2019_March_Fate_and_Transport_Rept_Geosyntec_.pdf

Analysis of Fate and Transport of Soil Contamination from the Broadkiln Sportsman’s Club to Prime Hook National Wildlife Refuge, Milton, Delaware

Prepared for

U.S. Fish and Wildlife Service Department of the Interior

Prepared by Geosyntec Consultants, Inc.

520 Pike Street, Suite 2600 Seattle, WA 90101

Project Number PNG0825

March, 2019

TABLE OF CONTENTS

Contents

1. INTRODUCTION

1.1 Site Description

1.2 Summary of Previous Investigations

1.2.1 Initial Service Response (2000)

1.2.2 USGS Groundwater Investigation (2001)

1.2.3 Interim Remedial Action (2002)

1.2.4 Ecological Risk Assessment (2004)

1.2.5 Groundwater and Surface Water Sampling (2007-2015)

1.2.6 Amphibian Investigations (2002-2011)

1.2.7 Supplemental Site Investigation (2017)

1.2.8 DNREC Site Investigation of BSC (2017)

1.3 Study Objectives and Scope

2. LEAD PELLET MASS

2.1 Methodology

2.2 Results

3. GEOCHEMICAL ANALYSIS

3.1 Methodology

3.2 Results

4. HYDROGEOLOGICS ANALYSIS

4.1 Methodology

4.1.1 Partitioning Coefficient (Kd) and Transport Velocity

4.1.2 Initial concentration (C0)

4.2 Results

4.3 Surface Water Transport

4.4 Air Transport

4.5 Preliminary Evaluation of Pb Immobilization

5. CONCLUSIONS

6. REFERENCES

LIST OF TABLES

Table 1 Estimated Pb Mass per Pellet Table 2 Estimated Pb Mass on Service and BSC Property Table 3 Water Quality Inputs to Geochemical Model Table 4a Source Area Water Volumes Table 4b Source Area Pb Mass Balance Table 5 Screening Level Summary of Lead Immobilization Technology

LIST OF FIGURES

Figure 1 – Site Vicinity and Surface Water Features Figure 2 – Estimated Density of Lead Shot Figure 3 – Geochemical Model Results: eh/pH Stability Diagram Figure 4 – Groundwater Elevation Map Figure 5 – Observed Lead Concentrations in Groundwater Figure 6 – Groundwater Transport Model Results Figure 7 – Potential Groundwater Remediation Approaches

LIST OF APPENDICES

Appendix A : Calculations and Groundwater Modeling Analysis

Page i

US Fish and Wildlife Service/Prime Hook NWR March, 2019

Executive Summary

This report evaluates lead contamination on the northwest portion of Prime Hook National Wildlife Refuge (NWR). The source of lead in soil and groundwater is an adjacent property called the Broadkiln Sportsman’s Club (BSC), which operated a shooting range from the early 1960’s until late 1990’s. The US Fish and Wildlife Service (the Service) has removed some lead-contaminated soils and installed a geotextile cap on approximately 0.6 acres of Service property adjacent to the BSC. No remediation has occurred on the BSC site. The Service is considering additional remedial actions at the Site and seeks to understand whether further remediation would be successful without also initiating remediation on BSC property. The goal of this study was to develop a conceptual level fate and transport analysis of lead contamination from the BSC.

Several studies have been conducted on the Service property beginning in the early 2000’s, but a site investigation had not been conducted on the BSC property until recently.

This investigation was completed by the Delaware Department of Natural Resources and Environmental Control (DNREC) in 2017 and included multiple soil borings, monitoring wells, soil sampling and groundwater quality analysis. In general, the higher lead soil concentrations lead were located on the eastern half of the BSC property, particularly adjacent to the property boundary. Only one of the monitoring wells showed elevated lead concentrations in groundwater above 10 ug/L, but this well was the only well located within 200 feet of the refuge boundary. Based on the available soil data and shot-count data. The effective source area for lead contamination is approximately 9.4 acres in size, straddling the property boundary between the BSC and the Refuge.

A comprehensive 3-dimensional fate and transport analysis of lead in groundwater and surface water could not be accomplished with the available data and resources. The analysis focused on mass balance approaches and a simple one-dimensional groundwater and geochemical analysis. The modeling results are summarized below:

• Pb remains dissolved in groundwater under the geochemical conditions observed in Site groundwater.

• The effective source area concentration of Pb in groundwater is between 3 and 12 mg/L. This corresponds to a total loading rate to groundwater of between 1,950 and 7,800 kg of Pb over a 50 year period.

Page ii

• Pb concentrations in groundwater are expected to decrease steeply below 2 ug/L beyond about 300 feet from the edge of the source area. The rate of decrease and concentration profile is sensitive to the hydraulic properties of the aquifer, the organic content of the aquifer material, and the redox conditions in the groundwater.

• Total lead that has entered Refuge property by a surface water pathway was estimated using a mass balance calculation based on the loading rate to groundwater determined from the groundwater model. It is estimated that between 4,550 kg and 18,200 kg of Pb lead has dissolved into surface run-off over the past 50 years.

• The total remaining mass of Pb on site is estimated to be approximately 36,500 kg (7,400 kg on Service property and 18,500 kg on BSC property).

• The potential for airborne transport of Pb to Service property is low since there is no active disturbance of Pb-contaminated soils and the property is heavily vegetated with humid conditions.

• Removal of the remaining mass of Pb on Service property would reduce, but not eliminate, continued leaching and long-term accumulation of Pb in both groundwater and surface water within the Refuge.

• If the mass of Pb on the BSC property is not removed, it will remain a long-term source of Pb to groundwater and surface water on Service property.

Based on these analyses, Geosyntec recommends:

1. Removal of the remaining Pb-contaminated soils on Service property. This will reduce, but not eliminate, the source of Pb to groundwater and surface water downgradient of the BSC property.

2. Evaluation of groundwater flowpaths and hydraulic gradients on Service property north of the area that has been excavated and capped and in the area of the un-named tributary east of the site.

3. Conducting a bench-scale laboratory test to evaluate whether phosphate amendments could potentially immobilize Pb in site groundwater and soils and be used as a remedial technology to prevent migration of Pb contamination from the BSC property.

1. INTRODUCTION

Geosyntec Consultants, Inc. (Geosyntec) has prepared the following report for the U.S.

Fish & Wildlife Service (the Service) summarizing our assessment of lead contamination on the northwest portion of Prime Hook National Wildlife Refuge (NWR). The source of lead in soil and groundwater is an adjacent property called the Broadkiln Sportsman’s Club (BSC), which operated a shooting range from the early 1960’s until late 1990’s.

The Service has already removed lead-contaminated soils and installed a geotextile cap on approximately 0.6 acres of Service property adjacent to the BSC. No remediation has occurred on the BSC site. We understand that the Service is in the process of evaluating potential additional remedial actions at the Site and seeks to understand whether further remediation would be successful without also initiating remediation on BSC property.

The goal of this study was to develop a conceptual level fate and transport analysis of lead contamination from the BSC.

1.1 Site Description

The Prime Hook NWR is located along Prime Hook Creek and the western shore of Delaware Bay in Sussex County, Delaware, approximately three miles northeast of the town of Milton. The 9,700-acre Refuge was established in 1963 and consists of approximately 7,400 acres of freshwater and tidal marsh, 1,000 acres of timber and brush, and 1,300 acres of grassland and croplands. Figure 1 shows the location, aerial photography and surface water features.

The BSC was founded in 1962 and maintained a trap shooting range on the property until 1998. The Club used five shooting stations, each built behind a small concrete trap house, which launched the clay targets into the air. Individual stations had an arrangement of five angled concrete walkways behind each trap house, with shooting stations and distances marked on the walkways.

The shooting range is an upland area that forms a drainage divide between two un-named wooded wetland tributaries to Prime Hook Creek. The highest land elevation on the upland is about 10 feet above sea level. Most of the wooded wetland areas are less than 5 feet in elevation. The Broadkiln River is tidally influenced through the Refuge, but Prime Hook Creek behind the flow-control structure (the dam) is not (Harding ESE, 2001).

The area downrange from the trap-shooting area consists of a forested wetland along a small, unnamed slough to Prime Hook Creek. The slough consists of a series of small, shallow ponds or pools that drain slowly to Prime Hook Creek. These pools are small (perhaps 3-12 feet wide) and no more than 3 feet deep. They are separated by saturated hummocks of mud, fallen trees, and organic matter. The overall size and shape of the slough varies seasonally, and it dries up completely on occasion. The slough is heavily forested, and is not used much by migratory waterfowl, which prefer open water. There are numerous small mammals, birds, amphibians, and insects that are common in and around the slough.

1.2 Summary of Previous Investigations

The trap-shooting range operated from 1962 until 1998, when a proposed land trade with the Service was initiated by the Club. As part of the proposed trade, the Service began a review of potential contamination of the site and, in 1998, the Service collected soil samples to determine the extent of lead shot deposition and lead concentrations in soil.

Results indicated that the surface soil in the shooting trajectory contained elevated concentrations of lead and high numbers of lead shotgun pellets. The Club discontinued shooting onto the Refuge property and in 2000 the Service funded the first year of a proposed 3-year cleanup project to delineate the magnitude and extent of soil contamination.

Since 2000, a number of studies have been conducted and interim remedial actions have been initiated, as summarized below.

1.2.1 Initial Service Response (2000)

In July of 2000, a synthetic geotextile cover membrane was placed over an approximately 0.5-acre area with the highest concentrations of lead pellets, with the primary objective of preventing waterfowl and other birds from ingesting the pellets. No soil was removed in the initial response. The textile material was initially anchored in place using 18-inch steel spikes. Following initial installation of the cover, 6-inch concrete blocks were placed on top of the cover would help it to resist wind damage.

A Preliminary Assessment (PA) was then completed in 2001 to determine the density of pellets on Service property. The PA concluded that lead shotgun pellets from the trap shooting range had accumulated on about 22 acres of Refuge property downrange from the Club (Harding ESE, 2001). The highest density of pellets on the Refuge was concentrated in a zone of approximately 26,200 square feet (0.6 acre) immediately adjacent to the Club property and directly downrange from the trap houses. As many as 57,868 lead pellets per square foot were counted in the center of this high-density area, which is referred to as the “drop zone” in previous reports.

1.2.2 USGS Groundwater Investigation (2001)

In 2000 and 2001, the USGS conducted a shallow groundwater investigation and installed 19 shallow wells on Service Property (Soeder and Miller, 2003). No wells were installed on BSC property. Well depths ranged from 2 to 8 feet below ground surface. Results from two sampling rounds indicated dissolved lead concentrations ranging from non-detect to 1,420 µg/L. Two surface water samples collected directly from the slough, several months apart and downgradient of the shooting range, contained 24 and 212 µg/L of lead, respectively.

1.2.3 Interim Remedial Action (2002)

An interim remedial action was initiated in 2002 to remove soil from the membrane-covered area. The geotextile was removed, and the top eight inches of soil and lead shot were excavated and transported off-site for disposal. The underlying 8-inch layer of soil was stabilized through the addition of Portland cement, and the area was backfilled with clean soil and re-seeded with a mixture of grasses (Shaw Environmental & Infrastructure, 2003). The cleanup was restricted to the upland area and a decision was made to monitor the wetland area for possible remediation.

1.2.4 Ecological Risk Assessment (2004)

In 2004, the Service completed an ecological risk assessment (ERA) for the site (Guy, et al, 2004). The ERA was intended to evaluate the potential threats to ecological receptors from direct exposure to lead contaminated soil and sediment as well, as from ingestion of lead shot. Toxicity tests, food chain accumulation models, and lead shot ingestion probability models were used to evaluate the risk to receptors that use the Site. This study recommended cleanup goals of 421 ppm Pb and 7-9 pellets/ft2 (Guy et al. 2004).

1.2.5 Groundwater and Surface Water Sampling (2007-2015)

In 2007, the Service began a monitoring program to assess lead concentrations in groundwater and surface water at the Site, and the potential for bioaccumulation and adverse effects in amphibians.

Groundwater was monitored periodically between 2007 and 2015 and showed infrequent detections of dissolved lead (at approximately 5 to 10 µg/L) in upland wells and no detections in upgradient and downgradient wells in the wetland (Pinkney and Callahan, 2014). However, surface water in the unnamed tributary had lead concentrations that exceed the chronic state ambient water quality standard (0.93 µg/L) as adjusted for water hardness of 40 mg/L calcium carbonate (CaCO3). For example, during the December 2015 sampling event, dissolved lead concentrations in the tributary ranged from 5.3 to 10,400 µg/L (Pinkney and Callahan, 2017).

1.2.6 Amphibian Investigations (2002-2011)

As part of the amphibian investigation, the wetlands were surveyed and adult green frogs (Lithobates clamitans), southern leopard frogs (L. sphenocephalus), bullfrogs (L.

catesbeianus), Northern cricket frogs (Acris crepitans), and Fowler’s toads (Anaxyrus fowleri) were observed in the wetlands, but tadpoles and/or egg masses were not observed (Sparling et al., 2006).

A laboratory sediment toxicity study was completed using southern leopard frog tadpoles that characterized various anatomical responses to elevated lead in sediment. It was reported that tissue residues increased with sediment concentration, but only the tadpoles exposed to concentrations of lead greater than 2,360 mg/kg were statistically elevated above all other groups including the control.

The Service (Krest et al., 2017) conducted an additional amphibian study to determine whether toxic effects and bioaccumulation of lead would occur under natural conditions in the wetland and if the results could be used along with literature and guidance to develop an updated PRG for possible sediment cleanup based on risk to amphibians. This study concluded that there were areas within the Refuge that showed risk to amphibian from exposure to lead. Based on the analysis, the Service concluded that the inconsistent sediment concentrations at the Warm Spot precluded the use of those results for developing a PRG. Instead, the Service proposed 312 ppm lead as a PRG for sediments in Prime Hook wetland locations where amphibians breed, in order to reduce the risk of serious malformations. T. Based on the CHF-funded laboratory study and Environmental Protection Agency guidance, we propose 312 ppm lead as a PRG for sediments in Prime Hook wetland locations where amphibians breed, to reduce the risk of serious malformations.

1.2.7 Supplemental Site Investigation (2017)

In 2017, a Supplemental Site Investigation (SSI) was completed by the Service at the Site to reduce remaining data gaps regarding the nature and extent of contaminants in soil and sediment in the forested wetland portion of the Site (Geosyntec, 2017a). The SSI confirmed that lead was the primary contaminant of potential concern (COPC) for the Site and that, in general, soil and sediment within the shot fall zone contain concentrations of lead above the PRGs. The SSI included additional soil and shot-count sampling on Service property to verify and refine historical sampling. The SSI showed that the highest concentrations of lead shot (greater than 10 pellets per square foot) were distributed across approximately 4 acres of Service property immediately adjacent to the BSC property.

In addition to the SSI, an Engineering Evaluation and Cost Analysis (EE/CA) was completed for the Site (Geosyntec, 2017b) using both historical data from the Site as well as data acquired for the SSI. The EE/CA developed cost scenarios for several remedial alternatives for soil at the Site including, no action, excavation with off-site disposal, excavation, treatment and off-site disposal, and soil and sediment capping.

1.2.8 DNREC Site Investigation of BSC (2017)

The Delaware Department of Natural Resources and Environmental Control (DNREC) prepared a Site Inspection (SI) Report for the Broadkiln Sportsman’s Club Site in 2017.

The purpose of the SI was to determine if historical and/or current site activities have contaminated groundwater, surface water, sediment, and/or soil onsite and into the surrounding area. The SI included a risk assessment. Based on the conclusions of the SI and the risk assessment, EPA and DNREC officials will determine whether the Site should be investigated further and/or remediated or obtain a “No Further Action” designation under the Federal Superfund and/or State Site Investigation & Restoration Section (SIRS) Programs.

A total of 58 soil borings were drilled across the BSC property. Forty (40) borings were drilled in the trap shooting field, 13 at the pistol range, and five (5) for monitor wells. A total of 135 soil samples were collected; 100 from the trap field, ten (10) from the monitor wells and 25 from the pistol range. Soil borings extracted by direct-push drilling in the trap field were advanced to approximately 2 feet bgs and soil samples were collected from the 0 to 6” (s), 6” to 12” (m), and 2” to 24” (d) intervals. Hand augured borings were advanced to approximately 2’ bgs and soil samples were collected from the 0 to 1’ and 1’ to 2’ intervals.

In general, the majority of the soil samples exceeding 1,000 mg/kg lead were located on the eastern half of the BSC property, with the highest soil concentrations adjacent to the Service property line. The highest observed concentration was 128,000 mg/kg about 120 feet east of the Refuge boundary. Two other samples adjacent to the Refuge boundary showed concentrations of 59,600 mg/kg and 12,100 mg/kg. These two samples were collected directly adjacent to Refuge boundary, but north of the Refuge area that was excavated and capped.

Only one of the monitoring wells (BSC-47) showed elevated lead concentrations above 10 ug/L. However, BSC-47 was the only well located within 200 feet of the refuge boundary.

1.3 Study Objectives and Scope

The objective of Geosyntec’s analysis is to:

1. Evaluate the fate and transport of lead from the Broadkiln Sportsman Club (BSC) to the Refuge property

2. Assess whether further remediation on Refuge property would be successful without remediation of the BSC property; and

3. Identify options for preventing contamination from entering the Refuge from the

BSC.

Geosyntec’s approach was to develop conceptual models based on available existing data to provide a framework for contaminant transport pathways from soils at the BSC to the Refuge. There are three components to the approach:

1. An estimate of lead pellet mass on both Service and BSC property was determined from previous soil investigations.

2. A geochemical conceptual model was developed to describes geochemical mobility and speciation of lead at the site. Data used to develop the geochemical model included observed soil chemistry (pH/eh), metals concentrations in soil, and groundwater chemistry.

3. A hydrogeologic conceptual model was developed to describes groundwater flow directions, travel time and attenuation factors applicable for metals in groundwater. Data used to develop the hydrogeologic model included observed groundwater elevations, location of surface water and topographic features, monitoring well details at or near the Site, and water quality data for both groundwater.

4. A surface water conceptual model was developed to describes general flow directions and a first-order estimate of lead mass transported into the Refuge.

2. LEAD PELLET MASS

2.1 Methodology

Lead pellet mass represents the primary source of lead to surface water, soil and groundwater. Lead pellet density (“shot count”), as well as soil lead concentration, was mapped as part of the SSI. However, this mapping was only conducted on Service property. The DNREC investigation mapped soil lead concentration, but did not map lead pellet density. The pellet density contour intervals mapped as part of the SSI were extended into the BSC property based on a hand drawn approximation using DNREC soil concentrations to create fully closed contours. The estimate was guided by soil sampling results on BSC property where lead concentrations greater than 400 mg/kg were identified (Figure 4 of the DNREC Report).

Lead pellet density was converted to lead mass based on estimates for various types of lead shot that would typically have been used at the shooting range. Depending on the type of shot used, a single lead pellet is estimated to contain between 0.080 and 0.048 grams of lead (Table 1).

2.2 Results

Figure 2 shows the estimated lead shot density on both Service and BSC property. In general, soil concentrations greater than 120,000 mg/kg correspond to shot count density greater than 10,000 pellets per square foot. Soil concentrations between 2,000 and 120,000 mg/kg correspond to shot count density between 1,000 and 10,000 pellets per square foot. Soil concentrations between 100 and 2,000 mg/kg correspond to shot count density between 100 and 1,000 pellets per square foot.

Based on estimated lead mass per pellet shown on Table 1, the estimated total mass of lead pellets on both Service and BSC property (prior to the initial 2001 remedial action) is shown on Table 2.

3. GEOCHEMICAL ANALYSIS

3.1 Methodology

Lead mobility in soil and groundwater is dependent upon the individual lead species in the source and downgradient areas. In order to provide a conceptual understanding of the likelihood of Pb mobility in Site groundwater, the speciation of Pb in groundwater at the Site was estimated in the software program The Geochemist’s Workbench [GWB] (Release 12; Aqueous Solutions, LLC). Average observed concentrations in Site groundwater in 2000-2001 (Soeder and Miller, 2002) were input to GWB in order to model the expected equilibrium concentrations of Pb species. The model does not account for the formation of Pb species that are adsorbed to naturally occurring minerals in soil, such as iron oxides. Table 3 shows the input geochemical parameters. GWB calculated the expected concentrations of Pb species, both aqueous and solid phase, as a function of reduction-oxidation potential (Eh) and pH.

3.2 Results

Figure 3 shows the predicted dominant lead species (i.e., greater than 50% of total lead) under equilibrium conditions as a function of Eh and pH. Under reducing conditions (Eh approaching negative values), formation of the solid phase Pb-sulfide mineral galena is predicted. At Eh greater than zero, the dominant Pb species are aqueous: Pb+2 and Pb-carbonate. The predicted Pb-carbonate species, PbCO3, is a neutrally charged aqueous complex, not a lead-carbonate solid. The green box on Figure 3 is the range of expected Eh and pH in site groundwater : pH is approximately 5 to 6, and Eh is mildly oxic based on detections of dissolved oxygen (Eh estimated to be 0.25 to 0.75 V).

Equilibrium modeling of Pb speciation indicates that Pb remains dissolved in groundwater under the geochemical conditions previously observed in Site groundwater.

The main Pb species is expected to be positively charged at pH values less than 4-6, and neutrally charged at pH values greater than 4-6. The pH of Site groundwater has been previously measured at 5-6, which means that pH is a key geochemical parameter to predict Pb mobility at the Site.

The modeling does not simulate the formation or dissolution of Pb species that may be kinetically limited. The charge of the surface of the aquifer solids would be expected to change from positive to negative across a similar pH range, which could result in the formation of adsorbed solid species, which are not included in the geochemical modeling performed here. In other words, the aquifer solids could be positively charged at low pH (not sorbing Pb) and negatively charged at higher pH (sorbing more Pb). Therefore, a simple retardation constant may not be effective to predict mobility of Pb in Site groundwater.

4. HYDROGEOLOGICS ANALYSIS

4.1 Methodology

The groundwater flow analysis was performed in several steps. First, groundwater elevations were evaluated to determine likely flow paths and hydraulic gradient along those flowpaths. Figure 4 shows the estimated groundwater potentiometric surface and groundwater flow directions based on available data. Figure 5 is based on available groundwater level monitoring collected by the USGS (2001), Service (2007-2017) and DNREC (2017). Groundwater flows outward from an apparent groundwater divide that is located on the upland area of the firing range. Groundwater flow directions are likely radial and fairly uniform, with flow toward the unnamed surface water features east of the firing range. Using these groundwater contours, a groundwater gradient of 0.0013 ft/ft is estimated. Previous investigations by the USGS indicate a hydraulic conductivity of between 5 and 10 feet per day (Soeder and Miller, 2001).

Groundwater transport along a flowpath was analyzed using a simplified 1-dimensional transport equation from Domenico & Schwartz (1990). This equation requires a variety of input parameters, and can generate plots of concentration versus time (at a specified distance) or concentration versus distance (at a specified time) from a source area with a specified initial concentration. A base case transport calculation was made using input parameters determined from site conditions or estimated using typical default parameters for wetland-type soils and groundwater. Guidance from the New Jersey Department of Environmental Protection (NJDEP, 2000) was used for the partitioning coefficient (Koc) and fraction of organic carbon (foc) inputs. The initial model input parameters for the base case were as follows:

• hydraulic conductivity (k) : 5 ft/day (based on USGS report)

• hydraulic gradient (i) : 0.0013 (based on observed groundwater levels)

• effective porosity (n) : 0.25 (estimated)

• dispersivity (d) : 300 (estimated). Typical range is 200 to 600

• bulk density (p) : 2.6 g/mL

• Pb partitioning coefficient (Koc) : 900 mL/g (NJDEP, 2000)

• Fraction of organic carbon (foc) : 0.002 (NJDEP, 2000)

As described above, because Pb speciation and transport geochemistry may vary depending on pH and other soil properties, a simple retardation factor based on a partitioning coefficient may not accurately predict transport distances and concentrations of Pb in Site groundwater. Therefore, sensitivity analysis was performed by varying input parameters to develop a range of potential lead transport times and distances along groundwater flowpaths.

4.1.1 Partitioning Coefficient (Kd) and Transport Velocity

The partitioning (or distribution) coefficient (Kd) is an important parameter that describes adsorption/retardation behavior of contaminants in groundwater. It is often correlated with organic carbon content and can be calculated as the product of Koc*foc. Koc varies with the contaminant, while foc is a soil property. For example, Koc for lead is typically higher than Koc for iron or cadmium. Default parameters for lead Koc and foc in wetland soils were taken from the New Jersey Department of Environmental Protection (NJDEP, 2000) at 900 mL/g for Pb Koc and 0.002 for soil foc. This results in a default Kd of 1.8 mL/g, which was used for initial base case. As described in the geochemical conceptual model, lead is also affected by eh/pH relationships that are not captured with a constant one-dimensional distribution coefficient. The sensitivity analysis brackets a range of transport conditions that could represent this eh/Ph dependence.

Groundwater transport velocity was calculated using the hydraulic conductivity, hydraulic gradient, and effective porosity. The retardation factor was based on the bulk density and partitioning coefficient (as described above). The base case transport velocity was 0.001 ft/day.

4.1.2 Initial concentration (C0)

The initial concentration (C0) for the source area over the history of the site is not known.

However, it is related to the mass of soluble Pb that leaches to groundwater. Figure 5 shows the various measured concentrations of Pb at the site since 1999. Prior to 2001, the USGS measured very high concentrations (up to 1,400 ug/L) in several shallow monitoring wells. The average of all samples collected in 2001 was 656 ug/L.

After the initial remedial action that removed and capped (with geotextile) a portion of the soil and lead pellets on Service property, lead concentrations in the immediate vicinity of the remedy decreased to below 10 ug/L. These observed Pb concentrations were used to calibrate the modeling results in a rough fashion. In other words, the initial concentration (C0), which is a required and sensitive parameter in the model calculation, was not allowed to vary outside these ranges.

4.2 Results

The results of the modeling are shown on Figure 6. The model predicts Pb concentrations as a function of distance from a fixed source location in one dimension (e.g. along a flowpath). The presumed source location is the area delineated on Figure 2, which is an area of about 9.4 acres and is two-dimensional. The centroid of this area was used as X0 in comparing observed concentrations to model predictions at various distances from the source.

Figure 6a shows the predicted Pb concentration as a function of time (for a specified distance from the source of 150 feet), and Figure 6b shows the predicted Pb concentration as a function of distance (for a specified time of 50 years after source release). The 50-year time factor is representative of the time since lead began accumulating at the site (beginning in 1960). The 150-foot was selected as a representative average distance from the centroid of the source area to well DG-8 and DG-9. This is also a representative distance from the edge of the highest soil concentrations on BSC property to the boundary with the Refuge.

There are three scenarios shown on Figure 6:

• The “base case” represents the default parameters described above, with a concentration in the source areas of 2,000 ug/L.

• The “high carbon” scenario represents a higher organic carbon content (foc) in the soil than the NJDEP default. A higher foc would result in a high Kd which would result in a higher retardation of Pb as a function of distance. This scenario also represents a higher Pb adsorption because of soil pH/Eh conditions, as discussed in Section 3. A source area concentration of 4,000 ug/L is needed to match the observed data at a distance of 100 to 150 feet.

• The “high transport” scenario represents higher advective transport velocity (a combination of hydraulic gradient, conductivity, and dispersivity). This scenario could also represent a lower Pb adsorption (higher transport) because of soil pH/Eh conditions, as discussed in Section 3. A source area concentration of 1,200 ug/L is needed to match the observed data at a distance of 100 to 150 feet.

Figure 6 shows that:

• The observed lead concentrations of 650 ug/L at USGS wells installed in 2001 can be reproduced using a source area concentration of between 3 and 12 mg/L (3,000 to 12,000 ug/L) over a time span of 50 years. Therefore, the range of transport properties and initial concentrations used in the model appear to be representative of site conditions.

• Pb concentrations are expected to decrease steeply below 2 ug/L beyond about 300 feet from a source area. This occurs for all sensitivity scenarios, and is consistent with observed reductions of Pb concentration after the initial remedial action (soil removal) on Service property. The distance of 300 feet should be considered from the edge of the source area depicted on Figure 2 (not the centroid).

An effective source area concentration of 3 to 12 mg/L corresponds to a total loading rate to groundwater of between 1,950 and 7,800 kg of Pb over a 50 year period. This calculation is based on a precipitation rate of 45 inches per year, with 70% run-off (30% infiltration). Based on the estimated total mass of lead pellet at the site (56,930 Kg combined on both BSC and Refuge property), this suggests that between 3.4% and 13.7% of the total mass of Pb in shot pellets deposited on the site since 1960 has leached to groundwater. This estimate is for groundwater only, and does not include mass that may have been transported via overland flow. Surface water runoff is discussed in Section 4.3 below.

A composite map showing estimated lead pellet density, groundwater flow direction, and Pb source area is shown on Figure 7. Note that a considerable area of un-remediated soil is likely contributing to groundwater that is flowing in a more northerly direction compared to the area that has already been remediated. There is limited groundwater flow data to confirm groundwater flow direction in this area.

4.3 Surface Water Transport

Surface water quality data (temperature, dissolved oxygen, pH and conductivity) and sediment lead concentrations are reported between 2002 and 2011 (Pinkney, 2017) at a reference location (known as “Fred’s Pond”) and two other locations (designated as “Hot Spot” and “Warm Spot”). These locations are between 35 and 100 meters southeast of the BSC property line. Unpublished surface water lead concentrations in an unnamed tributary are reportedly between 5.3 and 10,400 ug/L for a sampling event in December 2015 (Pinkney, 2017). Detailed modeling of lead transport via surface water is not possible using these data since the hydrology at the site (i.e. channel definition, flow rates, and connectivity to groundwater) is not well defined and water quality data is sparse.

However, the groundwater analysis provides a calibrated estimate of total lead mass that has leached to groundwater. Total lead that has entered Refuge property by a surface water pathway was estimated using a similar mass balance calculation based on the calibrated groundwater analysis. This calculation is based on the following assumptions:

• The processes that mobilize Pb to leach into groundwater are similar to the processes that mobilize Pb to surface water. This is described in Section 3, which showed that geochemical mobility of Pb is sensitive to redox and pH conditions, which can be assumed to be similar for both surface water and shallow groundwater.

• The estimated total Pb mass from pellets deposited to the source area from the BSC property since 1965 is 56,930 kg. This is described in Section 2 and based on pellet count data from 2017 (Geosyntec, 2017) and soil concentrations from

2017 (DNREC, 2017).

• The estimated total Pb mass that has entered groundwater from the source area since between 1965 and 2015 (50 years) is between 1,950 and 7,800 kg. This is described in Section 4 and based on a simple one-dimensional groundwater transport model.

• The estimated volume of surface water that has contacted the source area between 1965 and 2015 (50 years) is 358,370,951 gallons. This is based on an average of 44 inches of precipitation per year falling on a source area of 9.43 acres, with a run-off coefficient of 70% (30.8 inches of run-off).

• The estimated volume of groundwater that has contacted the source area between 1965 and 2015 (50 years) is 153,587,551 gallons. This is based on an average of 40 inches of precipitation per year falling on a source area of 9.43 acres, with an infiltration coefficient of 30% (13.2 inches of infiltration).

Table 4a and 4b summarize these calculations. Using these assumptions, the dissolution rate of Pb to groundwater can be calculated in terms of kilograms of lead dissolved per liter of infiltration. This rate can then be applied to surface water volumes to determine total dissolution of lead to both surface and groundwater since 1960. As described above, this assumes that the processes that mobilize Pb into groundwater are similar to those that mobilize Pb into surface water.

Assuming that between 1,950 and 7,800 kg of Pb have leached to 153,587,551 gallons of groundwater over the past 50 years, then the effective mass dissolution rate of lead pellets is between 9.0 x 10-5 and 3.6 x 10-4 kg per gallon of water. Applying these same dissolution rates to 358,370,951 gallons of surface water results in between 4,550 and 18,200 kg of Pb lead that has dissolved into surface run-off. Because the dissolution rate is proportional to water volume, the mass dissolution via surface water is higher.

While there are uncertainties associated with this estimate, the long-term mean concentration of lead in surface water that has contacted the source area would be between 3,300 and 13,500 ug/L (based on the mass dissolution estimate described above) is, which is similar to the range of unpublished 2015 concentrations reported by Pinkney (2017).

There is some potential for particulate transport and plant uptake that would reduce the total lead mass entering Refuge lands, but there are insufficient data to assess these additional sinks for lead that has entered the environment from the BSC property. Since the estimate falls within the estimated total mass of lead based on pellet counts and the range of observed lead in surface water, it is considered a reasonable order-of-magnitude estimate of mass balance via a surface water pathway. Further analysis of verification cannot be attempted without additional data.

4.4 Air Transport

Airborne transport of lead is often associated with active disturbances of lead contaminated soils that release particulates, such as excavation, construction, mining, or natural erosion. Factors such as high vegetative cover, wet/humid climate, and low topography/slope, which predominate at the site, would tend to minimize the potential for airborne transport of particulates with high concentrations of lead. For these reasons, the potential for airborne transport of Pb originating from the BSC property is considered minimal and airborne transport has not been evaluated quantitatively.

4.5 Preliminary Evaluation of Pb Immobilization

There are remedial alternatives that would passively immobilizes Pb in soil or groundwater entering The Service property from the BSC property. Application of phosphate has been demonstrated to immobilize lead in soils (Miretzky and Fernandez- Cirell, 2009). This technology could lessen the reliance on clean-up at the BSC property and protect downgradient ecological receptors without a major excavation or disruption to the Refuge. Lead forms an insoluble mineral when combined with phosphate and calcium. Phosphate amendments are an accepted remedial technology to immobilize lead in situ. Table 5 provides an overview and screening level analysis of this technology. A batch-scale laboratory test would be necessary to evaluate Pb immobilization and compatibility with a phosphate amendment that could be backfilled in a trench or injected as a slurry with small-footprint direct-push injection rigs. Site-specific groundwater and soil materials would be needed to conduct this test. As described in this report, field-based eh/pH relationships and organic carbon conditions will be needed to determine the ability of Pb to sorb to phosphate under natural environmental conditions. In addition, consideration of phosphate leaching to downgradient surface water will be a design constraint in order to avoid downgradient eutrophication. If the batch-scale test demonstrates Pb immobilization and manageable amounts of phosphate addition to groundwater, then a field-scale pilot test could be conducted along the Site property boundary, north of the area that has already been remediated (see Figure 7).

5. CONCLUSIONS

Based on our analysis, Geosyntec has the following conclusions:

• The total remaining mass of Pb on site is estimated to be approximately 36,500 kg (7,400 kg on Service property and 18,500 kg on BSC property).

• Pb has already impacted groundwater and surface water downgradient of the firing range, resulting in documented ecological impacts to amphibians.

• Geochemical analysis indicates that Pb leached from pellets in the source area will be mobile in waters similar to those observed at the site. The dynamics of Pb transport are sensitive to eh/pH conditions, which are not well documented on the site, but likely create a mobile environment for Pb transport.

• Based on soil sampling and a pellet count survey, the source area for Pb at the site is estimated to be about 10 acres in size, straddling the property boundary between the Refuge and the BSC property.

• Modeling of Pb transport in groundwater using a simple one-dimensional distribution coefficient approach indicates that between 3% and 13% of the total mass of Pb pellets deposited on the site since 1960 has leached to groundwater.

An additional 8% to 30% of the total mass of Pb pellets may have leached to surface water.

• Based on a range of transport properties that appear to be representative of site conditions, Pb concentrations are expected to decrease below 0.2 ug/L beyond about 300 feet from any source area. This is consistent with observed reductions of Pb concentration after the initial remedial action (soil removal) on Service property.

• The most likely groundwater pathway from areas of the BSC property with the highest lead pellet density to Service property is a more northerly flowpath compared to the area that has already been remediated (see Figure 7). This area currently has limited groundwater monitoring data.

• The potential for airborne transport of Pb to Service property is very low since there is no active disturbance of Pb-contaminated soils and the property is heavily vegetated with humid conditions.

• Removal of the remaining mass of Pb on Service property would reduce, but not eliminate, continued leaching and long-term accumulation of Pb in groundwater surface water and sediment within the Refuge.

• If the mass of Pb on the BSC property is not removed, it will remain a long-term source of Pb to groundwater and surface water on Service property.

6. REFERENCES

Domenico, P.A. and Schwartz, F.W., 1990. Physical and Chemical Hydrology. Wiley, New York

Geosyntec, 2017a. Supplemental Site Investigation for Wetlands Adjacent to the Former Trap Range, Prime Hook National Wildlife Refuge, Milton, DE. November.

Geosyntec, 2017b. Engineering Evaluation / Cost Analysis for Wetlands Adjacent to the Former Trap Range, Prime Hook National Wildlife Refuge, Milton, DE. November.

Guy, C.P., M. Huston, S. Krest, and D.R. Murphy, 2004. Ecological risk assessment for the Prime Hook National Wildlife Refuge lead shot site, Milton, Delaware. CBFO-C03-

04. U.S. Fish and Wildlife Service, Annapolis, MD.

Harding ESE, 2001. Preliminary Assessment – Site Investigation, Shotgun Range, Prime Hook National Wildlife Refuge, Milton, Delaware.

Krest, S., A.E. Pinkney, and M.A. Rutter, 2017. Use of an In Situ Tadpole Enclosure Study for Developing a Preliminary Remediation Goal for Lead in Wetland Sediments at Prime Hook National Wildlife Refuge, Milton, Delaware. CBFO-Cl6-0l. U.S. Fish and Wildlife Service, Chesapeake Bay Field Office, Annapolis, MD.

Miretzky, P. and Fernandez-Cirelli, A., 2009. Phosphates for Pb immobilization in soils:

a review. Environmental Chemistry Letters, 6(3):121-133.

New Jersey Department of Environmental Protection, 2000. Chemical Properties for Calculating Impact to Groundwater Soil Remediation Standards.

Pinkney, A.E. and C.R. Callahan, 2014. Ground Water and surface water monitoring of the Prime Hook Lead Shot Site, Prime Hook National Wildlife Refuge, Milton, Delaware (October 2007-December 2014). Final Report. CBFO-C14-0l. U.S. Fish and Wildlife Service, Chesapeake Bay Field Office, Annapolis, MD.

Pinkney, A.E. and C.R. Callahan, 2017. Ground Water and surface water monitoring of the Prime Hook Lead Shot Site, Prime Hook National Wildlife Refuge, Milton, Delaware (October 2007-December 2015). Final Report. CBFO-C16-04. U.S. Fish and Wildlife Service, Chesapeake Bay Field Office, Annapolis, MD.

Shaw Environmental & Infrastructure, 2003. Prime Hook National Wildlife Refuge, Milton, DE. Lead shot removal Phase II. Final report. Shaw Environmental & Infrastructure, Monroeville, PA.

Soeder, D., and C. Miller, 2003. Ground-water contamination from lead shot at Prime Hook National Wildlife Refuge, Sussex County, Delaware. USGS Water Resources Investigation Report 02-4282.

Westminster Ammunition Guide, 2004. https://www.chuckhawks.com/shot_info.htm https://www.chuckhawks.com/shot_info.htm

TABLES

Table 1 : Estimated Lead Mass per Pellet

Shot Type Avg Pellet

Count1 Lead Mass per pellet2 (grams)

#7-1/2 Shot 350 0.080 #8 Shot 410 0.068

#8-1/2 Shot 497 0.056 #9 Shot 585 0.048

Average 461 0.063 1 Winchester Ammunition 2004 Product Guide 2 Assumes 97% lead, 3% antimony per pellet

Table 2 : Estimated Lead Pellet Mass on Service and BSC Property

10,000 1,000 - 10,000 100 - 1,000 <10 TOTAL AREA

USFWS (sq ft) 39,204 79,859 49,797 50,965 219,826

BSC (sq ft) 33,585 88,850 31,289 37,113 190,836 Total 72,789 168,709 81,086 88,078 410,662

10,000 1,000 100 10 TOTAL MASS

USFWS (kg) 24,648 5,021 313 32 30,013

BSC (kg) 21,115 5,586 197 23 26,921 Total 45,762 10,607 510 55 56,934

10,000 1,000 100 10 TOTAL MASS

USFWS (kg) 4,930 4,268 313 32 9,542

BSC (kg) 21,115 5,586 197 23 26,921 Total 26,044 9,854 510 55 36,463

Assumes average lead mass of 0.063 grams/pellet

Current Mass (kg)

Contour Area (sq ft)

Mass (kg) prior to USFWS 2001 interim removal action

Shot Density (Pellets/sq. ft.)

Pellet Mass (kg)

Pellet Mass (kg)

Table 3: Summary of geochemical parameter concentrations used in geochemical modeling.

Component Concentration (mg/L)

Ca+2 5.89 Mg+2 4.25 Na+ 4.02 K+ 3.07

SO4

-2 11.2

Cl- 8.77

HCO3

-1 16 SiO2 (aq) 7.52

Pb+2 .25 Eh* .25-.75 pH 5-7

* Eh was estimated based on dissolved oxygen values of 1.3 to 3.1 mg/L.

Table 4a : Source Area Water Volumes

Table 4b : Source Area Pb Mass Balance

Precipitation 40 inches Source Area size 9.43 Acres

Groundwater Iniltration % 30% Total Groundwater Infiltration (50 years) 153,587,551 gallons

Surface Water Runoff % 70% Total Surface Runoff (50 years) 358,370,951 gallons

Groundwater

Surface Water

Source Area Properties

Total Initial Mass USFWS BSC Total Mass based on Pellet Count (kg) 30,013 26,921

Total

Min Max Groundwater-Leached Mass (kg)1 1,950 7,800 GW Leached Mass (%) 3.4% 13.7% GW Volume (gal) 153,587,551 Pb Leach Rate (per gallon) 1.3E-05 5.1E-05

Pb Leach Rate (per gallon) 1.3E-05 5.1E-05 SW Volume (gal) 358,370,951 Surface Water-Leached Mass (kg) 4,550 18,200 SW Leached Mass (%) 8.0% 32.0%

Total Mass Leached 6,500 26,000 % Initial Pellet Mass 11% 46%

1 Based on Groundwater model

Total Mass Leached

Groundwater

Surface Water

56,934

FIGURES

1212 12

P:\CAD_GIS\Projects\Prime Hook\GIS\Projects\Figure 1 Site Location and Surface Water.mxd 1/7/2019 10:34:13 AM

Legend Geotextile Location Range Features Stream Contour (2 ft NAVD88) Estimated Flow Path of Surface Water

Water Body Wetland

Notes:

1. Elevation contours from USGS National Elevation Dataset (Ft NAVD88).

0 500 1,000250 Feet

United States Fish and Wildlife Service Prime Hook National Wildlife Refuge, Delaware

Site Location Map and Surface Water Features

Figure

Seattle, WA January 2019

LewesLewes

MilfordMilford

Prime Hook

Slaughter Neck

Primehook Neck

Sussex

Site Location

59600

128000

12100

26110

14076.751.6

7.9 5922.4

37190

4.5

P:\CAD_GIS\Projects\Prime Hook\GIS\Projects\Figure 2 Shot DensityV2.mxd 5/8/2019 10:20:23 AM

Legend

Shot Density Contours

10000

DNREC Soil Sample Location > 11000 mg/Kg

DNREC Soil Sample Location <

11000 mg/Kg

USFWS Property Soil Concentrations

<1000 ug/L

Previous Soil Sample Location (Geosyntec 2017)

Un-named Tributary

Approximate Property Line

Concentration Greater than 11000 mg/kg (Source Area)

Geotextile Location

Range Features Sources:

1. Harding ESE1 2001

2. Geosyntec1 2017

3. DNREC1 2017

0 160

Feet

³ United States Fish and Wildlife Service Prime Hook National Wildlife Refuge1 Delaware

Estimated Shot Density

Seattle1 WA May 2019 in te rn a l i nf o : p a th

, d a te re vi se d , a ut ho r

Predicted Geochemcial Stability Diagram USFWS Prime Hook National Wildlife Refuge

Seattle, Washington March 2019

UG

2.63

UG-1

2.63

UG-2

2.36

DG-10

2.22

DZ

DZ-1

2.36

DZ-0

2.41

DZ-3

2.30

DG-9

2.18

DG-8

2.22

DG-8S

2.04

DG-5

2.02

DZ-2

2.33

DG-1

1.96

DG-4

2.01

DG-6

2.23

DG-2

1.01

DG-7

1.99

DG-3

2.71

BSC GW-50

4.00

BSC GW-47

4.10*

BSC GW-48

3.80

BSC GW-49

4.20

BSC GW-51

3.70

DZ4

DZ5

DZ6

UG1

0.00 piezometer*

SW-01

4.0 3.0

2.6

2.4 2.2

2.0

P:\CAD_GIS\Projects\Prime Hook\GIS\Projects\Figure 4 GW_Elevation.mxd 1/9/2019 11:04:31 AM

Legend Well Locations Groundwater Flow Direction Approximate Location of Groundwater Divide Estimated Groundwater Contour Property Line

Geotextile Location Range Features Stream Water Body Wetland

Note:

1. * = Groundwater level not used in contouring.

2. Well locations and names from USGS (2001), USFWS (2017), and DNREC (2017).

0 100 20050 Feet

United States Fish and Wildlife Service Prime Hook National Wildlife Refuge, Delaware

Groundwater Elevation…