Horizontal Discretization

A principal benefit of using MODFLOW 6 is the ability to locally refine the horizontal grid in areas where greater prediction accuracy is required, without the need to extend fine grid spacings to the edge of the model grid or use slowly converging nested grid refinement algorithms. Because MODFLOW 6 does not support use of both nested and quadtree grids in the same model, quadtree grid refinement was used. These grids, which represent tree data structures for which each parent cell has exactly four child cells, were used to refine the parent grid in areas where preservation of horizontal hydraulic gradients is critical to accurately predicting advective movement of dissolved PFAS. Using a quadtree level of 4, and a grid smoothing operation that adjusts cell size so that no cell shares an interface with more than two adjacent cells (Langevin and others, 2017), quadtree grid refinement of the parent grid produced square cells ranging in size from 125 ft x 125 ft in refined portions of the grid to 1,000 ft x 1,000 ft in areas far from locations where grid refinement occurs, with 250 ft x 250 ft and 500 ft x 500 ft intermediate cell sizes (fig. 5).

Quadtree grids were imposed both inside rectilinear and along curvilinear features (fig. 5). These grids were used to refine the parent grid within polygons representing the five PFAS reconnaissance areas and within multiple known AFFF source areas. The reconnaissance areas represent neighborhoods outside but adjacent to JBMDL boundaries on domestic wells and are considered to be at elevated risk for contamination. Source areas account for locations where known or suspected historical use of AFFF may have released PFAS into the environment. In addition to allowing for local increased spatial resolution in the critical reconnaissance and source areas, MODFLOW 6 enabled creation of quadtree grids along streams to more accurately account for hydraulic gradients adjacent to base-flow-dominated streams in the Coastal Plain and enhance the reliability of predicted base flows.

Vertical Discretization

Following delineation of the lateral no-flow boundary and local horizontal refinement of the three-dimensional grid, vertical layering was introduced by incorporating the hydrostratigraphic framework into the model (fig. 5). Discretization along the vertical dimension was achieved by importing top and bottom elevations of four layers based on the structure of hydrostratigraphic units defined in Fiore (2020). The Kirkwood-Cohansey aquifer system, Manasquan-Shark River confining unit, Piney Point aquifer, Vincentown aquifer, updip, shallow portions of the Navesink-Hornerstown confining unit, and undivided shallow aquifers and confining units lower in stratigraphy were represented across the four model layers. In addition, an upland gravel surficial unit that likely affects groundwater flow in the western part of JBMDL was also represented in the model. For further information about specific layers, see appendix 1.

Introduction of layer top and bottom elevations provided the basis for assigning depths to drain cells and apportioning supply-well withdrawals to layers intersected by specified screened intervals, making it possible to accurately specify internal boundary conditions. The four aquifer layers, which can further be refined using sublayering to better resolve steep vertical hydraulic gradients, impose known hydrogeologic structure on the model and allow aquifer properties unique to each layer, such as hydraulic conductivity (K), porosity, and chemical retardation factors, to be specified. As summarized in table 2, the four layers contain a total of 15 K zones, within which K and other aquifer properties are treated as uniform. Details of the K distribution are discussed later in this report.

In addition to varying in thickness across the model area, layers taper out to zero thickness at sedimentary sequence boundaries or geologic unconformities that vary in location with each layer. These boundaries cause discontinuities in the spatial extent of layers, generally referred to as pinchouts. Pinchouts are pervasive throughout the Coastal Plain aquifers that underlie JBMDL. In addition to allowing for local grid refinement, MODFLOW 6 supports use of unstructured grids to incorporate geologic pinchouts. Unlike the structured grid archetype supported by earlier versions of MODFLOW, which requires a fixed number of finite-difference connections for every model cell, an unstructured grid has an arbitrary number of connections to neighboring cells. One benefit of allowing for an arbitrary number of connections to each cell is that the model can include no lateral connections along the edges of pinched-out geologic layers. Although this produces an unbanded, non-symmetric solution matrix that requires a sparse matrix solver, it also makes it possible to design model grids with truncated layers (Langevin and others, 2017). In this MODFLOW 6 model of JBMDL and vicinity, cells in which a geologic layer pinches out were assigned top and bottom elevations of the layer equal to one another to obtain a layer thickness of zero.

Base Flow

Owing to the small topographic relief and low fluvial energy characterizing the Coastal Plain, groundwater contributions to streams are a dominant component of the groundwater budget and represent the single largest sink for groundwater (Charles and Nicholson, 2012).

Streams and wetlands cells in the model were delineated by the NHD-Plus dataset (U.S. Environmental Protection Agency and U.S. Geological Survey, 2012) and base flows were simulated using the MODFLOW 6 drain (DRN) package, which removes groundwater when the elevation of the water table exceeds the elevation of the streambed, conveying it instantaneously out of the groundwater flow system in the form of base-flow contributions to streams at a rate proportional to that elevation difference. The DRN package was applied to model cells along streams and within stream-connected wetlands to simulate the effect of base-flow losses in the base-flow-dominated streams that drain the low-relief Coastal Plain environment.

The DRN package requires inputs of streambed elevation and drain conductance. For streams, the streambed elevation was set to 90 percent of the thickness of the cell in layer 1 above the bottom of layer 1. Wetlands had a streambed elevation equal to the top of the model. Drain conductance is a measure of how easily groundwater seeps through streambed sediments along the sides and bottoms of the streams and hydraulically connected wetlands, where it instantaneously exits the model as base flow. Because drain conductance depends on a wide array of factors, including stream channel shape and bottom sediment thickness, its value is generally one of the most uncertain to specify. A constant conductance value of 5 square feet per day was assigned to streams, and a constant value of 2 square feet per day was assigned to wetlands. Drain conductances can also be estimated by calibration to more effectively reproduce base-flow losses, but these parameters were not chosen as calibration parameters in this model.

Fiore and others (2021) identified losing streams near JBMDL that are associated with dams and weir structures. These losing conditions are not simulated and are assumed to be local to that stream reach, with negligible effect on the regional system. However, low-flow discharge measurements have the potential to be affected by such conditions if a similarly engineered structure exists upstream, which may underestimate base-flow observations for that catchment.

Groundwater Withdrawals

Well pumping, which represents important losses from the flow system, was incorporated into the model using the multi-aquifer well (MAW) package, which can simulate well pumping from screened intervals that span multiple model layers, which is the case for many wells in the Kirkwood-Cohansey aquifer system that pump from layers 1 and 2. However, if a well extended across model layers that were in different hydrogeologic units, for example the Kirkwood-Cohansey aquifer system and the Manasquan-Shark River confining unit, the screened interval was edited to simulate pumping only from the aquifer layer and thus the screened interval in the model may differ from that in NJWaTr and (or) NWIS databases. For settings used for the 387 pumping wells simulated by the MAW package, refer to the model archive data release (Fiore and Colarullo, 2023).

Although losses from groundwater withdrawals within the JBMDL are small relative to base flow losses to streams, the varying rates and depths from which groundwater is pumped impose strong local controls on flow geometry that can fundamentally alter contaminant pathlines. If dissolved contaminants are present anywhere within the capture zones of active water-supply wells, it is likely that the contaminants may be present in the water pumped from those wells.

Recharge

The principal source of water influx to the subsurface flow model is effective precipitation recharge. Effective precipitation recharge is imposed as a boundary condition of type specific flux and represents the single largest source of water to the model. This recharge is what drives flow and advective transport of contaminants through the subsurface. Because its accurate quantification is essential to development of a reliable flow and transport model, recharge is treated as a calibration parameter and allowed to vary freely. Significant uncertainties associated with estimating abstractions from precipitation recharge caused by processes like evaporation and evapotranspiration also necessitate its inclusion as a calibration parameter.

Effective precipitation was incorporated into the model using the recharge (RCH) package of MODFLOW 6. Two generalized recharge zones were defined across the entire top layer of the active flow area to distribute infiltrating recharge through the top boundary of the model. One recharge value was used for wetland area, and another for non-wetland areas (fig. 8). Values for recharge in these zones were estimated by PEST and are described in the “Model Parameters” section of this report.

Bottom Fluxes

Along the model bottom, spatially distributed specified fluxes predicted by the RASA regional coastal plain model were assigned as boundary conditions (fig. 9). Specified fluxes along the bottom boundary of the model were assigned based on internal vertical fluxes extracted from the calibrated regional RASA groundwater flow model for the New Jersey Coastal Plain (Voronin, 2004). Although bottom boundary flows represent a small fraction of groundwater losses compared to base-flow losses, they can exert strong controls on the overall mass balance of the flow system. Given the large scale at which they were predicted, however, it is unlikely these fluxes substantially affect local patterns of flow and advective transport near supply wells. Upward influxes and downward outflows across the bottom boundary of the model were incorporated into the model using the well (WEL) package.

AFFF Area 1

AFFF area 1 contains three buildings—Hangar 2251, Building 1708, and Hangar 1837. Particle tracks from Hangar 2251 indicate flow toward the east and northeast and terminate at an unnamed tributary to South Run identified by the DEM (discussed hereafter as “tributary A” to differentiate from other unnamed tributaries in the study area). The location of the DEM-identified drainage portion of tributary A is consistent with the location of an “underground water feature” in Aerostar, S.E.S., LLC (2017). If this drainage receives groundwater discharge, the particle tracks indicate the surface water may be contaminated at this location. If tributary A does not receive groundwater discharge at this location and groundwater continues to disperse downgradient, the effects of other transport properties on PFAS would determine whether downgradient surface water features may be contaminated, the most likely being the main stem of South Run. Tributary A flows into another unnamed tributary of South Run (discussed hereafter as “tributary B”) that flows into the main stem of South Run, The DEM processing has tributary A divided into a northern stem and a southern stem (AFFF area 1 particle flows into the southern stem), but these may actually be part of the same drainage feature.

Particle tracks from Building 1708 and Hangar 1837 indicate groundwater flow to the east. Particle tracks terminate at another unnamed tributary of South Run (discussed hereafter as “tributary C”) that flows into the main stem of South Run downstream of tributary B. Aerial photos (Aerostar, S.E.S., LLC, 2017) indicate tributary C may be an ephemeral drainage; if so, its representation in the model may be overestimated and the eastward flowing PFAS from Building 1708, especially if these are ephemeral drainages that have water levels higher than the water table. The PFAS emanating from Building 1708 and Hangar 1837 may discharge farther to the east if not affected by transport processes. Observed groundwater levels (Tehama LLC, 2020) indicate eastward groundwater flow, but also a northward flow component. Analysis of a surface water sample collected at the confluence of South Run and tributary B had a combined PFOS and PFOA concentration greater than 70 ng/L; it is inconclusive whether this high concentration resulted from flow from Area 1 or from elsewhere at JBMDL.

AFFF Area 2

Particle tracks from AFFF area 2 all indicate eastward flow of groundwater. Particle tracks terminate primarily at South Run unnamed tributary B, with a small number of particles terminating in wetlands near the confluence of tributary B and two other unnamed tributaries to South Run that flow into tributary B (hereafter discussed as “tributary D” and “tributary E”), which indicates surface water at these locations may be contaminated with PFAS if they receive groundwater discharge from area 2. Single particles also terminate in tributary A, Bowkers Run east of the McGuire Air Force Base boundary, and the main stem of South Run. But given the large distance of these indicated flow paths, it is unlikely PFAS originating at area 2 would be detectable at these locations even if infiltration at area 2 discharges there. These portions of tributaries A and C, as well as downstream portions of tributary B, are drainages identified by the DEM processing. If these drainages do not receive groundwater discharge and groundwater continues to flow downgradient, the effects of other transport properties on PFAS would determine whether downgradient surface water features may be contaminated, the most likely being the main stem South Run. Analyses of two water samples collected upgradient of area 2, where tributary B is not contained in a culvert, indicate individual concentrations of PFOA and PFOS that were both less than 10 ng/L (Aerostar, S.E.S., LLC, 2017), which is consistent with the groundwater flow directions indicated by the water levels.

AFFF Area 3

Particle tracks from AFFF area 3 show a predominantly eastward flow and terminate at the northern stem of tributary A. If the drainage feature of tributary A does not receive groundwater discharge as discussed above and is overrepresented as a sink in the model, discharge of groundwater and possible PFAS contamination may occur at the main stem of South Run or downstream portions of tributary B, particularly if PFAS concentrations are higher in the northern portion of area 3. Some particle tracks from area 3 terminate at the western flank of area 3 at tributary B. This is also a feature identified by the DEM processing that does not appear on the NHD nor as a groundwater feature from Aerostar, S.E.S., LLC (2017) so is likely overrepresented in its effects on the hydrologic system. The termination of particles here is likely caused by its proximity to tributary B to the western border of area 3, and hydraulic heads indicate it will more likely flow to the east. Notably, despite the large size of source area 3, particles from this area terminate closer to their point of origin than do particles from source areas.

AFFF Area 4

Particle tracks from AFFF area 4 indicates flow to the southeast. Termination of these particle tracks is at the main stem of South Run, as well as another unnamed tributary to South Run (discussed hereafter as “tributary F”) that flows into the main stem of South Run. Area 4 is small relative to the model cell size that encompasses it, no nearby groundwater levels are available, and no streams are nearby, so at this resolution the flow paths emanating from area 4 are extremely generalized. Only 4 particles were released from area 4. Three of these particles terminate at tributary F, the downstream-most tributary to South Run, which indicates tributary F may be contaminated. The particle that terminates at main stem South Run is a feature identified by the DEM processing that aerial photos (Aerostar, S.E.S., LLC, 2017) indicate may not exist. Thus, this particle likely would terminate downstream at the main stem of South Run.

AFFF Areas 5 and 6

AFFF areas 5 and 6 are discussed together owing to their nearby location in the same wetland area. Particle tracks from these two areas terminate within the surrounding wetland, as well as at tributary E. Because of their location in the wetlands, flow paths in these areas are likely short and shallow, and thus contribute only minor amounts of recharge to groundwater.

Analyses of surface water samples collected from tributary E east of AFFF area 6 indicated that the combined concentrations of PFOA and PFOS exceeded EPA limits (Aerostar, S.E.S., LLC, 2017), which is consistent with particle track results. Tributary D to the north also had a surface water sample with concentrations exceeding EPA limits, but it is inconclusive whether this result is because of the wetlands representation in the model underestimating the flow paths or whether the high concentrations originated from a different PFAS source.

AFFF Area 7

Particle tracks from AFFF area 7 indicate eastward and southeastward flow. All particle tracks terminate at Jacks Run, which flows through area 7. This is consistent with surface water samples from Jacks Run near AFFF area 7 with concentrations that exceeded EPA limits for combined concentrations of PFOA and PFOS (Aerostar, S.E.S., LLC, 2017). Jacks Run eventually flows into Little Pine Lake, in which high concentrations PFAS have been reported by the NJDEP (Goodrow and others, 2018).

AFFF Area 8

Particle tracks from AFFF area 8 terminate to the east at Bowkers Run, which is a tributary to Jacks Run. No surface water samples were collected in Bowkers Run (Aerostar, S.E.S., LLC, 2017), but if PFAS are present in area 8, there is a strong possibility of PFAS contamination of Bowkers Run. Bowkers Run eventually flows into Little Pine Lake, in which high concentrations of PFAS have been reported by the NJDEP (Goodrow and others, 2018).

AFFF Area 9

Particle tracks from AFFF area 9 indicate eastward flow that terminates at Larkins Run, which is a tributary of Jacks Run. A drainage within area 9 was considered by the DEM processing to flow into South Run, but the water actually flows through a culvert under the runway into Larkins Run. This discrepancy is a consequence of the minor variation in topography in the wetlands between the runways where area 9 is located, combined with the elevated runways that the DEM processing believed to be topographic divides that did not account for the engineered structures. In a surface water sample collected from Larkins Run at a more upstream northern reach than where the model indicates discharge to Larkins Run, combined PFOA and PFOS concentration was less than the EPA limit, but individual PFOA and PFOS concentrations both exceeded NJDEP MCLs for those compounds (Aerostar, S.E.S., LLC, 2017). Because particle tracks indicate most discharge occurring at other reaches of Larkins Run than where this sample was collected, concentrations of PFAS that exceed the EPA limit may occur farther downstream in Larkins Run. Alternatively, the PFAS detected in the Larkins Run sample may indicate flow from area 9 may be more toward the northeast than the model indicates. Given the relatively few groundwater levels measured near area 9, the flow paths are not well defined.

AFFF Area 10

Particle tracks from AFFF area 10 terminate in the main stem of South Run, which flows through area 10. Contamination of South Run by PFAS originating in area 10 is consistent with two surface water samples collected at South Run within area 10, which both had concentrations exceeding EPA limits for combined PFOA and PFOS (Aerostar, S.E.S., LLC, 2017).

AFFF Area 11

Particle tracks from AFFF area 11 terminate at the main stem of South Run, and the northern branch of tributary A. A surface water sample was collected from South Run near area 11 with concentrations below EPA limits for combined PFOA and PFOS but above NJDEP MCLs for individual PFOA and PFOS (Aerostar, S.E.S., LLC, 2017). This sample was collected to the east-northeast of area 11, and within the flow paths from area 11. Thus, high PFAS concentrations at South Run at this sampling location likely was caused by sources in area 11.

AFFF Area 12

Particle tracks from AFFF area 12 indicate flow to the east and south, which is consistent with observed water levels. Particle tracks terminate at the unnamed tributary D of South Run that flows through area 12. This is consistent with a surface water sample collected in tributary D within area 12, which exceeded EPA limits for combined PFOA and PFOS (Aerostar, S.E.S., LLC, 2017).

AFFF Area 13

Particle tracks from AFFF area 13 indicate east and northeast flow that terminates at tributary D. This is consistent with a surface water sample collected in tributary D near area 13, which exceeded EPA limits for combined PFOA and PFOS (Aerostar, S.E.S., LLC, 2017).

AFFF Area 14

At AFFF area 14, the discharge of wastewater effluent into the subsurface via infiltration basins causes high hydraulic heads locally depending on which basin is in use (Fiore, 2016). As a consequence, groundwater moves radially away from the infiltration basins, which is consistent with the multiple flow directions indicated by particle tracks simulated by the model. However, the model does not account for the increased infiltration into groundwater associated with this site.

Most particle tracks terminate to the south at Cannon Run and its tributaries, which is consistent with two surface water sample with concentrations exceeding EPA limits for combined PFOA and PFOS (Aerostar, S.E.S., LLC, 2017). Additional particle tracks terminate at an unnamed tributary of North Branch Rancocas Creek to the east of area 14. If infiltration is focused on basins in the northern part of the site, flow may also occur toward Ong Run in the northeast, though no particle tracks terminated at Ong Run in this model. Which basins were in use during the 2018 groundwater level synoptic event was not reported. No samples were collected from the unnamed tributary to North Branch Rancocas Creek nor Ong Run (Aerostar, S.E.S., LLC, 2017), but high PFAS concentrations may be present in those stream locations.

AFFF Area 15

AFFF area 15 covers a large area, and particle tracks indicate multiple flow directions. Tributary D flows through area 15, and many particle tracks terminate in that tributary. Two surface water samples collected in tributary D within area 15 were below the EPA limit for combined PFOA and PFOS concentrations, but above the NJDEP MCL for individual PFOA and PFOS (Aerostar, S.E.S., LLC, 2017). The sampling locations were in the upstream portion of tributary D, and additional particle tracks terminate downstream of the sampling locations, so higher PFAS concentrations may be present in tributary D downstream of the sampling locations depending on where within AFFF area 15 the PFAS concentrations in groundwater are highest.

Substantial flow from area 15 also occurs to the southwest. Both observed and simulated water levels indicate a very steep hydraulic gradient in this direction, which is likely related to the upland gravel surficial deposit that underlies area 15. Particle tracks flowing west predominantly terminate at Indian Run and at unnamed tributaries to Indian Run. No surface water samples were collected in these streams (Aerostar, S.E.S., LLC, 2017), but model results indicate the streams may be contaminated by PFAS originating in area 15.

AFFF Area 16

AFFF area 16 consists of two historical fire training areas (note that AFFF areas 16-21 are on Lakehurst Naval Air Engineering Station and are shown on plate 2). Particle tracks from area 16 indicate groundwater flow to both the west and east and terminate at unnamed tributaries to Manapaqua Branch. The tributary to the west is discussed hereafter as “tributary G” and the tributary to the east “tributary H”. A surface water sample with concentrations that exceeded EPA limits for combined PFOA and PFOS concentrations was collected (Aerostar, S.E.S., LLC, 2017) in tributary G. The elevated concentrations of PFAS in this sample are to be expected with particle tracks terminating upstream of the sampling location. Tributary H is a drainage identified by the DEM processing method. If tributary H is a stream or other drainage feature at which groundwater discharges, particle tracks indicate it may be contaminated with PFAS. If tributary H is a drainage feature that does not receive groundwater discharge, PFAS may flow farther east, and its fate would be determined by effects of other transport processes.

AFFF Area 17

Particle tracks from AFFF area 17 terminate at pumping wells 291329 and 291330, which indicates those wells may be contaminated with PFAS. Wells 291329 and 291330 are listed as industrial-supply wells in the NJWaTr database (New Jersey Geological and Water Survey, 2018). A water sample collected from Ridgeway Branch, near and to the northeast of area 17 contained a combined concentration of PFOA and PFOS that exceeded the EPA limit (Aerostar, S.E.S., LLC, 2017). Although no particle tracks in the model terminate at Ridgeway Branch, it is likely that the PFAS in this sample originated in area 17, but the model may overestimate the effects of the pumping wells on groundwater flow directions near area 17. In addition, if these wells are not continuously in use, the sample may have been collected at a time when the wells were not pumping, a condition that is not represented by the steady-state nature of this simulation. This apparent discrepancy may also be caused by a resolution issue given the small geographic area of AFFF area 17 relative to the regional model cell size, such that heterogeneity is not accurately represented at this location.

AFFF Area 18

Particle tracks from AFFF area 18 indicate eastward flow terminating at unnamed Manapaqua tributary G as well as westward flow terminating at another unnamed tributary to Manapaqua Branch (hereafter discussed as “tributary I”). Eastward particle tracks are consistent with a surface water sample collected along tributary G that contained a combined concentration of PFOS and PFOA that exceeded the EPA limit (Aerostar, S.E.S., LLC, 2017). Given the proximity of particle termination locations at tributaries G and I to the main stem of Manapaqua Branch, it is also possible PFAS may enter Manapaqua Branch from AFFF area 18, but no particle tracks terminate at the main stem in the model. Similarly, pumping well 291227, which is listed as an industrial-supply well in the NJWaTr database (New Jersey Geological and Water Survey, 2018), is also near area 18 and may be vulnerable to PFAS contamination despite the absence of any particle tracks terminating at the well in this steady-state simulation.

AFFF Area 19

Particle tracks from AFFF area 19 indicate northeastward groundwater flow, and termination at Obhonan Ridgeway Branch, which is a tributary of Ridgeway Branch, and surrounding wetlands. A surface water sample was collected that was below the EPA limit for combined PFOA and PFOS concentrations, but above the NJDEP MCL for individual PFOA and PFOS (Aerostar, S.E.S., LLC, 2017). This sample was collected on a small tributary of Obhonan Ridgeway Branch that was not identified by either the NHD dataset or the DEM processing. The particle tracks from area 19 terminate downstream of the sampling location, so PFAS concentrations in Obhonan Ridgeway Branch may be higher downstream of that sampling location. A cluster of particles terminates along runways at a drainage identified by the DEM processing method. This result is likely an artifact of the processing method inaccurately identifying drainages around the runways, similar to AFFF area 9, so these particles will likely continue northeast and discharge at Obhonan Ridgeway Branch.

Some particles also terminate at pumping well 291270 near the northern border of JBMDL, and no particles terminate at pumping well 291126 near the northern boundary of area 19. However, given the steady-state conditions of the simulation, they should be considered vulnerable to contamination by PFAS if transient conditions cause differing hydrologic conditions. These wells are listed as industrial-supply wells in the NJWaTr database (New Jersey Geological and Water Survey, 2018).

AFFF Area 20

AFFF area 20 consists of 3 sites. These include Hangar 690, a helistat crash site directly west of the hangar, and a plane crash site to the south of the hangar. Some particle tracks from Hangar 690 and the helistat crash site terminate at tributaries G and H, downstream of the endpoints of particle tracks from AFFF area 16, which indicates additional PFAS discharges into these tributaries from AFFF area 20. However, Aerostar, S.E.S., LLC (2017) indicates that tributary G is within a culvert where it passes through the helistat crash site, and concentrations of PFAS in a water sample collected on the downstream end of the culvert were less than both EPA and NJDEP limits. This indicates the culvert is not receiving PFAS-contaminated groundwater, either because the culvert is not in hydraulic connection with groundwater or because the groundwater flows underneath the culvert and discharges to the south. The low concentration sample also indicates that the high PFAS concentration detected in the upstream sample from tributary G decreases during the approximately 3,500 ft stretch between the sampling locations, but the exact reason for this decrease is unknown. If the PFAS-contaminated groundwater originating at Hangar 690 and the helistat crash site flows underneath this culvert, it may discharge at tributary I, to another unnamed tributary to Manapaqua Branch (hereafter discussed as “tributary J”), to the main stem of Manapaqua Branch, or may be captured by one of the pumping wells to the south of these sites, such as 291333 at which 3 particles terminated.

Particle tracks from the plane crash site terminate at tributary J and at pumping wells 291331 and 291333, which indicates these locations may be contaminated with PFAS. Tributary J was identified by the DEM processing method but is not included in the NHD, so may not exist, in which case these particle tracks either would terminate at surface water locations farther southwest or be captured by 291331 and (or) 291333, which are industrial-supply wells, according to the NJWaTr database (New Jersey Geological and Water Survey, 2018). If effects of these two wells are overestimated in the simulation, groundwater flow would likely continue toward the south and discharge at the main stem of Manapaqua Branch or downstream on tributary J if transport processes did not lower concentrations over that distance.

AFFF Area 21

Particle tracks from AFFF area 21 indicate a southeastward flow direction and terminate at tributaries G and J. The tributary G discharge locations coincide with the collection site of a sample in which PFAS concentrations exceeded EPA limits for combined PFOA and PFOS, which indicates the contamination present in tributary G may be caused by area 21. These high concentrations, however, could have resulted not only from PFAS sources at area 21, but also from sources at AFFF area 16 that had discharge upstream of the sampling location. Tributary J has not been sampled and, as stated above, may not exist. However, if tributary J is a real drainage feature, it may be contaminated with PFAS given the termination of particles from area 21 as well as area 20 plane crash site.

Reconnaissance Area 4

Particles reverse tracked from reconnaissance area 4 terminate to the southwest. Reconnaissance area 4 is located north of AFFF area 4, which had flow paths to the southeast and is east of the flow paths that intersect reconnaissance area 4. Therefore, model results indicate that any PFAS that may be present in reconnaissance area 4 did not originate at AFFF area 4 and there may be an additional source of PFAS to the west of AFFF area 4 and to the east of AFFF area 3. MODPATH tracking results also show particles reversed tracked from the Vincentown aquifer underlying reconnaissance area 4, indicating that wells in the Vincentown aquifer at this location may be contaminated with PFAS, assuming other transport processes do not affect concentration of the PFAS in groundwater along that flow path. Reconnaissance area 4 is within the unconfined portion of the Vincentown aquifer. Fiore (2020) suggested, on the basis of downward vertical head gradients indicated by continuously monitored water levels, that the unconfined portion of the Vincentown aquifer may be contaminated and that such contamination may be expected when there is no overlying confining unit preventing downward migration of PFAS.

Reconnaissance Area 14

Many particles reverse tracked from reconnaissance area 14 terminated north of AFFF area 14 and south of AFFF area 15. Some particle paths are entirely within the Kirkwood-Cohansey aquifer system, and some that were reversed tracked from the confined portion of the Vincentown aquifer terminated at the water table after having passed through the unconfined portion of the aquifer, represented in the model as locations where thickness of the Manasquan-Shark River confining unit is less than 28 ft, as discussed earlier. These particle tracks may be associated with the representation of the upland gravel in the model, given their starting locations near the edge of the upland gravel. If a PFAS source area occurs in this general location (between AFFF areas 14 and 15), the likelihood of PFAS-contaminated groundwater reaching reconnaissance area 14 would depend on the effects of transport processes on PFAS. However, the similar tracks these particles take indicate these flow paths are regional at this model resolution rather than a specific identification of flow from a PFAS source area, so increased grid refinement would be needed for certainty.

Additional particle tracks in the Kirkwood-Cohansey aquifer system also originate in the vicinity of AFFF area 14. Thus, it is likely that PFAS present in reconnaissance area 14 originates at AFFF area 14, but model resolution is too limited to show a substantial number of reverse particle tracks that leave reconnaissance area 14 terminating in AFFF area 14. Additional particle tracks in the Kirkwood-Cohansey aquifer system originate within reconnaissance area 14 and in adjacent bordering locations, so shallower wells, particularly those near the edge of the reconnaissance area, likely intersect shallower groundwater flow paths compared to those that would emanate from AFFF area 14 or elsewhere. These shallower flow paths indicate that any PFAS in shallow wells may have a different source than AFFF area 14.

Particles reverse tracked from the Vincentown aquifer underlying reconnaissance area 14 that begin on JBMDL indicate that wells completed in the confined portion of the Vincentown aquifer may be contaminated with PFAS, assuming other transport processes do not affect concentration of the PFAS in groundwater along that flow path. However, lower concentrations are likely, considering the long particle flow paths from the outcrop area to the downdip limit of the confined portion of the aquifer.

Reconnaissance Area 16

Particles were reverse tracked from reconnaissance area 16 in both the Kirkwood-Cohansey aquifer system and the Vincentown aquifer.

Kirkwood-Cohansey aquifer system particle tracks in layers 1 and 2 originate southwest of reconnaissance area 16, many of which are north and east of AFFF area 16, which has flow paths to the south. Thus, the model indicates that any PFAS potentially present in reconnaissance area 16 may originate at a source other than AFFF area 16. The particle tracks that leave reconnaissance area 16 span a large adjacent boundary of JBMDL, which increases the likelihood that there may be another source in that location north of AFFF area 16. Many of these particle tracks leaving reconnaissance area 16 terminate in reconnaissance area 16, so if a relatively shallow well is contaminated with PFAS it may have captured water from a shallower flow path from an unknown source area in close proximity.

Some of the Kirkwood-Cohansey aquifer system and all of the Vincentown aquifer particle tracks originate far west from reconnaissance area 16, so any PFAS that flow along this path would likely be affected by other transport processes prior to its arrival, so it is less likely that those areas are contaminated with PFAS in reconnaissance area 16. Similar to reconnaissance area 14, the downdip limit of the Vincentown aquifer runs through reconnaissance area 16, so it is less likely that PFAS sources in its outcrop area would affect groundwater in the confined portion.

Reconnaissance Area 17

MODPATH particle tracking results show only a small number of reverse particle tracks from reconnaissance area 17 that terminate outside the borders of that reconnaissance area. Ridgeway Branch flows along the southwest border of the reconnaissance area, and the lack of particles beginning north of Ridgeway Branch that made it onto JBMDL to the south indicates Ridgeway Branch is acting as a hydrologic divide in the model at this location. It is inconclusive whether these effects of Ridgeway Branch acting as a flow barrier is only represented in the model as such, or whether the stream also provides a flow barrier in nature. Because of the possibly overestimated effect of pumping wells in this area, as evidenced by all forward particle tracks from AFFF area 17 being captured by pumping wells described earlier, the absence of particles terminating at AFFF area 17 are inconclusive.

On the southern edge of reconnaissance area 17, Kirkwood-Cohansey aquifer system flow paths are to the south and terminate along the easternmost border of JBMDL. This result indicates that any PFAS in the Kirkwood-Cohansey aquifer system in that location, if present, may emanate from a location where no AFFF source area has been identified on the easternmost border of JBMDL.

Many particles reverse tracked from reconnaissance area 17 terminated at the water table within reconnaissance area 17, which indicates that if a relatively shallow well within reconnaissance area 17 is contaminated with PFAS, it may have captured a shallower flow path from an unknown source area in close proximity.

Reconnaissance Area 18

Particle tracks that leave reconnaissance area 18 terminate to the west and northwest of that location. The cells where particles are terminating at the water table are more than 1,000 ft east of AFFF area 18, which indicates any PFAS that may be present in reconnaissance area 18 may originate from a source other than AFFF area 18, including similar area along the easternmost border of JBMDL also indicated by reconnaissance area 17 results. No additional AFFF source areas are located in that portion of JBMDL.