capacity of the saturate subsurface. Some practitioners recommend using dissolved organic carbon, excluding carbonates with acid, to more accurately estimate total electron donor capacity. Dissolved molecular hydrogen, methane, ethane and ethene are also necessary components of a natural attenuation monitoring program. The technique for quantifying dissolved molecular hydrogen involves bubbling extracted groundwater in the field using a flow through gas stripping device and analyzing with a properly equipped field gas chromatograph. The range of hydrogen concentrations is indicative of the particular biodegradation process occurring (nitrate reduction, iron reduction, sulfate reduction, methanogenic).

An additional element to consider is the completion of microcosm studies to assess if the microorganisms necessary for biodegradation are present and to help quantify rates of biodegradation. Simplistically, microcosm studies amount to collecting subsurface soil samples, placing the soil samples under a controlled laboratory environment, injecting known contaminants into the samples, and measuring the changes in concentration over time, allowing estimation of biodegradation rates. Microcosm studies are time-consuming, expensive, and are generally recommended only when there is a considerable need to demonstrate the presence of the microorganisms. However, if properly designed, implemented and evaluated, microcosm studies can provide very convincing evidence on the presence and rates of biodegradation. It has been suggested that biodegradation rate constants determined by microcosm studies may be much higher than actual in situ rates. Therefore, the preferable method is to determine the contaminant biodegradation rate constant via in situ field measurement.

CASE STUDIES

Naval Air Station Cecil Field, Florida ­ Documentation of RNA at Naval Air Station Cecil Field in Florida is provided by the segregated sequence of redox processes into distinct zones, where reductive dehalogenation of chlorinated ethenes occurs near the contaminant source, followed by oxidative degradation of vinyl chloride to carbon dioxide and chloride downgradient. Near the contaminant source, methanogenesis predominates. Moving downgradient, distinct sulfate-reducing, iron (II)-reducing, and oxygen-reducing zones are evident.

St. Joseph, Michigan Superfund Site ­ Groundwater flows at the St. Joseph site from the contaminant source toward Lake Michigan. The average hydraulic conductivity was estimated at 7.5 meters per day, the estimated travel time for TCE in groundwater is 18 years, and if the contaminant was released only as an aqueous phase, it should have reached the lake by now. The observed contaminant concentration and distribution suggests a continuing DNAPL source. Evidence of RNA is provided by segregated methanogenic and oxygenated zones of the saturated subsurface. Anaerobic degradation of TCE occurs through a successive dechlorination from TCE to DCE to vinyl chloride and ethene. High cis-DCE concentrations were generally associated with declines in oxygen and sulfate concentrations and appeared on the upper edge of the methanogenic zone. TCE concentration decreased by a factor of 50,000, and all contaminants decreased to values below the MCL before they reached the lake. Mass fluxes decreased by factors of 10 to 123. Apparent degradation rate constants were estimated from a two-dimensional model at 0.3 to 1.7 for TCE, 0.54 to 4.0 for cis-DCE, and 2.6 to 20 for vinyl chloride.

Plattsburg Air Force Base, New York ­ Soil and groundwater were contaminated by fuel hydrocarbons and chlorinated solvents at a former fire training area at Plattsburg Air Force Base, New York. Groundwater contaminants include TCE, cis-1,2-DCE, vinyl chloride, and BTEX. Depth to groundwater ranges from 45 feet bgs to the west and ground surface to the east, and groundwater flows to the southeast at approximately 140 feet per year. Available biogeochemical data indicate two separate plume behavior categories at the site. In the source area and extending approximately 1,500 feet downgradient, the presence of commingled BTEX and TCE is characterized by anaerobic conditions that are strongly reducing (Type 1 behavior). Dissolved oxygen concentrations are on the order of 0.1 mg/L (background 10 mg/L), nitrate approximately 0.1 mg/L (background 10 mg/L), iron (II) 15 mg/L (background 0.05 mg/L), sulfate 0.05 mg/L (background 25 mg/L), methane at 3.5 mg/L, hydrogen from 1.4 to 11 nanomoles. The hydrogen concentrations are indicative of sulfate reduction and methanogenesis even though there is no sulfate available and relatively little methane produced, thus suggesting that reductive dechlorination may be competitively excluding these processes. In this area BTEX is being used as the primary substrate and TCE is being reductively dechlorinated to cis-1,2-DCE and vinyl chloride. Extending from approximately 1,500 to 2,000 feet downgradient of the source, the majority of the BTEX has been degraded, and the plume is characterized by conditions suggestive of Type 3 plume behavior. Dissolved oxygen concentrations are on the order of 0.5 mg/L (background 10 mg/L), nitrate increasing to background level of 10 mg/L at plume end, iron (II) decreased to 1 mg/L (background 0.05 mg/L), sulfate increasing to 15 mg/L (background 25 mg/L), and hydrogen from 0.8 to 0.25 nanomoles. The hydrogen concentrations are indicative of iron (III) reduction.

Picatinny Arsenal, New Jersey ­ A solvent plume at this site occupies an area approximately 500 m long by 250 m wide, in a shallow, 20 m thick unconfined sandy aquifer in which groundwater flows at an estimated velocity of 0.3 to 1 m/yr. Thirty years of degreasing activities in one location have released an estimated 3,500 kg of TCE into the aquifer. Maximum concentrations of TCE in groundwater occur near the base of the aquifer, midway along the centerline of the plume, at 10,000 µg/L. Characterization of parameters indicative of natural attenuation found dissolved oxygen at less than 0.5 mg/L, nitrate less than 1 mg/L, ferric iron > 1 mg/L, abundant sulfate and carbon dioxide (measured as alkalinity), noticeable hydrogen sulfide smell, and dissolved methane concentrations at 1 to 85 µg/L. These findings strongly support the conclusion that anaerobic biologic processes are predominant at different stages, and their spatial distribution suggests different redox processes are present at different locations. Dissolved organic carbon is contributed from natural sources as fulvic and humic acids. The presence of cis-1,2-DCE and vinyl chloride, solvents not used at the site, are further indicative of reductive dehalogenation. Investigators at this site identified the distribution of the solvent mass in different phases, and estimated the means by which solvent mass is actively being reduced. The ratio of sorbed to dissolved solvent is about 3:1 to 4:1. An estimated 50 kilogram per year (kg/yr) is discharged to a stream at the toe of the plume, 1 kg/yr spreads laterally to concentrations below the plume boundary definition (10 µg/L), 360 kg/yr is removed by natural anaerobic biotransformation, 50 kg/yr is lost to advective-driven volatilization, and less than 1 kg/yr each are attributed to diffusion driven volatilization and to sorption. While 460 kg/yr are lost, primarily due to natural anaerobic bioattenuation, 550 kg/yr are added due to desorption of solvent from the most concentrated core of the plume. In contrast, traditional pump and treat technologies active at this site remove about 70 kg/yr, at an estimated cost of $700,000/yr. Natural attenuation has been proposed as the principal mechanism for the removal of TCE from the Picatinny Arsenal plume.

WILL REGULATORY AGENCIES ACCEPT RNA?
There have been over 90 U.S. EPA Records of Decisions with RNA as either the selected alternative, or an element of the selected remedy for the site (through 1995). About one-third of these sites are landfills. According to U.S. EPA, RNA is not a "no action" alternative for RP's to walk away from a site. In fact, demonstrating the effectiveness of remedies involving natural attenuation will require more thorough site characterization, detailed monitoring of the remedial progress, and contingency measures to ensure long-term reliability and protection of human health and the ecosystem. There are currently two states with RNA policies: North Carolina and New Jersey. Approximately one-third of the states favorably support RNA and are in the process of developing state policies - California to our knowledge is not one of these states. U.S. EPA will be issuing a federal policy on RNA either the end of 1996 or the beginning of 1997. The American Society for Testing & Materials (ASTM) is in the process of developing an ASTM Standard on RNA.

In the future, groundwater restoration strategies have to be better, smarter, and cheaper. This means better application of the scientific data and knowledge collected over the last fifteen years to utilize combined smarter strategies resulting in cheaper remedies to groundwater pollution problems. Better application of data may mean improved site characterizations with sophisticated sampling techniques, new analytical approaches, and state-of-the-art groundwater modeling to more accurately predict the contaminant fate and transport. Smarter strategies may mean combining source removal, source control/containment, and RNA to acceptably manage the risk of exposure. Cheaper as compared to pump & treat for decades may include assessing the net-benefit of actually restoring the groundwater resource to drinking water supply MCLs. In this industry, one thing is certain: "the future just ain't what it used to be."

Tim Parker, CEG, CH, is a Project Manager for Law Engineering and Environmental Services in Sacramento. Tom Mohr is Senior Hydrogeologist, Yolo County, and is President of GRA's Sacramento Branch.