Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Groundwater Contamination and Aquifer Reclamation at the Rocky Mountain Arsenal, Colorado 6 INTRODUCTION LEONARD F. KONIKOW U.S. Geological Survey DOUGLAS W. THOMPSON U.S. Army Corps of Engineers AB STRACT Groundwater contamination at the Rocky Mountain Arsenal, Colorado, is related to the disposal of liquid industrial wastes and to industrial leaks and spills that have occurred during the 40-yr history of operation of the Arsenal. From 1943 to 1956 the liquid wastes were discharged into unlined ponds, which resulted in contamination of part of the underlying alluvial aquifer. Since 1956, disposal has been accomplished by discharge into an asphalt-lined reservoir, which significantly reduced the volume of contaminants entering the aquifer. In the mid-1970s toxic organic chemicals were detected outside of the Arsenal in the alluvial aquifer. The Colorado Department of Health issued three orders, which called for (1) a halt to unauthorized discharges, (2) cleanup, and (3) groundwater monitoring. Subsequently, a management commitment was made to mitigate the problem. A pilot groundwater containment and treatment system was constructed in 1978; it consists of (1) a bentonite barrier and several withdrawal wells to intercept contaminated groundwater along a 1500-ft length of the northern Arsenal boundary, (2) treating the water with an activated carbon process, and (3) injecting the treated water on the downgradient side of the barrier through several recharge wells. Because of the success of the pilot operation, it is being expanded at present to intercept most of the contaminated underflow crossing the entire north boundary. However, boundary interception alone cannot achieve aquifer restoration at the Arsenal. It is anticipated that the overall final program will also have to include elements of source containment and isolation, source elimination, process modification to reduce the volume of wastes generated, and development of alternative waste-disposal procedures that are nonpolluting. A variety of alternatives have been proposed and are currently being evaluated to determine the most feasible for implementation. The research, planning, and design studies that are necessary to achieve the reclamation goal at the Arsenal illustrate that an effective aquifer restoration program is difficult to design and expensive to implement. The contamination of a groundwater resource is a serious prob- lem that can have long-term economic and physical conse- quences because in most cases the problem is neither easily nor quickly remedied. Wood (1972) concluded, "The most sat- isfactory cure for groundwater pollution is prevention." In many 93 cases where a serious groundwater contamination problem ex- ists, the single most important remedial action that can be taken is to eliminate the source of contamination. But even then, contaminants already in the aquifer will continue to migrate and spread unless some action is taken to immobilize, neu- tralize, or remove them. Hence, there is often a need to clean up or restore contaminated aquifers.
94 The "restorability" of a contaminated aquifer is dependent on the hydrogeologic and geochemical properties of the affected aquifer and on the chemical and physical properties of the contaminant. Restoration of a contaminated aquifer is neither technically nor economically feasible in many cases. Factors frequently hindering restoration include (1) the slow diffusive nature of groundwater flow, (2) the difficulty of defining sec- ondary permeability effects, (3) the generally low oxygen con- tent and lack of biologic reactivity in groundwater, (4) the re- tention of some chemicals in the aquifer because they tend to be sorbed by minerals in the rocks making up the aquifer, (5) the lack of transferability of some restoration techniques from one site to another, and (6) the lack of knowledge about the source of the contamination. Effective aquifer restoration programs, if technically feasible, are both difficult to design and expensive to implement. Never- theless, in response to public or governmental demands for positive action in clearly documented cases where groundwater contamination threatens public health, aquifer cleanup pro- grams are being required and instituted more frequently. Some programs are being financed and operated by the federal gov- ernment. Examples include the Rocky Mountain Arsenal, Col- orado, where irrigation and domestic water-supply wells in adjacent areas have been contaminated from industrial wastes stored at the Arsenal, and also Wurtsmith Air Force Base, Michigan, where toxic organic solvents used in aircraft main- tenance have entered and spread through the underlying aqui- fer. Other programs may be implemented because of violations of federal regulations. For example, a recent justice Depart- ment suit was filed in North Carolina under the imminent hazard provision of the Resource Conservation and Recovery Art the Alit acts that the rl~f~ncl~nt~ " ~ an- 7A%~7 FUJI AM . . . permanently restore the aquifer to a condition commensurate with safe human use" (Hazardous Waste News 2~2), [an. 21, 1980, p. 12~. As an example of an aquifer restoration program being initiated be- cause of state regulations, a chemical company in northern Michigan has come to an agreement with the state of Michigan to remove the contaminants from the soil and groundwater at their former dump site; the projected cost is $12 million to $15 million (The Wall Street fournal, Sept. 25, 1981, p. 48~. General management options for restoring water quality in aquifers currently available include the following: (1) eliminate the source of contamination but allow restoration to proceed only through natural flushing, dilution, and geochemical or biological reactions; (2) accelerate removal of contaminants through withdrawal wells, drains, or trenches; (3) accelerate flushing with artificial recharge; (4) install "impermeable" bar- riers to contain a contaminated area; (5) induce in situ chemical or biologic reactions that would neutralize or immobilize the contaminant; and (6) excavate and remove the contaminated part of the aquifer. The selection of the best approach for a particular situation requires the ability to predict changes in flow and chemical concentration in the aquifer for each possible management alternative. This in turn requires both adequate field data to describe the aquifer systems and the development of accurate simulation models to define the groundwater flow system, pollutant-transport mechanisms, and nature and rate of chemical or biological reactions. LEONARD F. KONIKO\V and DOUGLAS W. THOMPSON N 1 _q o o1 / ol JO u,1 Cal ~1 ~1 r--- q L ~ ,/ ~ Brighton DENVE; COLORADO A R A P A H O E C O . 1! ~ - _ _. _ _. _ __ PLY O ~ 10 MILES O 5 10 KILOMETERS ____ . / 1 red 1 FIGURE 6.1 Location of study area (boundaries are approximate). l 1 This chapter focuses on the groundwater contamination problem at the Rocky Mountain Arsenal, which is located near Denver, Colorado (see Figure 6.1~. This area is well suited for serving as a case study to illustrate data requirements, inves- tigative approaches, and management options related to the reclamation of contaminated aquifers because (1) the 40-yr his- tory of groundwater contamination is relatively well docu- mented in the scientific and engineering literature; (2) the geology and hydrology of the area are fairly well known; (3) adequate, though limited, water-quality data are available to calibrate numerical simulation models; (4) the locations and strengths of contaminant sources can be approximately recon- structed; (5) a management commitment has been made to aquifer reclamation; and (6) construction, operation, and eval- uation of a pilot reclamation system at the Arsenal have been completed. DESCRIPTION OF STUDY AREA History of Contamination The Rocky Mountain Arsenal has been operating since 1942, primarily manufacturing and processing chemical warfare prod-
Contamination and Aquifer Reclamation ucts and pesticides. These operations have produced liquid wastes that contain complex organic and inorganic chemicals, including a characteristically high chloride concentration that apparently ranged up to about 5000 mg/L. The liquid wastes were discharged to several unlined ponds (Figure 6.2), resulting in the contamination of the underlying alluvial aquifer. On the basis of available records, it is assumed that contamination first occurred at the beginning of 1943. From 1943 to 1956 the primary disposal was into pond A. Alternate and overflow discharges were collected in ponds B. C, D,andE. Much of the area north of the Arsenal is irrigated, both with surface water diverted from one of the irrigation canals, which are also unlined, and with groundwater pumped from irrigation wells. Some damage to crops irrigated with shallow ground- water was observed in 1951, 1952, and 1953 (Walton, 19617. Severe crop damage was reported during 1954, a year when the annual precipitation was about one half the normal amount and groundwater use was heavier than normal (Petri, 1961~. Several investigations have been conducted since 1954 to determine both the cause of the problem and how to prevent further damage. Petri and Smith (1956) showed that an area of contaminated groundwater of several square miles existed north and northwest of the disposal ponds. These data clearly indicate 39°50t _ 1 ,/ D ~B~ ~ I I r 1 i l i L. .. | J ROCKY MOUNTAIN ARSENAL BOUNDARY ~ ~ _ . _, _ . _ . . _ _ . _ . _ . O 1 2 MILES 1 I ~ 1 1 1 O 1 2 KILOMETERS EXPLANATI ON ............. 1 ::::::::::::::::::1 Irrigated area do Unlined reservoir - · Irrigation well ~ Lined reservoir FIGURE 6.2 Major hydrologic features; letters indicate disposal-pond designations assigned by the U.S. Army (Konikow, 1977). 95 that the liquid wastes seeped out of the unlined disposal ponds, infiltrated the underlying alluvial aquifer, and migrated down- gradient toward the South Platte River. To prevent additional contaminants from entering the aquifer, a 100-acre (0.045-km2) evaporation pond (reservoir F) was constructed with an asphalt lining in 1956 to hold all subsequent liquid wastes. Although the liner eventually failed, even if the lining were to have remained totally impervious, this new disposal pond in itself would not eliminate the contamination problem because large amounts of contaminants were already present in and slowly migrating through the aquifer. From about 1968 or 1969 through about 1974, pond C was maintained full most of the time by diverting water from the freshwater reservoirs to the south. This resulted in the inf~l- tration of about 1 ft3/sec (0.03 m3/sec) of freshwater into the alluvial aquifer. This artificial recharge had the effect of diluting and flushing the contaminated groundwater away from pond C faster than would have occurred otherwise. By 1972 the areal extent and magnitude of contamination, as indicated by chlo- ride concentration, had significantly diminished. Chloride con- centrations were then above 1000 mg/L in only two relatively small parts of the contaminated area and were almost at normal background levels in the middle of the affected area (imme- diately downgradient from pond C). In 1973 and 1974 there were new claims of crop and livestock damages allegedly caused by groundwater that was contami- nated at the Arsenal. Data collected by the Colorado Depart- ment of Health (Shukle, 1975) show that diisopropylmethyl- phosphonate (DIMP), a nerve-gas by-product, has been detected at a concentration of 0.57 part per billion (ppb) in a well located approximately 8 mi (12.9 km) downgradient from the disposal ponds and 1 mi (1.6 km) upgradient from two municipal water- supply wells of the city of Brighton. A DIMP concentration of 48 parts per million (ppm), which is nearly 100,000 times higher, was measured in a groundwater sample collected near the dis- posal ponds. Other contaminants detected in wells or springs ~ in the area include dicyclopentadiene (DCPD), endrin, aldrin, i dieldrin, and several organo-sulfur compounds. The detection of these chemicals, which were manufactured or used at the Arsenal, in areas off the Arsenal property led the Colorado Department of Health to issue cease and desist, cleanup, and monitoring orders in April 1975 to the Rocky Mountain Arsenal and Shell Chemical Company, which was leasing industrial facilities on the site. The Cease and Desist i Order called for a halt to unauthorized discharges of contam- i inants into surface water and groundwater just north of the Arsenal. The cleanup order applied to all sources of DIMP and DCPD located at the facilities. The third order called for a groundwater monitoring program, the results of which would be reported to the State Health Department on a regular basis. Consequently, a program that included groundwater monitor- ing and studies to determine a means to intercept contaminants flowing across the north boundary of the Arsenal was estab- lished by the U.S. Army. As a result of continued monitoring, additional contaminants have been identified in the groundwater at the Arsenal. The most widespread of those found are Nemagon (dibromochlo- ropropane) and various industrial solvents. Nemagon contam-
96 ination has been identified as probably resulting from Arsenal- related activities, whereas the industrial solvents identified are not unique to Arsenal activities. Extremely low concentrations of Nemagon (< 2 ppb) have been found in wells located im- mediately west of the Arsenal boundary. Other organic con- taminants associated with pesticide manufacturing have been found in wells located in a centrally located manufacturing plant area known as the South Plants area. These contaminants prob- ably entered the aquifer from accidental spills and leaks and appear to be migrating from this area very slowly. Hydrogeology The records of several hundred observation wells, test holes, irrigation wells, and domestic wells were compiled and anal- yzed to describe the hydrogeologic characteristics of the alluvial aquifer in and adjacent to the Rocky Mountain Arsenal. Kon- ikow (1975) presented four maps that show the configuration of bedrock surface, generalized water-table configuration, sat- urated thickness of alluvium, and transmissivity of the aquifer. These maps show that the alluvium forms a complex, sloping, discontinuous, and heterogeneous aquifer system. A map showing the general water-table configuration for 1955-1971 is presented in Figure 6.3. The assumptions and limitations of Figure 6.3 are discussed in more detail by Kon- ikow (1975~. The areas in which the alluvium either is absent or is unsaturated most of the time form internal barriers that significantly affect groundwater flow patterns within the aquifer and, hence, significantly influence solute transport. The general direction of groundwater movement is from regions of higher water-table altitudes to those of lower water- table altitudes and is approximately perpendicular to the water-table contours. Deviations from the general flow pattern inferred from water-table contours may occur in some areas because of local variations in aquifer properties, recharge, or discharge. The nonorthogonality at places between water-table contours and aquifer boundaries indicates that the approximate limit of the saturated alluvium does not consistently represent a no-flow boundary but that, at some places, there may be significant flow across this line. Such a condition can readily occur in areas where the bedrock possesses significant porosity and hydraulic conductivity or where recharge from irrigation, unlined canals, or other sources is concentrated. Because the hydraulic conductivity of the bedrock underlying the alluvium is generally much lower than that of the alluvium, groundwater flow and contaminant transport through the bedrock are as- sumed to be secondary considerations compared with flow and transport in the alluvial aquifer. Groundwater withdrawals in the area are predominantly from wells tapping the alluvial r aquifer. Contamination Pattern Since 1955 several hundred observation wells and test holes have been constructed to monitor changes in water quality and water levels in the alluvial aquifer. The areal extent of contam- ination has been mapped on the basis of concentration of chlo- ride, DIMP, and other inorganic and organic compounds in LEONARD F. KONIKOW and DOUGLAS W. THOMPSON .~" ~~ tROCKY MOUNTAIN ARSENAL BOUNDARY of O 1 2 MILES 1 ~ ~ O 1 2 KILOMETERS EXPLANATION 5/00 - WATER-TABLE CONTOURShows approximate altitude of water table, 1955 - 71. Contour interval 10 feet (3 meters). Datum is mean sea level Area in which alluvium is absent or unsaturated FIGURE 6.3 General water-table configuration in the alluvial aquifer in and adjacent to the Rocky Mountain Arsenal, 1955-1971 (Konikow, 1977). wells. Chloride concentrations ranged from normal background concentrations of about 40 to 150 mg/L to about 5000 mg/L in contaminated groundwater near pond A. Chloride data col- lected during 1955-1956 indicate that one main plume of con- taminated water extended beyond the northwestern boundary of the Arsenal-and that a small secondary plume extended beyond the northern boundary (see Figure 6.4~. The contam- ination pattern shown in Figure 6.4 clearly indicates that the migration of contaminants in this aquifer is also significantly constrained by the aquifer boundaries. Because chloride generally behaves as a conservative (that is, nonreactive) solute in groundwater, it is often assumed that chlorides can be used to indicate the maximum extent of con- tamination from a source that contains chloride. But this as- sumption is not always reasonable because chloride is also a common natural constituent in groundwater. At the Rocky Mountain Arsenal the extent of contamination as indicated by chloride concentration reflects a dilution ratio of about 33:1 from the contaminant source to the definable downgradient limit of contamination. However, the extent of contamination as indicated by some of the organic compounds, such as DIMP (see Robson, 1981), is much greater because they have a zero background concentration and can be detected to trace con-
Contamination and Aquifer Reclamation 39°50' 1 r- -- 'L O 1 2 MILES 1 1 , 1 1 1 O 1 2 KILOMETERS EXPLANATION Data point (Sept. 1955-March 1956) 500 Line of equal chloride concentration (in milligrams per liter). Interval variable 4~ Area in which alluvium is absent or unsaturated FIGURE 6.4 Observed chloride concentration in 1956 (Konikow, 1977). centrations that reflect a dilution ratio of about 100,000:1. Other organic contaminants exhibit a much smaller plume, or migra- tion distance, than does the chloride because of reactions that cause them to decay or to be adsorbed. Other differences among shapes and locations of plumes of different contaminants arise because they entered the aquifer at significantly different times and (or) locations within the Arsenal. For example, the Nem- agon plume occurs west of the chloride plume because the source of the Nemagon was not from the disposal ponds but apparently from a spill that occurred west of the ponds. Contaminants have also been detected in several shallow bedrock wells in or near the Arsenal. However, at present there are inadequate data to define the areal extent, depth of pen- etration, or rate of spreading of contaminants in the bedrock. APPLICATION OF SIMULATION MODELS The reliable assessment of hazards or risks arising from ground- water contamination problems and the design of efficient and effective techniques to mitigate them require the capability to predict the behavior of chemical contaminants in flowing groundwater. Reliable and quantitative predictions of contam- inant movement can only be made if the processes controlling 97 convective transport; hydrodynamic dispersion; and chemical, physical, and biological reactions that affect solute concentra- tions in the ground are understood. These processes, in turn, must be expressed in precise mathematical equations having defined parameters. The theory and development of the equa- tions describing groundwater flow and solute transport have been well documented in the literature. Perhaps the most important technical advancement in the analysis of ground- water contamination problems during the past 10 yr has been the development of deterministic numerical simulation models that efficiently solve the governing flow and transport equations for the properties and boundaries of a specific field situation. Although many of the processes that affect waste movement are individually well understood, their complex interactions in a heterogeneous environment may not be understood well enough for the net outcome to be reliably predicted. Thus, the analysis of groundwater contamination problems can be greatly aided by the application of deterministic numerical simulation models that solve the equations describing groundwater flow and solute transport. The solute-transport model described by Konikow and Bre- dehoeft (1978) was used to simulate the movement of chloride through the alluvial aquifer at the Arsenal in an effort to re- produce the 30-yr (1943-1972) history of contamination, to help test hypotheses concerning governing processes and parame- ters to develop an improved conceptual model of the problem, to aid in setting priorities for the collection of additional data, and to evaluate possible management alternatives (Konikow, 1977~. The model included an area of approximately 34 mi2 (88 Imp. The stringent data requirements for applying the solute- transport model pointed out deficiencies in the data base avail- able at the start of the study. Specifically, it was found that the velocity distribution determined from the water-table config- uration mapped in 1956 (see Petri and Smith, 1956) was in part inconsistent with the observed pattern of contaminant spread- ing. The subsequent quantitative analysis and reinterpretation of available hydrogeologic data, based partly on feedback from the numerical simulation model, led to a revised conceptual model of the aquifer properties and boundaries that incorpo- rated the strong influence of the internal barriers within the alluvial aquifer. The solute-transport model of Konikow (1977) was calibrated mainly on the basis of the chloride concentration pattern that was observed in 1956 (Figure 6.4~. Computed chloride patterns agreed closely with observed patterns, which during the 30-yr history were available only for 1956, 1961, 1969, and 1972. The calibrated model was then used to analyze the effects of future and past changes in stresses and boundary conditions. For example, comparative analyses illustrated that it would probably take at least many decades for this contaminated aqui- fer to recover naturally its original water-quality characteristics. But it was also inferred that appropriate water-management policies for aquifer reclamation can help to reduce this resto- ration time to the order of years, rather than decades, for the relatively mobile contaminants. Konikow (1974;) also noted that the simulation results showed that a reclamation scheme using a network of interceptor wells would aid in containing and removing the contaminated groundwater.
98 Robson (1981) developed and calibrated a solute-transport model for DIMP to help evaluate (1) the mechanisms and pa- rameters controlling DIMP migration, (2) future DIMP con- centrations in nearby municipal water supply wells, and (3) the effectiveness of various groundwater barrier configurations de- signed to halt off-Arsenal movement of contaminated ground- water. The model included an area of about 90 mi- (230 km-) and assumed that DIMP is conservative. Using the calibrated model, Robson was able to reconstruct the historical movement of DIMP in the aquifer between 1952 and 1975, to estimate DIMP concentrations in the South Platte River resulting from discharge of contaminated groundwater, and to predict future DIMP concentrations under a variety of assumed management alternatives. To evaluate more fully the range of engineering approaches or alternatives that would be feasible for construction along the north boundary of the Arsenal, Warner (1979) modeled a smaller part of the aquifer (2.5 mi2 or 6.4 ems) in that area in much finer detail. He predicted the impact on DIMP concentration of implementing a variety of interception schemes that incor- porated variants of a basic plan that included elements of groundwater withdrawal, a barrier, and reinfection of treated water. Among other findings, Warner (1979) showed that a properly operated hydraulic barrier, consisting of a line of pumping wells, would be just as effective as a bentonite barrier in stopping the movement of DIMP-contaminated groundwater across the northern boundary of the Arsenal. It is recognized that other organic contaminants of concern may be sorbed or altered by chemical and biological reactions as they move through the aquifer. The movement of a solute that is sorbed will be retarded relative to the movement of a conservative solute. This is beneficial in the sense that in a given time a contaminant that is sorbed will not migrate as far as a conservative contaminant. However, the sorption process could pose a significant obstacle to aquifer reclamation because even after the contaminant source has been eliminated, the sorbed organics could later desorb and continue to migrate through the aquifer, perhaps still posing a hazard after all con- servative contaminants have been flushed out of the aquifer. Sorption processes can and have been incorporated into solute- transport models (see Grove, 1976), and this allows a more realistic evaluation to be made of their behavior and response to imposed aquifer reclamation stresses. Although this then presents no great conceptual difficulty, in practice it is quite difficult to determine the coefficients that describe the rates of reactions and exchange capacity of the aquifer material for each individual contaminant. An overall systems-management model is currently in final development under the sponsorship of the U.S. Army. This computer model is expected to provide a valuable management and decision-making tool to aid in evaluating aquifer recla- mation alternatives at the Rocky Mountain Arsenal. The model will be composed of numerous modules, including (1) ground- water flow, (2) solute transport, (3) groundwater interception and control, (4) surface-water control, (5) groundwater and sur- face-water treatment, (6) cost estimation, and (7) report and graphics output. The model will be evaluated and verified using the Rocky Mountain Arsenal as a test case because of the abun- LEONARD F. KONIKOW and DOUGLAS W. THOMPSON dance of historical data there. After verification, selected al- ternatives for contamination control and elimination at the Ar- senal will be modeled with a goal of predicting long-term system responses and costs. If successful, this model will be applied to Installation Restoration programs under way at other loca- tions. AQUIFER RESTORATION PROGRAM Response to Cease and Desist Orders As a result of the Cease and Desist Orders, an Installation Restoration program was established at the Rocky Mountain Arsenal under the direction of the Program Manager for Chem- ical Demilitarization and Installation Restoration, Aberdeen Proving Ground, Maryland. This office was later reorganized into the U.S. Army Toxic and Hazardous Materials Agency (USATHAMA), which currently directs the Installation Res- "oration program at the Arsenal. The main objective of this program is to limit the migration of contaminants from the Arsenal to the degree required by applicable federal and state regulations. The program is primarily concerned with contam- ination problems resulting from historical activities on the Ar- senal as opposed to ongoing operations. The Installation Restoration program consists of three major parts or subprograms that include regional groundwater mon- itoring, contaminant migration control, and elimination of con- taminant sources. This program had been organized to allow a phased approach in developing and implementing contaminant control systems, thereby accelerating the reduction of potential environmental hazards. More than $25 million has been ex- pended to date in the Installation Restoration program, ex- cluding the costs associated with construction of the control systems. A comprehensive groundwater monitoring program was de- veloped based on historical contaminant distribution infor- mation and initiated late in 1975. It included sample collection from both on-site and adjacent off-site wells. This monitoring program has been continually updated since that time to in- clude additional wells and analytical parameters as required. Currently, it involves the collection and analysis of samples from 90 to 100 wells on a quarterly basis. The information generated from the monitoring program is used to define the distribution and track the migration of known contaminants, identify new contaminants, develop design criteria for contam- ination control and treatment systems, and evaluate the op- eration of existing systems. The subprogram concerning contaminant migration control at the Arsenal boundaries was initiated in late 1975 with the goal of rapidly eliminating the migration of contaminants off the Arsenal's grounds. Boundary control was the only viable option because of the already wide distribution of contami- nants, the long travel times associated with contaminant mi- gration from the sources to the boundaries, and the lack of precise definition of all source areas. Pilot and full-scale bound- ary control systems have been implemented at the northern Arsenal boundary, and plans have been developed to expand
Contamination and Aquifer Reclamation the treatment system along the northwestern Arsenal bound- ary. These systems will be discussed in more detail later in this chapter. Planning for the control and elimination of contaminant sources evolved several years later as additional data became available on specific source areas. The goal is to control or eliminate the contaminant sources on the Arsenal grounds and thereby elim- inate the need for boundary control in the future. Studies have been undertaken to aid further identification and definition of contaminant sources, to develop feasible source control and elimination alternatives, and to develop control and treatment systems. A summary of the strategy and progress of this sub- program is given at the end of this chapter. Contaminant Migration Control at Arsenal Boundaries Because the contamination that resulted in the issuance of the Cease and Desist Order was detected in surface water and groundwater immediately north of the Arsenal, the primary focus of the Installation Restoration program during 1976 and 1977 was the northern Arsenal boundary. A dike was con- structed to stop the migration off the Arsenal of contaminated surface water. Studies were initiated to determine a feasible alternative for stopping the flow of contaminated groundwater offthe Arsenal without significantly altering the normal ground- water flow pattern in the area. The concept selected involved interception of the groundwater a short distance south of the northern Arsenal boundary, treatment of the water to remove the contaminants, and reinfection of the treated water at the boundary. Two methods were proposed for intercepting the flow of groundwater. The first method involved the use of a hydraulic barrier, one or two lines of closely spaced pumping wells that would provide for dewatering of the aquifer along or between the lines. The permeability in the area is sufficiently high for this concept to have worked, but the gradient is shallow and concern was expressed over the potential for excessive recy- cling of water from the reinfection wells back to the withdrawal wells. As a result of this concern and to provide an additional safety factor, a second method was selected that involved the use of a slurry cutoff wall to form an impermeable barrier between the withdrawal and reinfection wells. Treatment Process Late in 1975 a laboratory study was initiated to evaluate various methods for removing organic compounds from representative groundwater samples from the area. Treatment processes in- vestigated include granular activated-carbon adsorption, pow- dered activated-carbon adsorption, chemical oxidation using ultraviolet (UV) light and ozone, and anionic exchange resins. Key chemical parameters for analysis included DIMP and DCPD. Extensive laboratory studies were conducted using standard isotherm tests for evaluating the carbons and resins and using batch reactor tests for evaluating the UV/ozone process. The anionic exchange resins were dropped from further consider- 99 ation because of low efficiency and high cost. A series of field studies was initiated on the carbon adsorption and UV/ozone oxidation processes to permit further evaluation. Powdered activated-carbon adsorption tests incorporating a polymeric coagulant were conducted using a standard Army Erdlator water-treatment unit (chemical addition, mixing, up- flow clarification)(Sweder, 1977~. Granular activated-carbon ad- sorption tests were conducted using a dynamic-flow, multi- column system (Sweder, 1977~. UV/ozone oxidation tests were conducted using a continuous-flow, mechanically mixed reactor (Buhts et al., 1978~. Granular activated-carbon was found to be more efficient (110 mg of carbon/L of water) in removing the contaminants than was the powdered activated-carbon (200 mg of carbon/L of water>. Cost estimates were developed for the carbon adsorption and UV/ozone oxidation processes based on treating 10,000 gallons of water per hour (37,850 L/h). The estimated cost of granular activated-carbon treatment was ap- proximately $2 per 1000 gallons, powdered activated-carbon treatment was approximately $4 per 1000 gallons, and UV/ ozone oxidation treatment was approximately $3 per 1000 gal- ions. As a result of these studies and the immediate availability of proven process equipment, granular activated-carbon was selected for use in the proposed treatment system. Installation and Operation of Pilot Containment System Hydrologic and chemical data indicated that the highest con- centrations of contaminants were crossing the northern Arsenal boundary in the alluvial aquifer in an area associated with a buried channel in the relatively impermeable bedrock of the Denver Formation. This area is located approximately 1 mile east of the northwest boundary and has a width of approxi- mately 1000 ft (305 m). Because little operational information was available on groundwater contamination control systems similar to the one proposed, the Army decided to install a limited pilot containment system in the area of high-contam- inant concentrations and evaluate the possibility of extending the treatment system across the entire affected part of the northern boundary. The North Boundary Pilot System (NBPS) was constructed and placed in operation in July 1978. It included the following five subsystems: a barrier, dewatering wells, reinfection wells, treatment plant, and monitoring wells. A schematic diagram of the system is provided in Figure 6.5. The barrier was constructed by filling a 3-ft-wide, 1500-ft- long trench, averaging 25 ft in depth, with a mixture of soil and bentonite clay. The barrier was anchored approximately 2 ft into the bedrock all along the alignment. The dewatering wells were installed south (upgradient) of the barrier approximately 225 ft apart on a straight line parallel to the barrier. There were six 8-inch-diameter wells placed within 30-inch-diameter gravel-packed holes. Each well was screened throughout the entire saturated portion of the alluvial aquifer. A submersible pump and flow control system were installed at each well site. Water from the wells was pumped through an underground manifold to a single sump at the treat- ment plant.
100 The injection wells were installed north (downgradient) of the barrier, approximately 100 ft apart on a straight line be- tween the barrier and the northern Arsenal boundary. There were twelve 18-inch-diameter wells, which were installed in 36-inch-diameter gravel-packed holes. The recharge wells were screened to a point above the water table. Treated water was continuously injected into the recharge wells by gravity flow through an underground manifold system. Sensors and flow control valves were installed in the wells to prevent overflow or surface discharge in the event that a well experienced an excessively high buildup of hydraulic head because of clogging of well screens or other factors. The treatment plant subsystem was designed to treat 10,000 gallons of water per hour. It consisted of two mixed-media pressure filters, each 4 ft in diameter, and two adsorber vessels (or columns), each 10 ft in diameter and 11 ft high, designed to contain about 20,000 lb (9100 kg) of granular activated car- bon. Water from the collection sump was pumped through the filters in parallel to remove suspended material, then through the carbon absorbers, and finally to the injection wells. Only one carbon adsorber was in operation at any one time. When the DIMP concentration approached 50 ppb, the carbon was replaced. During 1978-1981, replacement was required ap- proximately once every 9 months. The exhausted carbon was transported offsite for regeneration by a commercial vendor. Carbon usage rates ranged from 100 to 150 mg of carbon/L of water. The treatment system was designed to be largely au- tomatic and simple to operate by incorporating automatic back- washing of the filters and sensors for control of pumps and valves. Ten monitoring wells were installed both upgradient and downgradient of the pilot containment system. They were cased with small-diameter PVC pipe and screened in the alluvial PILOT TREATMENT PLANT INFLUENT FROM DEWATERING WELLS . ... _ _ ..... _ REINJECTION - DEWATERING WELL WELL GROUND SURFACE ~ ,,,~,,~, t~ ~ ,,,~-,;-- _ /~ FIGURE 6.5 Schematic diagram of north boundary contamination control pilot system. LEONARD F. KONIKOW and DOUGLAS W. THOMPSON ~ ~ ~ ~ ~ ~ ~ ~ I ~ ~ ~ , An- T T 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 RETENTION TIME FIGURE 6.6 Typical gas chromatography/mass spectrometry scan of north boundary pilot treatment system influent. aquifer. Water levels and chemical quality were monitored periodically to provide information on the effectiveness of the operation of the system. The cost of the barrier and the wells as constructed in 1978 was $450,000. The facility for housing the treatment system cost approximately $40,000. The treatment equipment was ob- tained under a lease/service contract agreement with a com- mercial vendor with an initial cost of approximately $100,000 and a yearly fee ranging from $135,000 to $150,000. The NBPS operated successfully for a period of approxi- mately 3 yr. For example, during fiscal year 1979, downtime was less than 1 percent of operating time. The granular acti- vated carbon effectively removed the organic contaminants from the groundwater, generally to a level of less than 10 ppb, as illustrated by a comparison of typical gas chromatography/mass spectrometry analyses of the influent (Figure 6.6) and effluent (Figure 6. 7) of the treatment system. The flow of groundwater downgradient from the NBPS was essentially unchanged (D'Appolonia Consulting Engineers, Inc., 1979~. Preliminary data indicate that the concentration of organics in the ground- water downgradient from the pilot system has diminished sig- nificantly. Expanded Containment System As a result of the successful operation of the pilot containment system, construction of the expanded containment system was begun in early 1981. The expanded system consists of a 6800- ft barrier ranging from 25 to 50 ft deep, 54 withdrawal wells, and 38 reinjection (or recharge) wells. The expanded barrier effectively intercepts all the contaminated groundwater flowing across the northern Arsenal boundary in the alluvial aquifer. The expanded treatment system is designed to treat 36,000 gallons (136,000 L) of water per hour. The adsorbers used in the pilot operation have been replaced with three pulsed-bed adsorbers designed to contain 30,000 lb (13,600 kg) of carbon
Contamination and Aquifer Reclamation LU In g cat to a UJ Tl - a) ~ ~a _ ~o A n Q ~ .. 1 ~ _ ma_ I r r ~T~TT~ 1 2 3 4 s 6 7 8 9 10 11 12 13 14 15 16 RETENTION TIME FIGURE 6.7 Typical gas chromatography/mass spectrometry scan of north boundary pilot treatment system effluent. each. The new adsorbers should be much more efficient than the old ones because the anticipated carbon usage rate is only 25 to 30 mg of carbon/L of water. The mixed-media filters have been replaced with cartridge filters, which are easier to main- tain. The whole system is highly automated and will require only intermittent monitoring by a single operator. The esti- mated cost for the expanded system is approximately $6 million. The expanded system became operational in 1983. Other Contaminant Migration Control Systems Concepts have been developed for two additional boundary contaminant migration control systems located along the north- western Arsenal boundary (Figure 6.8~. One system will be located at the southern end of that boundary and the other midway along that boundary. Both systems have been devel- oped primarily to control the migration of low concentrations of Nemagon across the boundary. Both systems will be similar in size to the NBPS and will incorporate granular activated- carbon treatment of the groundwater. The system to be located on the southern end of the boundary (Irondale System) was constructed under the direction of Shell Chemical Company and incorporates a hydraulic barrier for interception of the groundwater, along with the injection wells. It became oper- ational in 1983. The other system, to be constructed by the Army, will incorporate a slurry cutoff wall, withdrawal wells, and reinfection wells, similar to the pilot system. It is scheduled to be operational in 1985. Planning for Control and Elimination of Contaminant Sources Contaminant migration control at the boundaries of the Rocky Mountain Arsenal was initiated to stop or severely limit the migration of contaminants off the Arsenal grounds as soon as possible. Owing to the size of the Arsenal and the extent of 101 the source areas, the boundary control systems could be re- quired to operate for an indefinite period of time. The only way to limit this requirement and the associated cost is to control or eliminate the contaminant sources. Therefore a study was initiated in 1980 to identify and assess existing and inno- vative control or elimination alternatives that are capable of bringing the Arsenal into compliance with all applicable federal and state environmental laws and regulations. Another study objective was to develop preliminary cost data and technical data for use in a subsequent detailed evaluation and comparison of alternatives. A study team made up of 12 government and independent scientists and engineers was established to con- duct and manage the study. A review of historical operations, past study reports, and data from ongoing studies was made to identify, where possible, potential sources of contaminant mi- gration problems. The next phase of the study involved the development of control strategies. Guidelines and criteria for development of the strategies were required because of the complexity of and relationships between the contaminant sources and migration characteristics. In addition, some degree of commonality of structure or organization among the strategies was needed to enable a comparison and ranking of the alternatives to be de- veloped. As a result, a hierarchical approach and structure for generation and classification of control strategies were devel- oped incorporating five levels of detail ranging from concept to unit operation (Rocky Mountain Arsenal Contamination Con- trol Study Team, unpublished report, August 1981~. Each team member individually developed a number of strategies using the hierarchical approach and determined the problem defi- nition and technical data-base deficiencies associated with each scheme. The schemes were then submitted to the group as a whole for integration and evaluation. Screening criteria were developed to aid in evaluating and comparing the alternative schemes. The goal was to produce EXPLANATION · · ·Dewatering wells / ~ _ ~ E] Liquid Treatment / tt NORTH BOUNDARY ~ '\ A Recharge wells ~~ SYSTEM 7 Slurry trench NORTHWEST BOUNDARY '/ /9~,/ SYSTEM (Proposed) ,~1 / ~ Basins ,,* ~ Basins B ~*7 v//////////////////////////////////, SOUTH PLANTS AREAWAY Lakes I ~iRONDALE SYSTEM (Shell Chemical Co.) l l ~1\ COMPLEX ~ r- l I 1 0 ~ 2 KILOMETERS 0 1 2 MILES l 1 l 1 am,. 1 ! _ ~ FIGURE 6.8 Location of existing and proposed boundary control systems.
102 a set of criteria that could be applied at the various hierarchical levels, thereby enabling a general screening of the schemes rather than a detailed evaluation of each one. The major criteria selected for use are as follows: 1. Availability of technology, 2. Amount of additional data required, 3. Cost and time needed to fill data gaps, 4. Life cycle costs capital and O&M, 5. Compatibility between systems, 6. Degree of risk environmental and technological, Compliance with regulatory requirements. The individual schemes developed by the study team mem- bers were integrated, evaluated, and screened by the study group as a whole. This work resulted in the presentation of 14 alternative schemes that were recommended for detailed eval- uation by the Contamination Control Study Team. The schemes incorporate various aspects of the technologies listed in Table 6.1. The schemes address only the known contaminant sources at the Arsenal and therefore may have to be expanded if ad- ditional sources are identified in the future. In addition to the development of the alternative schemes, the study team identified a number of data gaps concerning both problem definition and technology development that must be filled before final selection of a control or elimination al- ternative can be made. Studies have been included in the overall Installation Restoration program to fill these data gaps. They include additional hydrogeologic definition of certain areas on the Arsenal, surface-water hydrology definition, technology development for water treatment, and technology development for disposal of contaminated soil and residue. As the data from these additional studies become available, the study team will further evaluate and revise the alternatives as required with the goal of selecting one alternative for implementation. The implementation of the selected alternative will be con- ducted using a phased approach. As soon as a particular part TABLE 6.1 Contaminant Source Control and Elimination Technologies Groundwater Interception Hydraulic barrier Slurry trench Dewatering trench (French drain) Water Treatment Adsorption (carbon and resin) Chemical addition/coagulation/precipitation Filtration Membrane separation Chemical oxidation Activated sludge Volatile stripping Ion exchange Contaminated Soil and Residue Treatment Incineration Fixation/stabilization In situ forced leaching Excavation and disposal LEONARD F. KONIKOW and DOUGLAS W. THOMPSON of the alternative is defined and design criteria are developed, construction will be initiated. For example, the elimination of Basin F will probably be one of the first major actions initiated because it is known to leak and because the extent and nature of the contamination associated with this area of the Arsenal have been better defined than elsewhere. The control and elimination of known contaminant sources at the Rocky Moun- tain Arsenal are currently expected to involve a 5-yr construc- tion program that is scheduled to start in 1985. A final cost estimate for the construction program has not been developed, but preliminary estimates range from $50 million to $100 mil- lion. SUMMARY AND CONCLUSIONS Removing pollutants from a contaminated aquifer may seem to be an almost impossible task. While this may be true for some contaminated aquifers, others may be amenable to one or more plans for artificial reclamation that could significantly accelerate the rate of water-quality improvement in the aquifer. The feasibility of any such reclamation plan would be strongly dependent on the hydraulic and chemical properties of the aquifer, on the type and source of contamination, and on the duration and areal extent of contamination. Because a variety of reclamation plans can be proposed for any one problem, an accurate model of flow and contaminant transport in the aquifer is an invaluable tool for planning an efficient and effective program. The control and elimination of contaminant migration and contaminant sources at the Rocky Mountain Arsenal represent a large, complex, and costly undertaking (over $25 million has been spent in the Installation Restoration program). An exten- sive well-monitoring program has been required to define the extent of the contamination and the relationships between the sources and contaminant migration patterns. Control of con- taminant migration at the Arsenal boundaries has proved fea- sible using a system involving groundwater interception, treatment, and reinfection. The system was operated success- fully without adversely affecting the flow and distribution of groundwater downgradient from the treatment system, and it has resulted in a significant decrease in the concentration of organic contaminants in groundwater downgradient from the pilot system. Although boundary-control systems can be used successfully to stop or restrict the migration of contaminants off the Arsenal's grounds, they cannot solve the problem of continued contam- inant migration from the source areas to the environment. The overall solution thus involves the control or elimination of the contamination at the sources. A program has been successfully initiated at the Rocky Mountain Arsenal to develop and assess source control and elimination strategies. Through additional data collection and feasibility studies, a single strategy will be selected and implemented using a phased construction ap- proach. The ultimate goal of these activities is to bring the Arsenal into compliance with all applicable federal and state environmental laws and regulations. The great difficulty and expense involved in mitigating
Contamination and Aquifer Recla7nation groundwater contamination problems do not lessen the need to do so; they do illustrate the long-term benefits of planning and designing waste-disposal activities to prevent or minimize future contamination hazards. AC KN OWLE D G M E NTS The Installation Restoration program at the Rocky Mountain Arsenal (RMA) is being funded and directed by the U. S. Army Toxic and Hazardous Materials Agency, Aberdeen Proving Ground, Maryland, in cooperation with the Rocky Mountain Arsenal, Denver, Colorado. The authors wish to thank the personnel from these organizations for their support. Special thanks are extended to Carl Loven, Chief, Process Develop- ment and Evaluation Division, RMA, and Donald Hager, Ru- bel and Hager, Inc., Tucson, Arizona, for providing operational and cost data on the RMA contaminant control systems. REFERENCES Buhts, R. E., P. G. Malone, and D. W. Thompson (1978). Evaluation of ultraviolet/ozone treatment of Rocky Mountain Arsenal ground- water, U.S. Army Engineer Waterways Experiment Station, Tech- nical Report Y-78-1, 78 pp. D'Appolonia Consulting Engineers, Inc. (1979). Evaluation of north boundary pilot containment system, RMA, Denver, Colorado, Proj- ect Number RM79-389, 60 pp. Grove, D. B. (1976). Ion exchange reactions in groundwater quality models, in Advances in Groundwater Hydrology, Am. Water Re- sour. Assoc., pp. 144-152. Konikow, L. F. (1974). Reclamation of a contaminated aquifer, Geol. Soc. Am. Abstr. Programs 6, 830-831. Konikow, L. F. (1975). Hydrogeologic maps of the alluvial aquifer in 103 and adjacent to the Rocky Mountain Arsenal, Colorado, U.S. Geol. Surv. Open-File Rep. 74-342. Konikow, L. F. (1977). Modeling chloride movement in the alluvial aquifer at the Rocky Mountain Arsenal, Colorado, U.S. Geol. Surv. Water-Supply Pap. 2044, 43 pp. Konikow, L. F., and J. D. Bredehoeft (1978). Computer model of two- dimensional solute transport and dispersion in ground water, U.S. Geol. Surv. Techniques of Water-Resources Inv., Book 7, Chap. C2, 90 pp. Petri, L. R. (1961). The movement of saline ground water in the vicinity of Derby, Colorado, in Ground Water Contamination Symposium, Robert A. Taft Sanitary Eng. Center Tech. Rep. W61-5, pp. 119- 121. Petri, L. R., and R. O. Smith (1956). Investigation of the quality of ground water in the vicinity of Derby, Colorado, U.S. Geol. Surv. O pen-File Rep ., 77 pp. Robson, S. G. (1981). Computer simulation of movement of DIMP- contaminated groundwater near the Rocky Mountain Arsenal, Col- orado, T. F. Zimmie and C. O. Riggs, eds., in Permeability and Groundwater Contaminant Transport, American Society for Testing and Materials, Philadelphia, Pa., pp. 209-220. Shukle, R. J. (1975). 1974-75 Groundwater Study of the Rocky Moun- tain Arsenal and Some Surrounding Area, Colorado Department of Health, Denver, Colo., 20 pp. Sweder, R. G., Jr. (1977). Carbon adsorption treatment of contami- nated groundwater at Rocky Mountain Arsenal, Rocky Mountain Arsenal, Denver, Colo., 97 pp. Walton, G. (1961). Public health aspects of the contamination of ground water in the vicinity of Derby, Colorado, in Ground Water Con- tamination Sy7nposium, Robert A. Taft Sanitary Eng. Center Tech. Rep. W61-5, pp. 121-125. Warner, J. W. (1979). Digital-transport model study of diisopropyl- methylphosphonate (DIMP) ground-water contamination at the Rocky Mountain Arsenal, Colorado, U.S. Geol. Surv. Open-File Rep. 79- 676, 39 pp. Wood, L. A. (1972). Groundwater degradation causes and cures, in Proceedings 14th Water Quality Conf., Urbana, Ill., pp. 19-25.