Site Remedial Actions

The history of environmental investigations and regulatory activity related to the Central Chemical facility has been summarized by the Maryland Department of the Environment (1989, 2017) and the U.S. Environmental Protection Agency (2009) and is synopsized here. Briefly, guthion concentrations measured in air samples collected in 1962 in response to odor complaints from residents were deemed nonhazardous by the Maryland Department of Health. In 1970, the Washington County Health Department inspected the facility and directed the consolidation and coverage with soil of refuse disposed onsite and regrading to direct surface runoff away from the lagoon disposal area. DDT was detected in sediments in Antietam Creek in 1976. Soil samples collected from the Central Chemical facility and vicinity, as part of subsequent investigations later that same year, contained DDT, lead, and arsenic at concentrations as high as 1,650 parts per million (ppm), 395 ppm, and 300 ppm, respectively. Because of a subsequent hydrogeologic investigation ordered by the Maryland Water Resources Administration, waste disposal areas at the facility were closed and covered with clay, soil, and stabilizing vegetation. Further environmental investigation was ordered by the Maryland Department of the Environment (MDE) following the discovery of buried “chemical materials” near the onsite former lagoon waste disposal area as part of sewer construction in 1987 and the detection of pesticides, naphthalene, and volatile organic compounds (VOCs) in soils. Soil and groundwater samples collected as part of subsequent studies contained VOCs, pesticides, and heavy metals and borings through the former waste lagoon (fig. 1) included powders and other materials of various colors and petroleum odors. Results of a site inspection published by the MDE in 1996 determined that pesticide contamination in soils at and near the facility posed no significant cancer risk to pedestrians but did pose a slightly increased health risk for young children. The Agency for Toxic Substances and Disease Registry (2005) similarly concluded that exposure to chemicals from the site’s soils, sediments, and water resources represented no apparent hazard to public health.

Following the inclusion of the Central Chemical facility on the NPL in 1997, remedial investigations were conducted in two parallel units (Maryland Department of the Environment, 2017). Soils at the facility and waste material in the former disposal lagoon were included in operable unit 1 (OU-1), whereas groundwater was included in operable unit 2 (OU-2; Maryland Department of the Environment, 2017). Remedial activities related to the soil and former waste lagoon at the Central Chemical facility (OU-1) are described by the U.S. Environmental Protection Agency (2009), and included permanent sequestration of onsite contamination through (1) the solidification and stabilization or excavation and removal of material in the former waste lagoon, (2) consolidation of contaminated soils from elsewhere at the facility to the area of the former lagoon, (3) installation of a low permeability cover over the consolidation area at the former lagoon, and (4) monitoring and treatment of groundwater to prevent the movement of contaminants from the consolidation area. The groundwater treatment system was expected to be completed during 2020 and the solidification and stabilization of the former waste lagoon was expected to begin in 2021 (U.S. Environmental Protection Agency, 2019). Remedial investigations related to groundwater at and near the Central Chemical facility (OU-2) are ongoing.

Purpose and Scope

A conceptual model of the hydrogeology and groundwater contaminant fate and transport at and near the Central Chemical facility (fig. 1) is presented and described in this report. The conceptual model was developed through review, synthesis, and interpretation of the results of hydrogeologic, soil, and environmental investigations conducted at and near the facility in recent decades and is intended to support plans for environmental remediation of OU-2. The geologic setting, groundwater hydrology, and environmental fate and transport of pesticides and other contaminants at the facility are discussed, as well as limitations of the current understanding of contaminant transport within the groundwater system. This report does not include an assessment of risk associated with the COC.

Hydrologic Cross Sections

Three cross sections were constructed for analysis of groundwater flow pathways at the Central Chemical facility. Section A-A' occurs from northwest to southeast along the approximate bedrock dip direction (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009; Brezinski, 2013) from monitoring wells MW-K-80 and MW-K-440 in the northwest to MW-B-55, MW-B-400, and M-B-598 in the southeast (fig. 1). Because of the complexly folded nature of the site’s geology, dip direction and angle vary depending on location on the fold limb, but the bedrock dips primarily perpendicular to the dominant regional northeast-southwest strike. Lagoon piezometer cluster nests PZ-1, PZ-2, and PZ-3, monitoring well MW-A-51, and the sinkhole used in EPA Tracer Test 3 (TT3; Field, 2017; fig. 1) are projected onto section A-A' along strike. Cross sections B-B' and C-C' (fig. 1) occur along the approximate strike direction of bedrock (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009; Brezinski, 2013). Cross section B-B' extends from offsite wells OW-2-65, OW-2-115, and OW-8-230 in the northeast to MW-M-50 located near the southwest flank of the lagoon (fig. 1). Projected onto this section are monitoring wells MW-J-71, MW-O-145, MW-N-83, and MW-N-113, and lagoon piezometers PZ-2S-13 and PZ-2D-34 (where section B-B' intersects section A-A') (fig. 1). Cross section C-C' starts at lagoon piezometers PZ-4D-44 and PZ-4S-17, which are located approximately 135 feet (ft) southeast of MW-M-50 on section B-B'. Monitoring wells MW-P-235, MW-I-55, and MW-Q-150 and the sinkhole used as the injection location for the Environmental Protection Agency (EPA) Tracer Test 2 (TT2; Field, 2017) are projected onto section C-C', which terminates at monitoring well MW-E-46 to the southwest (fig. 1). Despite the proximity of well EW-1-110 to sections B-B' and C-C', this well was omitted from the sections because its longer 50-ft open-interval length limits the usefulness of its groundwater level data.

Reported groundwater level data were contoured on cross sections and two maps to observe the distribution of hydraulic heads in the subsurface and infer characteristics of the groundwater flow system beneath the site. Tables 1.1 and 1.2 contain all reported groundwater data from the site, as well as relevant well construction information. Construction information and groundwater data from 2014, 2016, and 2017 were reported by Amec Foster Wheeler (2018). Data from 2018 were reported by Wood Environment & Infrastructure Solutions, Inc., (2019). Appendix 1 also contains the total head drawdowns during 2019 aquifer tests (Geosyntec, 2019) and tracer-test detections in 2014–2015 (Field, 2017). No data were collected by the USGS. Data from the October 2016 synoptic groundwater measurement event were the lowest groundwater levels reported for much of the site (appendix 1). Data from the October 2018 synoptic groundwater level measurement event were generally the highest groundwater levels reported (appendix 1). Thus, contouring data from both October 2016 and October 2018 allows for comparison of differences in conceptual groundwater flow pathways between dry (October 2016) and wet (October 2018) conditions. The October 2018 dataset was collected during the wettest calendar year and wettest fall season on record for Maryland (National Oceanic and Atmospheric Administration, 2019), which is likely what contributed to the higher groundwater levels during that measurement event compared to earlier events. After contouring reported groundwater levels on cross sections, maps were created to observe subsurface groundwater level conditions at specific depths or elevations in the subsurface. Potentiometric surface maps were produced at the bottom of the epikarst and at an elevation of 52 ft (North American Vertical Datum of 1988 [NAVD 88]). Further explanation of the approach used in contouring the cross sections and maps is discussed later in this report.

Annual Evapotranspiration Calculations

Estimates of evapotranspiration (ET) were produced using the operational remote-sensing-based Simplified Surface Energy Balance (SSEBop) model (Senay and others, 2013) from 2003 to 2019. The SSEBop is based on the Simplified Surface Energy Balance (SSEB) approach (Senay and others, 2007, 2011). The global ET is derived from the integration of Moderate Resolution Imaging Spectroradiometer (MODIS)-based land surface temperature, maximum air temperature from WorldClim, and reference ET derived from global data assimilation systems (GDAS). A comprehensive validation of the SSEBop model was conducted by Velpuri and others (2013) over the conterminous United States at both point and basin scales.

Contaminant Transport

To evaluate contaminant transport in the epikarst, four wells (MW-A-51, MW-B-55, MW-M-50, and MW-E-46) were selected based on their depth below the surface and location in the study area. Along cross section A-A' (fig. 1), MW-A-51 and MW-B-55 were determined to be completed within the epikarst of the site (described later in the Epikarst section of this report). MW-A-51 is located to the southeast of the lagoon waste disposal area, whereas MW-B-55 is located at the eastern edge of the site boundary. Dye from the 2017 EPA tracer test study was detected in MW-A-51 demonstrating a connection between the surface where the dye was released and the well (Field, 2017). MW-M-50 is located just to the southwest of the lagoon waste disposal area along the transect B-B'. MW-E-46 is shown on cross section C-C' and is located to the southwest of the waste disposal lagoon at the boundary of the site (fig. 1).

Concentrations of COC measured in samples collected from MW-A-51, MW-B-55, MW-M-50, and MW-E-46 are presented in appendix figures 2.1 to 2.16. Concentrations of COC and water quality measurements were extracted from Wood Environment & Infrastructure Solutions, Inc. (2019) 2018 Groundwater Monitoring Report for OU-2 and are available through the U.S. Environmental Protection Agency (2021). COC concentrations measured during sampling events were plotted then fitted with a local polynomial regression fitting (LOESS) as a solid line with corresponding standard errors in gray. Years 2016 and 2018 are highlighted with red bars corresponding to the “dry” and “wet” conditions previously described. These two years provide an opportunity to evaluate how changes in the water table affect COC.

To evaluate contaminant transport within the conduit system or dissolution voids, three compounds were selected to compare between the different wells. Wells MW-I-55, MW-J-71, MW-M-50, MW-N-113. MW-O-145, MW-P-235, OW-2-65, OW-2-115, and OW-8-235 were selected for comparison due to proximity to possible dissolution voids reported by Weston, Inc. (1989) and Amec Foster Wheeler (2018) (table 1). Chlorobenzene was identified as a compound of interest for comparison due to its frequent detections in many of the wells (appendix 2). For comparison, 1,2,4-trichlorobenzene (1,2,4-TCB), n-octanol-water partition coefficient (log Kow = 4.02), was selected due to its similarity to chlorobenzene (log Kow = 2.84) and was selected due to its similarity to chlorobenzene (log K_ow=2.84) but is more hydrophobic and more likely to sorb to matrix materials than chlorobenzene. 1,2,4-TCB has a hydrophobicity closer to benzene hexachloride (BHC) isomers (log K_ow=3.72), commonly grouped as technical-BHC. COC concentration data are plotted by well with respect to time and fitted with a LOESS as a solid line (appendix 3). Years 2016 and 2018 are highlighted with red bars to compare groundwater level effects as previously described. Box and whisker plots were generated for each well and grouped by compound to evaluate variability in concentration by compound and well (appendix 3).

Soil

Soil at the Central Chemical facility is generally fine grained, varies in thickness, and is classified as Hagerstown silt loam, which is 80 percent silt with some clay and laterally discontinuous thin sand lenses (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009). The soil is derived from chemical weathering of the underlying carbonate bedrock, is generally well drained, and varies in thickness at the facility from 0 ft at bedrock outcrops to more than 40 ft (fig. 2) (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009; Amec Foster Wheeler, 2018). Surface permeability ranging from 0.06 to 0.6 inches per hour (in/h) have been reported for the Hagerstown silt loam (Maryland Department of the Environment, 1993). Overlying the natural soil in some locations is as much as 13 ft of fill, including brown silt with some coarse-to-fine sand and gravel (U.S. Environmental Protection Agency, 2009).

Geologic Formations

The Central Chemical facility and vicinity are underlain by the Conococheague Formation, Stonehenge Limestone of the Beekmantown Group (Stonehenge Limestone), and Rockdale Run Formation of the Beekmantown Group (Rockdale Run Formation; figs. 3 and 4), which include primarily carbonate rocks (Brezinski, 2013, 2018). The lithology, stratigraphy, and nomenclature applied to rocks of these formations in the Hagerstown Valley and wider Great Valley are described in detail in Brezinski (2018) and are briefly summarized herein.

The Central Chemical facility is underlain by the Zullinger and Shady Grove Members of the Conococheague Formation (Brezinski, 2018; figs. 3 and 4). The Zullinger Member, previously described as the “middle member” of the Conococheague Formation, includes 1,600 to 1,900 ft of alternating sequences of thick beds of thrombolitic limestone and ribbony, laminated dolomitic limestone. These sequences alternate on the scale of 200 to 300 ft. Above these alternating sequences of the Zullinger Member are younger ribbony limestone, sandy dolomite, and chert of the Shady Grove Member, which was previously described as the “upper member” of the Conococheague Formation. The Shady Grove Member of the Conococheage Formation is between 350 and 500 ft thick.

Overlying the Conococheague Formation to the southeast and northwest of the Central Chemical facility are the younger rocks of the Stonehenge Limestone (figs. 3 and 4), which include three members with a combined thickness of approximately 850 ft (Brezinski, 2018). At the base of the Stonehenge Limestone is the Stoufferstown Member, which is 175 to 275 ft thick and composed of thin (less than or equal to 1 inch) beds and ribbons of siliceous lime mudstone. Included among the thin beds of the Stoufferstown Member is an atypical 3- to 9-ft-thick massive thrombolitic lime mudstone. The Stoufferstown Member is overlain by the Funkstown Member of the Stonehenge Limestone, which is 350 to 450 ft in thickness and composed of thick beds of organically bound limestone. Above the Funkstown Member is the Dam Five Member of the Stonehenge Limestone, which is composed of medium bedded, locally ribbony mechanically deposited limestone that varies in thickness from 330 to 450 ft. The Funkstown Member and Dam Five Member were previously mapped as the “middle” and “upper” members of the Stonehenge Limestone, respectively.

The Rockdale Run Formation overlies the Stonehenge Limestone in the vicinity of the Central Chemical facility (figs. 3 and 4) and exceeds 2,700 ft in thickness (Brezinski, 2018). Included within the lower 350 to 450 ft of the formation are cyclical sequences of organically bound limestone, limestone ribbons, laminated dolomite, and chert (Brezinski, 2018). These rocks are overlain, in turn, by interbedded oolitic lime packstone and dolomite, approximately 200 ft thick, and by more than 1,000 ft of additional cyclical carbonates (Brezinski, 2018).

Geologic Structure

The complex folding and faulting of the carbonate geologic units in the Hagerstown Valley reflect compression in a generally westward direction that occurred during the orogenic event that created the Appalachian Mountains (Maryland Department of the Environment, 1989; Brezinski, 2018). This compression created folds in the rock units spanning several orders of magnitude in scale with axes that trend generally to the north-northeast (Brezinski, 2018; fig. 5). Fractures, including bedding-plane partings, joints, and faults, are common in the rock units in the valley and are commonly oriented parallel or perpendicular to the north-northeastern strike of the rock formations (fig. 5); faulting reflects movement over scales from inches to hundreds of miles (Brezinski, 2009, 2018).

The Central Chemical facility is located to the southeast of the axis of a structural anticline that trends to the north-northeast and plunges to the southwest (fig. 4) (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009). Bedding planes at and near the facility generally dip more steeply (between 55 and 90 degrees) on the northwestern limb of this anticline than on the southeastern limb (between 25 and 45 degrees) (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009).

Secondary meso-scale folds also occur at the site and have been named the Western Boundary Anticline, the Former Plant Syncline, the Former Plant Anticline, the Drainage Swale Syncline, and the Mitchell Avenue Anticline (Amec Foster Wheeler, 2018). Bedding planes dip more steeply to the west than to the east on either side of the secondary anticlines, similar to those of the primary anticline (U.S. Environmental Protection Agency, 2009). Section A-A' of figure 6A, 6B is aligned with the approximate dip of the regional anticline and intersects several of the meso-scale folds beneath the facility. Although no attempt at bedding correlations was made, the generalized fold structure reported in the cross sections in Amec Foster Wheeler (2018) was checked against reported acoustic televiewer logs collected during construction of boreholes MW-K-440 and MW-B-400 in Amec Foster Wheeler (2018). Borehole MW-K-440 had reported dominant fracture azimuths to the north and west-northwest in the upper 100 ft and to the east and east-southeast at a depth from 100 to 250 ft. These orientations are consistent with MW-K-440 being located within the Western Boundary Anticline, where the upper portion dips toward the northwest near the tilted axial trace and the deeper portion along the lower southeast-dipping limb. For borehole MW-B-400, the dip azimuths are predominantly toward the northwest and southeast directions, which is consistent with the location of MW-B-400 near the middle of the axial trace of the Mitchell Avenue Anticline. In the upper 200 ft, the orientation is more dominant to the southeast, so MW-B-400 was drawn on section A-A' as being southeast of the axial trace (fig. 6A, 6B).

The axial trace of the Drainage Swale Syncline was mapped by Amec Foster Wheeler (2018) as occurring approximately 150 ft northwest of MW-B-55 at the screen elevation and was incorporated on cross section A-A' (fig. 6A, 6B). Although the reported Former Plant Syncline (Amec Foster Wheeler, 2018) occurs along section A-A', it was omitted because of the unclear structure of this fold and the Former Plant Anticline that is reported to terminate south of this cross section. The cross sections interpreted by Amec Foster Wheeler (2018) that cross this area also do not include the Former Plant Syncline.

The geologic structure depicted on the cross sections was not determined by any other means than as represented in Amec Foster Wheeler (2018). Therefore, the actual dip angles and overall structural geometry may differ from those depicted here, which should be considered generalized approximations. On the northwest-dipping limb of the Western Boundary Anticline and the limb between he Drainage Swale Syncline and Mitchell Avenue Anticline on section A-A' (fig. 6A, 6B), the dips are approximately 45 degrees. Beds located on the limb between the Western Boundary Anticline and Drainage Swale Syncline are drawn by Amec Foster Wheeler (2018) as dipping approximately 27 degrees southeast (100 ft in elevation for every 200 ft in distance). Sections B-B' and C-C' are aligned approximately in the northeast-southwest strike direction and no meso-scale folds are included on these sections (fig. 7A, 7B, 7C). Amec Foster Wheeler (2018) shows beds tilted 14 degrees southwest along the strike direction (decreasing approximately 100 ft in elevation over 400 ft of distance). The same bed orientations were incorporated on cross sections B-B' and C-C' as those in Amec Foster Wheeler (2018), but whether these orientations correspond to the strikes and dip angles mapped by Brezinski (2013) is unknown. Since many fewer structural data are available along the strike direction, the true structure along strike is unclear.

Rock fractures observed in the vicinity of the Central Chemical facility are generally oriented in a manner similar to those in the wider Hagerstown Valley. Fractures near the land surface generally strike to the north-northeast or to the northwest and are steeply inclined or nearly vertical (Maryland Department of the Environment, 1989). These fractures may be wider near the crest of anticlines due to extensional stresses during convex folding (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009). Rock strata are offset along a reverse fault to the west of the facility near the axis of the major anticline (fig. 3; Brezinski, 2013; Field, 2017).

Karst Features

Karst features, including closed depressions, sinkholes, springs, and caves, are common in carbonate rocks of the Hagerstown Valley (Duigon, 2001; Brezinski, 2018). Because these features are formed through the dissolution of the rock matrix by groundwater, their occurrence is related to the solubility of individual lithologic units and to the distribution of fractures available as conduits for groundwater flow (Brezinski, 2018). Karst features occur in all rock formations in the vicinity of the Central Chemical facility, although the Rockdale Run Formation is generally more susceptible to karst development than the Conococheague Formation, but less so than the Stonehenge Limestone. In the Hagerstown Valley, springs and sinkholes commonly occur near the surface trace of faults and rock dissolution is common along bedding planes. Solution cavities that form along joints in the rock tend to be wider in formations with more massive bedding and more purely carbonate lithology. Karst features are also common in areas where groundwater flow may be increased by a locally depressed water table or in areas of stormwater impoundments or unlined drainage.

Karst features have been observed at and near the Central Chemical facility. Numerous sinkholes and closed depressions have been reported at and within a few kilometers of the facility (Maryland Department of the Environment, 1993; Field, 2017). Caverns and voids in the bedrock exceeding 5 meters (m) in height have been encountered during subsurface drilling at the facility and in-situ permeability as high as 1,042 feet per year has been reported (Maryland Department of the Environment, 1993). Fractures enlarged by solution have been observed in outcropping rocks of the Conococheague Formation at the facility (Field, 2017).

Overlying the Conococheague Formation at the Central Chemical facility is highly fractured bedrock of the epikarst (U.S. Environmental Protection Agency, 2009). The contact of the bedrock with the overlying soil and epikarst is highly irregular and the soil thickness consequently varies substantially (Maryland Department of the Environment, 1989; U.S. Environmental Protection Agency, 2009). Vertical pathways of high permeability have been reported in the epikarst at the facility (U.S. Environmental Protection Agency, 2009).

Surficial and Epikarst Hydrogeologic System

The shallow hydrogeologic system encompasses the unsaturated vadose zone through the depths at which the regional water table fluctuates within the bedrock and consists of two portions: (1) the surficial sediments, and (2) the subcutaneous epikarst portion of bedrock.

Bedrock Hydrogeologic System

The site is underlain by the Zullinger and Shady Grove Members of the Conococheague Formation. In the site area, the Stonehenge Limestone overlies the Conococheague Formation. Amec Foster Wheeler (2018) categorized the bedrock at the site into three general lithologic units: a dark gray, fine-grained limestone with a high degree of fracturing; a light to medium gray, coarse-grained limestone with few fractures but large apertures when present; and a light gray to tan coarse-grained limestone/dolomite interbedded with fine, dark gray limestone. Amec Foster Wheeler (2018) mapped the contact between the Conococheague Formation and Stonehenge Limestone at the site as a thrust fault. Bedding correlations between boreholes from Amec Foster Wheeler (2018) were assumed to be correct and no attempt was made to recategorize these geologic units or reinterpret bedding correlations between boreholes. Bedding on cross sections A-A', B-B', and C-C' in this report (figs. 6 and 7) are included to delineate the generalized schematic bedrock structure geometry only and do not imply any location-specific or borehole-specific lithologic information.

Dissolution of carbonate bedrock will occur preferentially within joints and bedding plane fractures determined by the geologic structure (Duigon, 2001; Brezinski, 2018); in other words, the structure of the dissolution voids (tertiary porosity) follows the structure of joints and fractures (secondary porosity). Groundwater flows more easily through fractures than through the primary porosity of the rock matrix. As groundwater flows through fractures over time, acidic water will cause dissolution, enlarging fractures. Rock matrix not adjacent to fractures has lower permeability and porosity than fractures, thereby more quickly neutralizing dissolved acids in the groundwater. Groundwater flow is slow through the rock matrix and acids can be more quickly neutralized and dissolution of the bedrock stops. The amount of time to replace the “used up” neutralized groundwater with incoming acidic groundwater that will restart dissolution will subsequently be longer than in fractures. Conversely, acids in groundwater flowing through fractures may not be neutralized and the replacement of these acids in the fractures and joints will be faster than in the rock matrix because groundwater in fractures is being replaced continuously, and thus these features will become enlarged by dissolution more readily and to a greater degree than in the rock matrix. The prevalence of dissolution-enlarged fractures typically decrease with depth (Duigon, 2001), but groundwater flow directions in bedrock occur predominantly along the structure of the bedrock, regardless of whether that feature is associated with secondary or tertiary porosity (Amec Foster Wheeler, 2018).

Three scales at which structural bedrock folds occur each affect groundwater flow differently. Regionally, the site’s location on the limb of the large anticline whose axis is located to the northwest of the site (Brezinski, 2013, 2018) will affect direction(s) of groundwater flow away from the site and may affect interpretation of tracer-test detection at springs. Too few offsite data are available for this study to accurately discuss groundwater flow directions at the regional scale. Imprinted on this regional fold are smaller meso-scale folds that can be considered “site-scale” structural features with the greatest effect on groundwater flow and contaminant transport within the boundaries of the site. These folds include the Western Boundary Anticline, Drainage Swale Syncline, Mitchell Avenue Anticline, and other named folds (Amec Foster Wheeler, 2018). Micro-scale folds also occur with the potential to affect groundwater flow at the borehole scale. Micro-scale folds are important for packer test and geophysical log interpretations, but likely have negligible effect on site-scale processes.

Dissolved Phase Aqueous Transport

Contaminants transported within the dissolved phase of groundwater will primarily follow the dominant groundwater flowpath (advective transport) as discussed within the Hydrology section of this report. COC dissolve into groundwater either through recharge events where rainwater dissolves soluble COC from contaminated soils or desorb hydrophobic contaminants from contaminated soils and sediments. Depending on the chemical properties of the COC, dissolved phase transport can be rapid (in the case of soluble, freely available COC) or slow (due to diffusion from within particles). The concentrations of water-soluble compounds are generally limited by the solubility of those compounds if raw source material is present such as in the waste lagoon. As groundwater moves from the source material, dilution and dispersion occur as the groundwater moves along the dominant flow path. While solubility limits do exist for more hydrophobic compounds, the dissolved phase within groundwater is generally limited by partitioning between the dissolved (aqueous) phase and the solid phase where compounds are sorbed to a solid matrix. Dissolved hydrophobic contaminants within the groundwater will sorb and desorb from solid materials based on the thermodynamic chemical activity of the water and the solid materials. The chemical activity of a compound in a mixture is used to define chemical potential and determine directional flux as a system moves to equilibrium. A compound is at equilibrium between states when the chemical activities are equal. When contaminated groundwater with a high activity coefficient is in contact with solid materials with a lower activity coefficient, the contaminants will sorb to the solid material due to the directional flux from high to low. Contaminants will continue to sorb until equilibrium is established. If the concentration within the groundwater changes to a lower activity coefficient, such as a period of high recharge, the thermodynamic flux is reversed and COC are mobilized into the dissolved phase. The contact time of the groundwater with the source material can have an impact on the concentration. Slow desorbing hydrophobic (log K_ow>4) compounds may increase in concentration within the groundwater during periods of low recharge due to the longer contact time.

Inorganic contaminants and metals found within the groundwater at the study site are affected both by solubility and the redox chemistry of the groundwater. Changes in the redox chemistry or pH of the site may mobilize or immobilize contaminants within the groundwater. The carbonate rock found within the bedrock system has a buffering effect on the pH. Changes in groundwater chemistry within the epikarst layer would be more variable as recharge enters the groundwater and water levels change within the unconsolidated overburden. Fast recharge events coupled with a change in redox state may result in mobilizing contaminants to deeper portions of the aquifer via conduit flow.

As mentioned in the hydrology discussion of this report, the water table periodically rose above the disposal lagoon. During high water levels, groundwater was likely in direct contact with raw materials such as DDT that would typically be bound to soils and sediments, increasing the potential for mobilization.

Colloidal (Particle-Bound) Transport

White (2018) identified two potential transport mechanisms for hydrophobic COC in a karst system. First, water-soluble contaminants and dissolved COC generally move with the dominant flow path of the water in an aquifer. Second, hydrophobic compounds move and are stored within the system sorbed either to particles or colloids that do not necessarily correspond with the movement of water in an aquifer. Dense non-aqueous phase liquids (DNAPLs), such as the chlorinated organic contaminants (for example, 1,2-dichlorobenzene and 1,2-dichloroethane), can settle into pockets, such as voids or caves, within the conduit system where the contaminants can accumulate with settled particulate materials (Vesper and others, 2001). Slow mobilization of hydrophobic contaminants into the dissolved phase then can occur through a slow diffusion process as less contaminated water flows over deposited contaminated sediments. In the event of higher energy flows within the conduit, a second mobilization of the sediment can occur rapidly transporting the sediment to a new location. This type of mobilization, caused by high energy flows within the conduit system, would be difficult to observe and measure since this type of flow can be episodic and unpredictable. Additionally, wells are traditionally installed and screened to reflect the dominant groundwater flow and not to target a conduit with irregular flows patterns. Due to the hydrophobicity of many of the COC, the mass associated with the dissolved phase will likely be many orders of magnitude less than the concentration associated with the solid phase.

Various epikarst features, such as mud-filled voids and caves, were observed during well installations as part of the Remedial Investigation and other OU-2 remedial actions at the site. These features were observed between 18 and 106 ft depth. The source of the mud and sediment filling the voids was not determined at the time of the installation, however, three of the wells (MW-A-51, MW-N-113, and MW-O-145) connect with surface-level conduits (sinkholes) demonstrated by the detection of dye in these wells during the dye tracer study conducted by the EPA in 2017 and will be discussed in a later section of this report. It is important to note that the dye for the EPA study was not injected into a well but into sinkholes present at the site. Without analysis of the material found within the voids, it is not possible to conclusively determine if COC are present, but there is evidence that these voids are connected hydraulically with a surface sinkhole at the site. The connection between the surface and an epikarst conduit system could allow for particle-bound COC to enter and accumulate within these voids. As discussed in the Epikarst section of this report, perched epikarst water tables can form vertical “dead ends” for groundwater flow before reaching the aquifer. These features can trap COC that may mobilize to the deeper aquifer under specific conditions or water levels. Storm sampling events in June 2003 observed site-related pesticides and heavy metals within the stormwater leaving the site via stormwater sheet flow that also may enter the various sinkholes present on the site (U.S. Environmental Protection Agency, 2009). An initial site investigation report observed some onsite surface-water travel to a depression that contains a sinkhole approximately 2 ft in diameter located to the north of the entrance to Central Chemical Corporation (Maryland Department of the Environment, 1989). COC entering the groundwater system from contaminated material at the surface will not contain the same composition as the COC from the disposal lagoon. High levels of COC were found throughout the site from improper handling during the manufacturing process and disposal which included dumping material down sinkholes (Maryland Department of the Environment, 1989). Well locations and screen intervals were selected to monitor the site contaminant transport based on identification of the disposal lagoon as a primary source of groundwater contamination. As discussed in the Hydrology section of this report, the only location where the overburden was sampled was in piezometers associated with the lagoon. Without additional sampling locations within the site, it is not possible to determine the extent and transport of COC associated with non-lagoon sources.

Groundwater monitoring activities for OU-2 in 2018 and 2019 reported detectable quantities of total suspended solids (TSS) that lessened in quantity with depth, as shown in figure 11. Most TSS were detected within or near the epikarst zones but also were detected in deeper parts of the system indicating the potential for sedimentation at depth to occur. Without analysis of the TSS, it is not possible to determine the source of the TSS, although they provide additional evidence of sedimentation occurring within the groundwater flow system.

Contaminant Transport Near the Water Table

Wells MW-A-51 and MW-M-50 are located on the periphery of the waste disposal lagoon. Concentrations of VOCs within MW-A-51 do not show a decreasing trend but are highly variable (appendix 2; fig. 2.1). During 2016 when groundwater levels were low, concentrations of 1,2-dichloroethane, 1,2,4-TCB, benzene, chlorobenzene, ethylbenzene, and toluene increased. Concentrations of these compounds decreased during 2018 and fit closer with the LOESS. The increase in 2016 may be the result of VOCs being released from contaminated material with longer contact time to enable material to diffuse out of the matrix. Concentrations of VOCs in MW-M-50 (fig. 2.5) do not share the same variability as MW-A-51 with lower concentrations measured in 2016. COC in MW-M-50 (figs. 2.5–2.8) generally follow a trend of elevated concentrations prior to 2010, followed by lower concentrations after 2010. The general pattern within MW-M-50 follows a contamination source based on desorption-dissolution from contaminated soils in which levels would decrease with time as the source material is depleted. The exception to this pattern is observed in chlorobenzene, diphenamid, and bis(2-ethylhexyl)phthalate (fig. 2.6). These three COC remain elevated and may be the result of a different source material not associated with the waste disposal lagoon. The presence of elevated levels of bis(2-ethylhexyl)phthalate may be related to an indirect release of plastics including polyvinyl chloride (PVC). Solvents disposed of at the site could leach bis(2-ethylhexyl)phthalate into the groundwater where they came into contact with various plastics.

Wells MW-B-55 and MW-E-46 are located at the edge of the study area boundary. VOCs and SVOCs generally decrease in concentration over time within these wells (appendix 2; figs. 2.9, 2.10, 2.13, 2.14). Pesticide concentrations of DDX (combined isomers of DDT, DDD, and DD), endrin, gamma-BHC, alpha-BHC, and pentachlorophenol all decrease with time in well MW-E-46 (appendix 2, fig. 2.15). A bell-shaped pattern is observed for beta-BHC and heptachlor epoxide that is not observed in MW-B-55 (appendix 2, fig. 2.11). The concentration of the SVOC diphenamid in well MW-B-55 (appendix 2, fig. 2.10) increases with time possibly caused by recharge desorbing contaminants from surface soils then reabsorbing to the deeper materials. Desorption could result in a steady increase in diphenamid until either the source material is removed or the matrix reaches equilibrium.

The variable COC concentrations and patterns observed near the water table do not provide conclusive evidence of dominant transport mechanisms or sources. The variability observed between different COC within the same well may be the result of additional source materials unconnected to the disposal lagoon. Storage of contaminated material within the epikarst system has the potential to hinder remediation efforts if not considered in the remedial action. Gaps in the sampling frequency add uncertainty to COC transport if affected by water-level changes within the epikarst as observed in MW-A-51. Increased sampling frequency and a targeted approach to capture changes in water level and recharge events may provide clarity in understanding COC transport within the epikarst.

Contaminant Transport Within Dissolution Voids

Specific wells depicted in sections B-B' and C-C' in figure 7 contain reported dissolution voids with possible connections, as seen in the figure. Additionally, these cross sections contain wells with tracer detections from the EPA tracer test study (Field, 2017). Therefore, these wells indicate possible contaminant transport within dissolution voids. A possible conduit connection from MW-N-113 to MW-I-55 on section C-C' may be aligned with the sinkhole used as the injection location during EPA Tracer Test 2. Dissolution features also are present in MW-N-83 and may also be connected to this sinkhole. Wells MW-J-71 and MW-N-113 also have a possible void connection within section B-B’. Although no void is depicted in wells MW-P-235 and MW-N-113, there is some groundwater connectivity based on detection of dye during the EPA tracer test study (Field, 2017). Wells OW-2-65, OW-2-115, and OW-8-230 located offsite also were evaluated due to the possible dissolution void connecting MW-J-71 to OW-2-115.

TSS concentrations in MW-I-55 measured in 2018 and 2019 were between 15 and 55 microgram per liter (mg/L), which may result from direct recharge through the sinkhole used in the tracer test study. MW-J-71 also had elevated levels of TSS measured in 2018 and 2019 with concentrations between 17 and 67 mg/L. MW-J-71 and MW-I-55 both had the highest concentrations of TSS of the wells sampled that was not a shallow piezometer. Elevated chlorobenzene levels were measured in MW-J-71 with fluctuating concentrations and no observed decrease over the sampling period (fig. 12A). 1,2,4-TCB exhibits similar elevated levels but at an order of magnitude lower concentration (fig. 12B). The more hydrophobic 1,2,4-TCB (log K_ow=4.02) has higher sorption to solid materials (like TSS) than chlorobenzene (log K_ow=2.84), resulting in approximately an order of magnitude difference in the aqueous concentration if the initial release were the same. BHC isomers have a similar hydrophobicity (log K_ow=3.72) to 1,2,4-TCB, but concentrations decrease with respect to time. Chlorobenzene and 1,2,4-TCB may be associated with accumulated sediments within the dissolution void acting as a reservoir for the contaminants.

Evaluating wells within cross section B-B’ (fig. 7) that may be connected to MW-J-71, elevated levels (greater than the RSL) of chlorobenzene and 1,2,4-TCB were observed in MW-O-145 and OW-2-115 though there was no significant pattern change with time (appendix 3). TSS concentrations in MW-O-145 were between 12 and 21 mg/L and in OW-2-115 between 9.8 and 32 mg/L. MW-O-145 is not depicted with a possible void connection and is within the vicinity of MW-J-71 (fig. 7). MW-O-145 also had detections of dye during the EPA tracer test with some connection to surface conduit flow systems. The elevated levels of both chlorobenzene and TSS in OW-2-115 support the possible void connection depicted in cross section A-A' of figure 7C. Within the well cluster containing OW-2-115, TSS were measured in OW-8-235 to range from 4.0 to 7.8 mg/L and had dye detections during the 2017 study. The shallowest well within the cluster, OW-2-65, had TSS concentrations between 4.0 and 7.0 mg/L. Elevated levels of contaminants generally were not observed in OW-2-65, as concentrations were significantly lower than deeper wells within the cluster. Chlorobenzene concentrations within OW-8-230 increased with time. This increase is not observed with 1,2,4-TCB or technical-BHC. Chlorobenzene associated with particles may be a transport mechanism offsite via the possible dissolution void connecting MW-J-71 to OW-2-115.

Well MW-N-113 is also depicted with a possible void connection to MW-J-71 and has some connection to MW-P-235 based on detection from the EPA tracer test (fig. 7). Chlorobenzene concentrations in both MW-N-113 and MW-P-235 decreased with time supporting the hydraulic connectivity observed in the tracer test (fig. 3.1). Elevated levels of TSS were not observed in MW-N-113 during this same period with concentrations near the detection limits, which may be the reason for the decrease in concentration with respect to time if there is no significant storage of contaminant within sediments allowing for dilution of the contaminants. MW-P-235 had detectable quantities of TSS between 5.4 and 8.0 mg/L and may have buffered reduction with time due to TSS compared to MW-N-113. 1,2,4-TCB also decreased with time in MW-N-113 but was variable in MW-P-235 (fig. 3.3). Technical-BHC decreased with respect to time in MW-N-113 until the wetter period of 2018, where the concentration increased to earlier levels (fig. 3.5). Although the void connecting MW-J-71 to MW-N-113 may exist, it does not seem to be contributing to contaminant transport associated with particles.