Installation of Monitoring Well MG 2220

Well MG 2220 was drilled by the USGS during April 4–5, 2016, to monitor the shallow part of the aquifer near the extraction well MG 2201 (S-1) and along a possible route of contaminant migration between extraction well MG 2201 (S-1) and the abandoned NPWA production well MG 665 (S-9) (fig. 2). Well MG 2220 is 58 ft southwest of well MG 2201 (S-1) (fig. 2) and was drilled to a total depth of 82 ft bls. It was initially completed with 6.5-inch (in.) inside-diameter steel casing to 19 ft bls and with a 6-inch nominal open hole from 19 to 82 ft bls. A borehole video log was collected on April 11, 2016, and geophysical and flowmeter logs were collected on August 22, 2016. On October 2, 2017, a 2-in.-free-standing (ungrouted) polyvinyl chloride (PVC) casing was installed with a 10-ft screened interval from 72 to 82 ft bls. Permanent installation of the casing and screen with sand pack and grout was postponed until further investigations could determine the most contaminated depth interval in the borehole to place the screen. The borehole currently (2024) remains an open borehole with a temporary screen and no grout or sand pack.

Reconstruction of Well MG 668 (GKM)

Well MG 668 (GKM) is a former industrial production well that was initially drilled to a depth of 208 ft bls on an unknown date sometime before 1957 (Longwill and Wood, 1965) and was installed in a well pit. During the RI in 1992, packer testing of isolated intervals to a reported depth of 187 ft showed that the greatest concentration of PCE (330 µg/L) was in the shallowest isolated zone, from 0 to 28 ft (CH2M-Hill, Inc., 1993a). The reference datum for depths was not reported with these findings (CH2M-Hill, Inc., 1993a), so it is not known whether those depths are measured from land surface or from top of casing in the well pit (about 5.7 ft below land surface).

Well MG 668 (GKM) is in a pit (fig. 5A) that frequently collects water during storms, although how water enters the pit during storms is not specifically known. Before reconstruction of this well in October 2017, a channel in the concrete pit floor and slot in the steel well casing (fig. 5B) diverted water from the pit into the well. Inflow from surface water that collects in the pit was likely the cause of observed sharp spikes in rising water levels in well MG 668 (GKM) during storms, discussed later in the section “Water Levels.” The rising water-level spikes may temporarily increase the downward hydraulic gradient in the well from shallow to deep water-bearing fractures, thus increasing the potential for downward flow of the most contaminated water identified as present at shallow depths less than about 28 ft (CH2M-Hill, Inc., 1993a) into deeper fracture zones. Although no downward flow was measured in the well during geophysical logging by the USGS under ambient conditions in July 2012 (discussed in section “Borehole Geophysical Logging and Video Surveys”), the water-level spikes indicate potential for transient downward gradients.

To eliminate or reduce this possible connection for contaminant movement from shallow to deep water-bearing zones, well MG 668 (GKM) was filled with grout from its total depth of 208 ft bls to 35.4 ft bls, the slot in the casing was plugged, and the casing was extended to 2.37 ft above the concrete pit cover on October 2–3, 2017 (fig. 5C). Despite efforts to prevent pit water from entering the well, a video log collected when water was standing in the pit on July 28, 2018, documented that pit water leaked to some extent into the well at the joint between the old well casing and the casing extension added by the driller (fig. 5D). Although the target depth for well reconstruction was about 50 ft bls to include all water-bearing fractures to that depth, the water-bearing fracture at about 37–38 ft bls was inadvertently grouted during reconstruction.

MG 665 (S-9)

Well MG 665 (S-9) is a former production well that was initially drilled to a depth of 300 ft bls in 1954 (tables 1 and 3) (Longwill and Wood, 1965). The USGS collected borehole geophysical logs in well MG 665 (S-9) in 1992, and some results of that logging are described in the RI report (CH2M-Hill, Inc., 1993a). At the time of logging by the USGS in April 1992, the well had 46.5 ft of 8-in.-diameter casing and was an 8-in.-diameter open hole to a depth of 300 ft bls; however, the well was partly caved at 280 ft bls and partly blocked from 293 to 300 ft bls (Charles Wood, U.S. Geological Survey, written commun., May 1, 1992). The original digital log data were not available for the geophysical logs, but scans of the paper logs (fig. 1.2) were digitized resulting in the log suite shown in figure 6. Although originally selected to serve as an extraction well as part of the EPA 1996 remediation plans, in 1997 the EPA decided not to pump well MG 665 (S-9) because contaminant concentrations in water from the well were low, and the well was abandoned (plugged and sealed) sometime during the period 1998–2000 (U.S. Environmental Protection Agency, 2003).

The caliper logs showed openings at about 55, 65, 110, 120, 130–140, and 160 ft bls that likely are high-angle fractures as openings appear to span more than 1 ft (fig. 6). Openings near 85–95, 173, 210, and 236 ft bls may be low-angle or bedding-plane fractures because these openings are thin, spanning less than 1 ft. Inflections on the fluid-log profiles that may indicate water-bearing fractures are apparent at about 210 ft bls on the water-temperature log and about 235 ft bls on the fluid resistivity log, corresponding to openings at similar depths on the caliper log (figs. 6 and 1.2). Smaller inflections on fluid logs are apparent near openings shown on caliper logs at or near depths of 55–65 ft bls and 85–95 ft bls. The natural gamma log shows about four relatively thin zones (2–5 ft in thickness) of elevated activity near about 18, 70, 210, and 280 ft bls and a thicker zone (10 ft in thickness) at 290–300 ft bls that are typical of marker beds and, if present in other wells, can be possibly used for lithologic-log correlation.

MG 679 (S-11)

Well MG 679 (S-11) is a former production well that was drilled in 1960 with an 8-in.-diameter casing to 41.5 ft bls and was reported to be completed as an open hole below the casing to a depth of about 308 ft bls (tables 1 and 3) (Longwill and Wood, 1965). Reported lithologies on drilling logs indicate predominantly red shale from about 0 to 180 ft bls, blue-gray argillite interbedded with red shale from 180 to 200 ft bls, sandstone from 200 to 210 ft bls, blue-gray argillite and siltstone from 210 to 285 ft bls, and reddish-brown shale interbedded with blue-gray argillite from 285 to 308 ft bls (Longwill and Wood, 1965, p. 56–57). Depth to static water level at the time of logging (July 17, 2012) was about 8.4 ft bls, and logs extended to a depth of about 286 ft bls indicating some collapse or filling of the borehole since initial drilling.

The caliper and acoustic-televiewer (ATV) logs show numerous fractures throughout the borehole that are generally more closely spaced in the intervals from about 40 to 140 ft bls and 195 to 235 ft bls than in other intervals of the borehole and a very large (borehole diameter greater than 16 in.) opening from about 268 to 275 ft bls (fig. 7). Many of the fractures on the caliper and ATV logs appear to be high angle. However, low-angle fractures, likely parallel to bedding planes, are also common. Natural gamma logs show slightly lower activity from about 195 to 230 ft bls, which overlaps and is consistent with the reported sandstone lithology from 200 to 210 ft bls. The fluid-temperature log shows relatively constantly increasing values with depth, but the observed water-temperature gradient is less than the natural geothermal gradient (reported by Paulachok [1986] to be about 1 degree Celsius per 100 ft in Philadelphia, 24 miles southeast of Souderton) and thus may indicate small amounts of vertical flow throughout the borehole. However, borehole flow was not measured so the presence of flow under ambient conditions at the time of logging can only be inferred from the temperature log. The fluid-resistivity log shows several inflections (the largest at about 60, 85, 140, 195, and 220 ft bls) that generally correspond to changes in lithologies, notable fractures, or both, and indicate possible water-bearing zones (fig. 7).

The borehole video log collected by the USGS on July 10, 2012, when the depth to water was 8.2 ft bls showed some low-angle and numerous high-angle fractures throughout the borehole (fig. 8). High-angle or mixed high- and low-angle probable water-bearing fractures were noted at depths of about 76–87, 101.9, 113–118.5, 165.8–168.8, 173–174, 192.5–209.4, 232.5–235, and 264.7–276.8 ft bls. Other high-angle or mixed high- and low-angle possible water-bearing fractures were noted at depths of about 130–134, 151–155, 162.5, 182, 188–190, 210–220, 221–229, 278.9–282.2, and 284.2 ft bls. Low-angle possible water-bearing openings were noted at depths of about 89.1, 93.5, 134.5, and 138 ft bls. Apparent turbulence in borehole water was noted at depths near 76–87, 101.9, 173–174, and 232.5–235 ft bls in association with high-angle fractures. Large amounts of particulates that appear to be bacterial matter are present along borehole walls and in the water column, especially in the upper 150 ft of the borehole. Images of selected fractures and features in well MG 679 (S-11) from the 2012 video log are shown in figure 8.

MG 668 (GKM)

Well MG 668 (GKM) is a former production well that was drilled on an unknown date, but sometime before 1957, to a reported depth of 208 ft (Longwill and Wood, 1965) as discussed in the section “Reconstruction of Well MG 668 (GKM).” At the time of logging by the USGS on July 18, 2012, the well head was located in a well pit (fig. 5A) about 5.7 ft below the top of the concrete pad forming the top of the pit; the well had an 8-in.-diameter casing to 9.4 ft bls and was an 8-in.-diameter open hole below the casing to a depth of about 97 ft bls and a 6-in.-diameter hole from about 97 to 195 ft bls (tables 1and 3). Depth to static water level at the time of logging on July 18, 2012, was about 22.9 ft bls. The top of the concrete pad was approximately flush with land surface at the edge of the parking lot and was about 1 ft higher than land surface of the grassy slope to the southwest (fig. 5C). The top of the concrete pad was assumed to be at land surface for logs and water levels measured in the well unless otherwise specified.

The caliper and ATV logs show numerous fractures throughout the borehole that are generally more closely spaced in the intervals from about 10 to 97 ft bls and 120 to 130 ft bls than in other intervals of the borehole (fig. 9). Natural gamma log shows little variation with depth, with a single peak of slightly elevated activity at about 53 ft bls. The single-point resistance log shows values higher in magnitude at depths above about 96 ft bls that may result from lithology (sandstone more resistant than shale), but these values also may be affected by changes in borehole diameter and fluid-resistivity values at depths above 96 ft bls and so should be interpreted with caution. The fluid-resistivity log shows an inflection at about 30 ft bls, a slope change at about 80 ft bls, and a larger, sharp inflection at about 96 ft bls that had higher fluid-resistivity values above 96 ft bls than below 96 ft bls, indicating water that is relatively more dilute than deeper water enters the borehole through fractures at or near 96 ft bls. A small inflection at 96 ft bls is also indicated on the fluid-temperature log, which additionally shows small changes in slope at about 59, 39, and 32 ft bls. Little to no vertical borehole flow was measured under ambient conditions (fig. 9). Under pumping conditions when the pump was set at 35 ft bls, an increase in upward borehole flow was measured at depths just above the fractures near 96 ft bls, indicating along with the other logs, that the fractures near 96 ft bls produce water and are likely among the most productive water-bearing features in the borehole. The flow rate of about 0.10 gal/min measured at 55 ft bls under pumping conditions was much less than the pumping rate of about 0.8–1.25 gal/min, indicating possible undermeasurement of flow because of leaky seals on the flowmeter diverter in borehole with rough walls and (or) that the fractures above 55 ft bls produce a substantial amount of water being pumped.

The borehole video log collected by the USGS on June 26, 2012, was referenced to depths below the top of the concrete pad estimated to be at land surface. The June 2012 video log showed some low-angle and numerous high-angle fractures throughout the borehole. Possible water-bearing high-angle fractures and associated openings were noted at about 29–33, 60–64, 74.7–76, 78.2–81, 96–98.8, 121–126, 127.4–129.7, 142.8–145, and 184–188 ft bls. Possible water-bearing low-angle fractures and associated openings were noted at about 37, 51, 59, and 136.6 ft bls. Of these features, openings near 30, 37, 59, 80, and 96 ft bls are associated with inflections on fluid logs or increases in flow that can be inferred to indicate water-bearing zones. Images of selected fractures from the 2012 video log of well MG 668 (GKM) are shown in figures 10 and 11. Above the static water level of 23.2 ft bls, borehole walls were wet and low-angle fractures were noted at about 16.2–18.4 and 19.5 ft bls (figs. 11A, 11C, and 11E).

The borehole video log collected by the USGS on July 26, 2018, was referenced to depths below the grassy slope next to, and about 1 ft lower than, the top of the concrete pad of the well pit (fig. 5C). Therefore, all depths for the 2018 video log are about 1 ft less than if referenced to depths below the top of the concrete pad and are referred to in feet below grassy slope (ft bgs). For example, the depth to the casing’s bottom in the 2018 video log is about 8.3 ft bgs, which is about 9.3 ft bls (and below the top of the concrete pad). When the borehole video log was rerun in 2018 after the borehole had been backfilled to a depth of 34.4 ft bgs (or 35.4 ft bls) and reconstructed to extend casing above the well pit, the depth to water was 7.9 ft bgs, which was about 15 ft higher than in 2012 when the borehole was open to 208 ft bls. Possible water-bearing low-angle fractures and associated openings in the 2018 video log were noted at about 14.9 and 18.3–20.9 ft bgs, depths which were above static water when the well was logged in 2012. Possible water-bearing high-angle fractures and associated openings were noted at about 27.7–32.6 ft bgs, similar to logs collected in 2012. Images from the 2018 video log for selected fractures in well MG 668 (GKM) are shown paired with same features from the 2012 video log in figure 11. The extensive vertical fractures apparent on the 2012 video log from about 29–33 ft bls appear reduced in extent on the 2018 video log (figs. 11E and 11F), possibly because of grouting when well was backfilled to 34.4 ft bgs (about 35.4 ft bls). In the 2018 video log, which was collected after several days of moderate to heavy rain (as recoded at Doylestown Airport, about 10 miles east of NP1 Site [National Oceanic and Atmospheric Administration, 2024b]) and several feet of water had collected in the well pit, water was noted to enter the borehole at the joint between the existing and extended casing at about 4.7 ft bgs (fig. 5D), indicating some surface water can still enter the borehole despite reconstruction efforts.

MG 2201 (S-1)

Well MG 2201 (S-1) is a monitoring well installed by the EPA that, at the time of logging by the USGS on July 19, 2012, had a 6-in.-diameter casing to 16.3 ft bls and was a 6-in.-diameter open hole below the casing to a depth of about 59 ft bls (tables 1 and 3). Depth to static water level at the time of logging on July 19, 2012, was about 23.1 ft bls.

The caliper and ATV logs show numerous fractures throughout the borehole that are generally more closely spaced in the intervals from about 20 to 32 ft bls and 50 to 57 ft bls than in other intervals of the borehole (fig. 12). The natural gamma log shows relatively small variation with depth, with peaks of slightly elevated activity at about 32.5 and 48 ft bls. The single-point resistance log shows values slightly higher in magnitude at about 26–27 ft bls and 34–50 ft bls that may result from lithology (sandstone is more resistant than shale). The fluid-resistivity log shows small inflections at about 27, 40, and 53 ft bls. Small inflections at about 26 and 28 ft bls are indicated on the fluid-temperature log. A small amount of vertical borehole flow (about 0.03 gallon per minute [gal/min]) was measured at about 38 ft bls above fractures at about 51–55 ft bls under ambient conditions (fig. 12). Under pumping conditions (about 1 gal/min), an increase in upward borehole flow of about 0.43 gal/min was measured at about 48 ft bls above the fractures at about 51–55 ft bls, indicating along with the other logs, that the fractures at about 51–55 ft bls and above about 28 ft bls likely are the most productive water-bearing features in the borehole.

The borehole video log collected by the USGS on July 19, 2012, showed a few low-angle and high-angle fractures throughout the borehole (fig. 13). Above the static water level of 21.6 ft bls, the borehole walls were wet and several high-angle fractures were noted at about 18.4 and 19.4–20.7 ft bls. Possible water-bearing low-angle fractures and associated openings were noted at about 25–28 ft bls. Possible water-bearing high-angle fractures were noted at about 51–55 ft bls. Of these features, openings near 25 and 51 ft bls are associated with inflections on fluid logs or increases in flow that can be inferred to indicate water-bearing zones. The water in the borehole was turbid (with apparent particulate matter) and had a reddish-brown haze shown in figure 13B, especially in the interval from static water level to about 40 ft bls, suggesting little borehole flow in that interval under conditions at the time of logging. The cause of turbidity is unknown but could be related to bacterial activity. Turbulence was noted at the fractures near 51 ft bls, indicating that inflow or outflow through those fractures is likely. Images of selected fractures and features from the 2012 video log of well MG 2201 (S-1) are shown in figure 13.

MG 2202 (S-2)

Well MG 2202 (S-2) is a monitoring well installed by the EPA that, at the time of logging by the USGS on July 17, 2012, had a 6-in.-diameter casing to 15.5 ft bls and was a 6-in.-diameter open hole below the casing to a depth of about 60 ft bls (tables 1 and 3). Depth to static water level at the time of logging on July 17, 2012, was about 10.3 ft bls.

The caliper and ATV logs show several fractures throughout the borehole that are generally spaced about 5–10 ft apart (fig. 14). The natural gamma and single-point resistance logs show relatively small variation with depth, which likely indicate minimal variation in lithology in intervals spanned by the borehole. The fluid-resistivity log shows relatively small inflections at about 16 and 23–26 ft bls. The fluid-temperature log also shows relatively small inflections at about 16 and 23–26 ft bls, in addition to a larger inflection at about 44 ft bls and other small inflections at about 36 and 52 ft bls. No borehole flow logs were collected in well MG 2202 (S-2). The available geophysical logs indicate that the fractures near 16, 23–26, and 44 ft bls are likely the most productive water-bearing features in the borehole.

The borehole video log collected by the USGS on June 26, 2012, at a static water level of about 10.2 ft bls, showed mostly low-angle fractures throughout the borehole. Images of selected fractures and features from the 2012 video log of well MG 2202 (S-2) are shown in figure 15. There appears to be a large opening just below the casing’s bottom at 15.5 bls, where an unknown object (which could be a plastic bottle) rests (fig. 15A); this opening is shown as an increase in borehole diameter to about 10 in. on the caliper log (fig. 14). Possible water-bearing low-angle fractures and openings were noted at about 27.4, 36.1, and 44.8 ft bls. Of these features, openings near 23–27 and 44.8 ft bls are associated with inflections on fluid logs that can be inferred to indicate water-bearing zones.

MG 2203 (S-3)

Well MG 2203 (S-3) is a monitoring well installed by the EPA that, at the time of logging by the USGS on July 16, 2012, had a 6-in.-diameter casing to 20 ft bls and was a 6-in.-diameter open hole below the casing to a depth of about 50 ft bls (tables 1 and 3). Depth to static water level at the time of logging on July 16, 2012, was about 13.9 ft bls.

The caliper and ATV logs show several fractures throughout the open borehole above about 43 ft bls (fig. 16). The natural gamma and single-point resistance logs show small variation with depth that likely indicate minimal variations in lithology in intervals spanned by the borehole. The fluid-resistivity log shows small inflections at about 20 and 41 ft bls. The fluid-temperature log also shows inflections at about 20 and 41 ft bls, in addition to inflections near about 25 and 27.5 ft bls. No borehole flow logs were collected in well MG 2203 (S-3). The available geophysical logs indicate that the fractures near 20 and 41 ft bls likely are the most productive water-bearing features in the borehole.

The borehole video log collected by the USGS on June 26, 2012, when the static water was about 13.1 ft bls showed several high-angle and some low-angle fractures throughout the borehole. Images of selected fractures and features from the 2012 video log of well MG 2203 (S-3) are shown in figure 17. There appears to be a high-angle fracture just below the casing’s bottom at about 19.7 to 23 ft bls. Other possible water-bearing high-angle fractures and openings were noted at about 39.1 to 41.2 ft bls. Possible water-bearing low-angle fractures and openings were apparent at about 28.5 and 43.5 ft bls. Of these features, openings near 20 and 41 ft bls are associated with inflections on fluid logs that can be inferred to indicate water-bearing zones.

MG 2204 (D-3)

Well MG 2204 (D-3) is a monitoring well installed by the EPA that, at the time of logging by the USGS on July 16, 2012, had a 6-in.-diameter casing to 15 ft bls and was a 6-in.-diameter open hole below the casing to a depth of about 198 ft bls (tables 1 and 3). The well was flowing at the time of logging on July 16, 2012.

The caliper and ATV logs show numerous fractures throughout the borehole, especially in the interval from the casing’s bottom at 15 ft bls to about 90 ft bls (fig. 18). The natural gamma log shows relatively small variation with depth, with some intervals with slightly elevated activity from about 122 to 137 ft bls and 155 to 160 ft bls separated by an interval of relatively lower activity. The single-point resistance logs are generally the inverse of the natural gamma logs, a relation which is common but more pronounced in logs of well MG 2204 (D-3) than in logs for other wells at the NP1 Site. Used in conjunction, the natural gamma and single-point resistance logs indicate the largest differences in lithology observed in the length of the borehole are in the shale-dominate intervals of 122–137 ft bls and 155–160 ft bls (relatively elevated gamma, lower single-point resistance), which are separated by a sandier interval (lower gamma, higher single-point resistance). The fluid-resistivity log shows relatively small inflections at about 27, 40, 50, 60, 70, 80, 98, 130, 154, 180, and 195 ft bls and a larger inflection at about 112 ft bls. The fluid-temperature log is nearly vertical, possibly indicating upward flow throughout the borehole in this flowing well. No borehole flow logs were collected in well MG 2204 (S-4). The available geophysical logs indicate that the fractures near 112 ft bls are likely among the most productive water-bearing features in the borehole but other fractures throughout the borehole also produce water.

The borehole video log collected by the USGS on July 10, 2012, when the well was flowing showed mostly high-angle and a few low-angle fractures throughout the borehole. Images of selected fractures and features from the 2012 video log of well MG 2204 (D-3) are shown in figure 19. Possible water-bearing high-angle fractures and openings were noted at about 24.9, 61.7, 77.3, 81–87, 181.5, 173.7, 187.4, 190.4, and 194 ft bls, with other probable water-bearing large vertical fractures at about 89 and 110–118 ft bls. Clumps of matter, likely related to bacterial activity, are present in the water column, especially at depths greater than about 74 ft bls. Turbulence as indicated by rapid nonlinear movement of these clumps was noted near the large fractures at about 86 and 110–118 ft bls. Possible water-bearing low-angle fractures were noted near 154 ft bls where there is a change in lithology indicated by the natural-gamma log and rock color. Of these features, openings near 62, 85, 110, 154, 180, and 194 ft bls are associated with inflections on fluid logs that can be inferred to indicate water-bearing zones. Intervals from 122 to 137 ft bls and 155 to 160 ft bls appear grayish green, as shown for selected depths in figure 20, in contrast the other intervals in the borehole, which are reddish brown (fig. 19). The grayish green intervals correspond to intervals of relatively elevated natural gamma activity (fig. 18).

MG 2220

Well MG 2220 is a monitoring well installed by the USGS and the EPA in 2016 that, at the time of logging by the USGS on August 22, 2016, had a 6-in.-diameter casing to 19 ft bls and was a 6-in.-diameter open hole below the casing to a depth of about 82 ft bls (tables 1 and 3). The depth to water at the time of logging August 22, 2016, as about 25.2 ft bls. Well MG 2220 was logged by the USGS with different equipment in 2016 than what was used for other logs collected in 2012. Some logs, such as natural gamma, electrical, and fluid logs collected using different equipment may have comparatively different specific values but similar relative values.

The caliper and ATV logs show numerous fractures throughout the borehole, especially in the interval from the casing’s bottom at 19 ft bls to about 62 ft bls (fig. 21). The natural gamma log shows relatively small variation with depth and a peak with slightly elevated activity at about 18 ft bls in the casing. The peak would likely be greater if it was not in the casing, which reduces the gamma signal. The electrical conductivity log (inverse of electrical-resistivity log) also shows relatively small variation with depth, except for an interval of relatively elevated conductivity from about 54 to 61 ft bls. The fluid logs were collected under ambient and pumping conditions. Under ambient conditions, the fluid-resistivity and temperature logs show small inflections at about 35 and 50 ft bls. When the well was pumped at a rate of about 5 gal/min, the fluid-resistivity and temperature logs show small to very small inflections at about 42 and 55 ft bls; the fluid-resistivity log also shows a small inflection at about 68 ft bls. The fluid-temperature log, other than very small inflections, is nearly vertical, which is consistent with upward flow throughout the borehole under pumping conditions. Little (about 0.12–0.055 gal/min) to no vertical borehole flow was measured under ambient conditions, but all measured flow was downward (fig. 21); downward flow increased below fractures near 42 ft bls, indicating possible inflow through those fractures, and then decreased below fractures near 49 ft bls, indicating possible outflow through the fractures near about 44–49 ft bls. These flow measurements should be interpreted with caution because rough borehole walls may have reduced effectiveness of diverter seals for the heat-pulse flowmeter in this well, which would result in reducing the accuracy and magnitude of measured flow values. Under pumping conditions, upward borehole flow was measured throughout the borehole; increases in upward flow were measured at depths above the fractures near 78, 70, and 42 ft bls, indicating possible inflow from or near those fractures, and a decrease in upward flow was measured above the fractures near 49 ft bls, indicating possible outflow through fractures near about 44–49 ft bls (fig. 21). Flow measurements in the highly fractured interval from 20 to about 62 ft bls should be qualified as more uncertain than typical because of possible issues in obtaining a borehole seal for the flowmeter diverter during this measurement. The fractures above 36 ft bls seem to be the most productive in the well because flow measurements indicate the balance of water being pumped from the well at 5 gal/min is derived from fractures in that interval. Transient pumping in nearby extraction well MG 2201 (S-1) was noted to affect water levels, and possibly borehole flow, in well MG 2220. The results of flow measurements, along with the other logs, show that the fractures in the interval above 62 ft bls, especially above 36 ft bls (about 25–35 ft bls) and at or near 38–42 and 62 ft bls, and near 70 and 78 ft bls produce water; additionally, the fractures near 50 ft bls appear to be hydraulically active because flow measurements indicate possible outflow at or near that depth, and fractures near 55 ft bls may also be hydraulically active as indicated by inflections of fluid logs.

The borehole video log collected by the USGS on April 1, 2016, when depth to water was 22.8 ft bls, showed numerous and extensive fractures in the interval from the casing’s bottom at 19 ft bls to about 62 ft bls and a few fractures below 62 ft bls. Images of selected fractures and features from the 2016 video log of well MG 2220 are shown in figures 22 and 23. Probable water-bearing large high-angle fractures and openings appear with planar features from about 25 to 42 ft bls and with less planar features from about 42 to 58 ft bls. Other possible water-bearing fractures and openings are at about 68.5–70 ft bls and 76.8–78 ft bls. In the interval from about 31–35 ft bls, the texture and form of the borehole indicates possible faulting in the bedrock, with apparent crushed rock near 31.6 ft bls and twisted texture in the rock near 33.5 ft bls (fig. 22).

Bedding Orientation of Outcrops

Field observations indicate that the local orientation of bedding at outcrops in the study area differs from the regional trend of N60°E and that the structure of rocks at and near the NP1 Site is more complex than a simple north-west dipping homocline. Measurements of strike and dip were made by the USGS in 2018 on rock outcrops (fig. 24) in the unnamed tributary of Skippack Creek between Wile Avenue and West Street at approximate locations shown on figure 2. The average strike of beds was approximately due north (measurements ranged from N6°W to N13°E) and the average dip was about 20 degrees west. Local differences in bedding orientation may be related to faulting. A zone of fractured rock indicating possible faulting that separates outcrop with apparent differences in bedding orientation can be seen in figure 24A. Typically, the beds are cut by numerous high-angle fractures oriented orthogonally such as those shown in figure 24B, with strikes of about N30°W and N60°E.

Additionally, one measurement shown on the geologic map of Longwill and Wood (1965, pl. 1) in the southern part of the NP1 Site gives a strike about N20°W and dip of 60 degrees to the southwest, much different than that of the regional homocline. Differences between local and regional structure may be attributed to nearby faulting, as indicated by mapped faults and changes in bedding orientation associated with the Chalfont Fault, a major regional fault trending east-west, and related minor faults located about 4,000–5,000 ft south of the NP1 Site. The geologic map of Longwill and Wood (1965, pl. 1) shows that displacement along the Chalfont Fault altered the strike and dip of geologic units north of the fault as well as an isolated northwest trending gray-shale unit about 5,000 ft southwest of the Site without mapped faulting but likely striking to the northwest because of faulting (fig. 1.1A). Alternate, unpublished mapping from Joseph Smoot (U.S. Geological Survey, written commun., 2005) shows this unit southwest of the NP1 Site is part of a continuous gray-shale unit with structure apparently deformed in the vicinity of the Chalfont Fault (fig. 1.1B)

Borehole Deviation

Borehole deviation logs may provide some information about orientation of geologic structure. Brown and others (1981) note that field observations frequently indicate that gradual borehole deviation in one direction is associated with anisotropy of the rock properties (such as bedding or schistosity) and abrupt changes in deviation direction are associated with sudden changes in rock properties (such as faults, unconformities, or large differences in lithology). In some cases where bedding dips less than 45 degrees, it has been reported that the direction of borehole deviation tends to migrate up dip (normal to bedding planes); however, where dips of bedding are steeper, the direction of borehole deviation tends to migrate down dip (McLamore, 1971; Wilson, 1976; Brown and others, 1981). Near the NP1 Site, the deviation logs for numerous wells completed in the Brunswick and Lockatong Formations about 6 miles to the southeast of Souderton had a linear trend normal to bedding (Senior and others, 2008). Plots showing borehole deviation logs collected by USGS for seven wells as part of NP1 Site investigations are shown in figures 1.3–1.9.

The borehole deviation logs for three wells nearest the center of the NP1 Site (MG 668 [GKM], MG 2202 [S-2], and MG 2220) generally are linear or curvilinear and mostly trend to the northeast (figs. 25, 1.3, 1.6, and 1.9). However, the deviation log for well MG 2220 initially trends north from land surface to about 30 ft bls before changing directions (fig. 1.9), which may indicate the presence of faulting. Broken and apparently twisted rock possibly indicative of faulting were apparent on the borehole video logs for well MG 2220 at depths from about 31 to 34 ft bls (fig. 22). The deviation log for well MG 2201 (S-1) also has a linear section that trends to the northeast, but the deviation abruptly changes direction at about 35 ft bls to trend to the west and northwest (fig. 1.5). These four wells are near and straddle a stream reach that trends northeast-southwest with a strike of about N50°E. In contrast to these four wells, the borehole deviation logs for three wells southwest and farther from the center of the NP1 Site (MG 2203 [S-3], MG 2204 [D-3], and MG 679 [S-11]) are also linear and curvilinear, with some abrupt changes, that trend in different directions ranging from the northwest to the southeast (figs. 25, 1.3, 1.7, and 1.8); these three wells are near, and straddle, the stream reach that trends north-northeast-south-southwest, with a strike of about N20°E.

Possible interpretations of these deviation logs include probable presence of faulting, indicated by abrupt changes in direction of deviation, and general differences in deviation direction between wells nearest the center of the NP1 Site and wells to the southwest of the NP1 Site that likely are related to spatial differences in orientation of bedding or fractures.

Geophysical-Log Correlation

Gamma and single-point resistance logs were examined to determine differences in lithology that could be correlated among wells and possibly used to infer structure. The vertical and horizontal extents of lithologic data that can be used for correlation are limited because there are relatively few wells at the NP1 Site, and four of the eight wells with logs are relatively shallow (49–82 ft bls in depth). A few distinctive natural gamma features but no distinctive single-point resistance features were present on available logs for eight wells at the NP1 Site. The lack of distinctive single-point resistance features in the logs indicates small differences in lithology (sand or shale component) in intervals of rock intercepted by the wells. The natural gamma logs for all eight wells are shown on a section line discussed in the section “Conceptual Model of the Groundwater System.”

Elevated natural gamma marker beds were present on geophysical logs for the 300-ft-deep well MG 665 (S-9) collected in 1992 (figs. 6 and 1.2) but not in logs for any of the other seven nearby wells at the NP1 Site. Elevated natural gamma activity from radioactive minerals in relatively thin (commonly less than 5 ft in thickness) black shale beds in the Lockatong Formation and some sections of the Brunswick Formation have been used as markers for correlating geologic units at other sites in Montgomery and Bucks Counties (Barton and others, 2003; Senior and others, 2008). The lack of elevated natural gamma marker beds that could be used for correlation in most of the wells at the NP1 Site, other than well MG 665 (S-9), may be related to differences in well depth, disruptions in geologic structure, or lateral depositional discontinuities, such that these marker beds are not intercepted in these other wells.

Other distinctive lithologic features that could be potentially correlated include the 5- to 10-ft-thick intervals of slightly elevated natural gamma activity associated with grayish-green beds near depths of 135 and 155 ft bls in well MG 2204 (D-3) (figs. 18 and 20). However, like the thin zones of elevated natural gamma activity in well MG 665 (S-9), these features do not appear on logs for nearby wells (MG 2203 [S-3], MG 679 [S-11], and MG 2204 [D-4]) in the southwest part of the study area (fig. 2), thus providing additional evidence for lack of lateral continuity in bedding that may indicate local disruption in structure, such as faulting or indicate local-scale variability in depositional environment that results in some beds pinching out.

In the three closely spaced wells in the northeast part of the study area, one to two zones with slightly elevated natural gamma activity are present on logs at depths of about 18–52 ft bls in well MG 668 (GKM), 48 ft bls in well MG 2201 (S-1), and 18–52 ft bls well MG 2220 (figs. 9, 12, and 21). The elevated gamma units are not present in all logs, possibly because of lateral changes in bed thickness or structural displacement (faulting). If these zones can be correlated (fig. 26), the strike of the dipping beds would be consistent with the northeast trend of well distribution parallel to topography and the stream (fig. 2), and any displacement along a possible fault at about 33 ft bls in well MG 2220 as shown by images of crushed and twisted rock in the borehole-video log (fig. 22) would be minor.

Fracture Orientation

The optical-televiewer and ATV logs provided data on the depth and orientation of fractures in each monitoring well. All wells encountered numerous fractures. The most notable characteristic was the large frequency of high-angle fractures, because it is more likely for a vertical well to penetrate a low-angle bedding-plane fracture than a high-angle fracture. The frequent occurrence of high-angle fractures penetrated by wells is consistent with the observation of numerous high-angle fractures at outcrops in the unnamed tributary of Skippack Creek between Wile Avenue and West Street.

Fractures were most notable in well MG 2220, which was drilled through a zone that was extensively broken with high-angle fractures, possibly related to faulting. The 82-ft-deep well encountered the highly fractured zone from 25 to 60 ft bls, and still images from the borehole video of well MG 2220 show broken rock and high-angle planar fractures in that interval (fig. 22). The orientation of the fractures above 60 ft bls in well MG 2220 is difficult to discern from the ATV log because of the density of the fracture openings, but two high-angle fractures below 60 ft bls appear to dip to the southeast (figs. 21 and 23). Depending on its extent, this fracture zone or fault zone could provide an avenue for contaminant movement.

High-angle fractures also were predominant in logs of wells MG 2201 (S-1) and MG 2204 (D-3) and present, but not predominant, in logs of well MG 679 (S-11), MG 668 (GKM), MG 2202 (S-2), and MG 2203 (S-3). Where discernable from ATV logs, the high-angle fractures were oriented mostly about north-south in wells MG 2201 (S-1), MG 2202 (S-2), and MG 2203 (S-3). In contrast to those wells, the high-angle fractures appear to dip to the southwest in well MG 679 (S-11) and below 120 ft bls in well MG 2204 (D-3) and to southeast above 120 ft bls in well MG 2204 (D-3). The difference in fracture orientation above and below 120 ft bls in well MG 2204 (D-3) may be related to the presence of a small fault, as evidenced by an abrupt change in direction at about 95 ft bls in the borehole deviation log (fig. 1.8). Large, fractured openings were evident from about 110–113 ft bls as shown by the caliper, ATV, and borehole video logs (figs. 18 and 19). From review of video logs of well MG 668 (GKM), the high-angle fractures at same or similar depths are oriented in different directions, complicating interpretation of the ATV logs.

The low-angle fractures shown in ATV image logs, interpreted as bedding-plane partings, were predominant in logs of wells MG 668 (GKM), MG 2202 (S-2), and MG 2203 (S-3). The low-angle fractures do not show a consistent orientation.

Water-Level Fluctuations

The timing and magnitude of water-level fluctuations were similar among monitoring wells, except at well MG 668 (GKM), which frequently exhibited large, rapid rises following storms (fig. 27A). Seasonal changes in water levels were observed in all wells. The levels for all wells except well MG 668 (GKM) fluctuated less than about 4 ft during the period of study. Annual fluctuations ranged from about 1 to 2 ft in wells MG 2203 (S-3), MG 2204 (D-3), and MG 679 (S-11), which are in the southwest part of the NP1 Site, and from about 3 to 4 ft in wells MG 2201 (S-1), MG 2202 (S-2), and MG 2220, which are in the northeast part of the NP1 Site (figs. 2 and 27A). Transient water-level fluctuations of up to about 20 ft in wells MG 668 (GKM) were superimposed on seasonal water-level fluctuations that were similar to those in well MG 2220.

Water levels are shown as depth below land surface (fig. 28A) and altitude referenced to the North American Vertical Datum of 1988 (NAVD 88) (fig. 28B) for the period March through September 2017. Water-level altitudes represent the composite value of hydraulic heads from discrete water-yielding fractures throughout each open hole. Among wells MG 2203 (S-3), MG 2204 (D-3), and MG 679 (S-11) in the southwest part of the study area, the water-level altitudes during the 2017 monitoring period were equal or very similar despite differences in depths to water below land surface (figs. 28A). The water-level altitudes in these wells were also about 15–20 ft lower than those of the source area in the northeast part of the NP1 Site. The water-level altitudes were highest Site-wide in wells MG 668 (GKM) and MG 2202 (S-2) in 2017 (fig. 28B). Both wells generally had similar water-level altitude despite differences in depths to water below land surface, except for transient storm-related rises in well MG 668 (GKM). The water-level altitudes were nearly equal in the closely spaced (about 56 ft apart) well MG 2220 and the extraction well MG 2201 (S-1), which was being pumped almost continuously during the monitoring period. The accuracies of the water-level altitudes were calculated using land-surface altitudes determined by lidar (table 1) and estimated to within 1 ft.

Water levels in well MG 2204 (D-3) remained above land surface for the entire 2017 monitoring period, ranging from 2.8 to 4.7 feet above land surface (figs. 27A and 28A). Well MG 2204 (D-3) is an artesian well and overflows the casing if unsealed, but it has a cap with an expandable seal that could be temporarily removed to allow installation of a transducer to measure water pressures for this study. Water pressures measured using a transducer inside the well casing are converted to equivalent hydraulic heads. The hydraulic head of this shut-in well does not represent the water-table altitude but is a composite value of hydraulic heads in numerous water-yielding fractures throughout the open hole from 15 to 198 ft bls. The well shows the overall upward vertical hydraulic gradient near the unnamed tributary to Skippack Creek at this location.

The rapid changes in water levels in well MG 668 (GKM) were apparently affected by water present in the well pit at the top of the surface casing. From March 23 through August 23, 2017, rapid water-level rises of 10–15 ft in a few hours were recorded at well MG 668 (GKM) followed by relatively rapid water-level declines (fig. 28A). Each rapid water-level rise in well MG 668 (GKM) peaked at about 5 ft bls during that period, which is similar to the approximate depth to the opening in the well casing below the concrete cover of the well pit that allows water in the pit to drain into the well. Thus, it is probable that the rapid water-level rises were mostly caused by surface water in the pit draining into well MG 668 (GKM). Some rapid water-level rises peaked at values as high as about 3 ft bls during the period from spring 2015 to spring 2016 (fig. 27A), indicating the presence of standing water in the pit above the top of the casing. Although well MG 668 (GKM) was reconstructed in October 2017 to prevent surface water in the pit from entering the well, water could still enter the pit itself after October 2017. A water-level measurement during a site visit on May 25, 2018, showed the depth of water in the well was 9.01 ft bls, and the water standing in the pit was 5.29 ft bls. This difference indicates that water in the pit no longer has direct entry to the well. However, a video log collected when water was standing in the pit on July 28, 2018, documented that pit water can slowly leak into the well after the pit fills up during storms (fig. 5D).

The large, rapid water-level rises recorded in well MG 668 (GKM) before reconstruction in October 2017 are not present in the hydrographs from the other wells that were monitored (fig. 27A and 28A). The lack of correlating sharp water-level increases in two wells (MG 2201 [S-1] and MG 2220) with good hydraulic connection to well MG 668 (GKM) is consistent with surface-water inflow as the cause of rapid water-level rises in well MG 668 (GKM) before reconstruction in October 2017. The hydraulic connections among these three wells are indicated by water-level response in well MG 668 (GKM) to pumping of MG 2201 (S-1) and MG 2220, as discussed in the section “Hydraulic Connections.” After reconstruction of well MG 668 (GKM) in October 2017, the water-level response to precipitation changed in magnitude and pattern (fig. 29) from the period prior to reconstruction (figs. 27 and 28). As discussed in section “Reconstruction of Well MG 668 (GKM),” water can leak into the reconstructed well MG 688 (GKM) through the joint between existing and extended casing. The observed water-level record for 2018 shows that water levels in well MG 688 (GKM) can rise to nearly land surface and seem to fall slowly (fig. 29A), indicating infiltration of surface-water into the well and low transmissivity of fractures in the open interval of the well below the casing’s bottom after grouting (9.4–35.4 ft bls); the difference in pre- and post-reconstruction water-level response to precipitation was caused by grouting up the deeper, more transmissive fractures in well MG 688 (GKM). Although the record is affected by several periods when water levels declined below the transducer at about 19.7 ft bls in 2018 (fig. 29A), water-level altitudes in well MG 668 (GKM) are generally higher in 2018 after reconstruction (fig 29B) than before reconstruction in October 2017 (fig. 28B).

Hydraulic Connections

Continuous monitoring of water levels was used to identify hydraulic connections among the wells. Hydraulic connections among wells near the source area were indicated by water-level responses (declines) to pumping from wells MG 2201 (S-1) and MG 2220. The increase of the pumping rate in well MG 2201 (S-1) on April 19, 2016, caused water levels to decline in wells MG 2201 (S-1) and MG 668 (GKM), which is 183 ft to the northeast of well MG 2201 (S-1) (figs. 2 and 30A). The water-level response in the well drilled on April 6, 2016 (MG 2220), to an increased pumping rate in well MG 2201 (S-1) on April 19, 2016, is not known because no water-level data were collected in well MG 2220 on that date. Pumping of well MG 2220 at a rate of about 1 gal/min during flow-meter logging on August 22, 2016, caused drawdown in wells MG 2201 (S-1) and MG 668 (GKM) (fig. 30B). No apparent response was observed in the other wells with water-level monitoring that were farther away from well MG 2220 and the source area. Pumping of well MG 2220 at a rate of about 4.4 gal/min as part of purging prior to sampling on May 2, 2018, caused a drawdown of 1.7 ft in well MG 2220 and 1.1 ft in well MG 2201 (S-1), but no water-level response was observed in reconstructed well MG 668 (GKM) as shown by discrete manual measurements of water levels in well MG 2220 and continuous water-level data for wells MG 2201 (S-1) and MG 668 (GKM) (fig. 30C). These data indicate that the reconstruction of well MG 668 (GKM) in 2017 likely closed off the hydraulic connection between wells MG 2220 and MG 668 (GKM).

Hydraulic connections were also indicated by water-level responses (rises) to pumping shutdowns. The temporary cessation of pumping in MG 2201 (S-1) on October 16 and 18, 2016, for 45 and 165 minutes, respectively, caused water levels to rise in wells MG 668 (GKM) and MG 2220 (fig. 31A). When extraction well MG 2201 (S-1) was shut down May 15–16, 2018, the water level quickly rose about 5 ft in well MG 2201 (S-1) and 4 ft in well MG 2220 (fig. 31B). The rapid response and similar magnitude of rise indicates a strong hydraulic connection between the two wells. However, in contrast to effects of pumping in 2016, the May 2018 shutdown of extraction well MG 2201 (S-1) did not affect water levels in the reconstructed well MG 668 (GKM), which had been backfilled to a depth of 34.4 ft bls in October 2017. Prior to reconstruction, when well MG 668 (GKM) was open to at least 189 ft, changes in pumping rate at the 59-ft-deep well MG 2201 (S-1) affected the water level of well MG 668 (GKM) (fig. 30A). Thus, it appears that the well MG 668 (GKM) reconstruction probably closed off the hydraulic connection between those wells; the hydraulic connections(s) between the two wells likely through fractures in the interval from 35 to 60 ft bls in well MG 668 (GKM) and through the fracture near 51–55 ft bls in well MG 2201 (S-1).

Historical water-level data also have documented hydraulic connections among wells. Drawdown in wells during pumping tests or during operation of production wells at or near the NP1 Site has been noted in the past. Observations on these hydraulic connections are shown in figure 32 and summarized below:

Water-Level Contours and Gradients

Although there are not enough wells to provide water levels sufficient for reliable mapping of the water-table surface for the NP1 Site, preliminary potentiometric contours based on water-level altitudes in seven wells are presented on a map (fig. 32) to better represent recent (2018) conditions than the maps prepared during the RI (CH2M-Hill, Inc., 1993a, fig. 3-8) or as part of 2016–17 monitoring reports to PADEP (AECOM, 2016, fig. 3; 2017, fig. 2), which were also based on limited data. The available monitoring wells are aligned northeast-southwest along the stream, but additional well coverage would be needed to the southeast and northwest of existing wells. Additionally, water-levels in open holes with multiple water-bearing fractures represent composite hydraulic heads, such water-level data from the available wells, which range in depth as open holes from about 34 to 300 ft bls may be too limited to accurately define both horizonal and vertical groundwater gradients. Well depths range from about 34 to 60 ft bls for the five wells northeast of West Street and from about 208 to 300 ft bls for the two wells southwest of West Street (tables 1 and 2; figs. 2 and 33). Factors including topography, pumping, recharge, and permeability, and geologic structure can affect the distribution of horizontal and vertical gradients. Thus, additional monitoring wells at different depths and open to shorter sections of the aquifer may be needed for reliable potentiometric maps.

Assuming the stream is a sink (area of natural discharge) for shallow groundwater, an approximation of the water-table contours is configured for measurements made on April 24, 2018, as shown in figure 33. The map was drawn to show a horizontal hydraulic gradient toward the stream and the influence of pumping from extraction well MG 2201 (S-1) on the water-table configuration. Water-level altitudes in wells MG 2204 (D-3) and MG 679 (S-11) in the southwest part of the NP1 Site are composite head values for deep wells that probably are higher than the water table. The contours in figure 33 show one possible configuration of the water table. The configuration could be tested if additional data become available from new monitoring wells, especially if those are not aligned between the existing monitoring wells.

Groundwater Monitoring

The PADEP has conducted long-term monitoring of VOC concentrations in groundwater by sampling wells MG 668 (GKM), MG 2201 (S-1), MG 2202 (S-2), MG 2203 (S-3), and MG 2204 (D-3) semiannually to quarterly since 1999 (AECOM, 2012a, b; 2013; 2014; 2015; 2016; 2017; 2018; U.S. Environmental Protection Agency, 2023). The five wells range in depth from about 50 to 208 ft bls (tables 1 and 3). The USGS sampled the 82-ft-deep well MG 2220 once in 2018. All sampled wells are open boreholes below the casing’s bottom and have more than one water-bearing fracture, such that samples may represent a composite of water quality from the contributing water-bearing fractures. Factors that can affect results of sampling include differences in sampling methods and varying contributions from multiple water-bearing fractures, and, for low-flow sampling, depth of pump intake.

Long-Term Monitoring, 1999–2022

Analytical results of groundwater samples collected for the PADEP from monitoring wells are shown for PCE (fig. 32) and TCE (fig. 33) for the period 1999–2022. The data for 1999–2017 are from annual operation and monitoring reports (AECOM, 2012a, b; 2013; 2014; 2015; 2016; 2017; 2018) and were compiled in 5-year review reports by the U.S. Environmental Protection Agency (2008a, 2013, 2018a, 2023). The data for 2018–22 are from annual operation and monitoring reports as summarized in U.S. Environmental Protection Agency (2023); these reports use the names assigned by the EPA to identify wells at the NP1 Site (table 1). The sampling methodology as described in these reports varies among wells; the methodology described in AECOM (2017) states that samples were collected after low-flow purging for wells MG 2202 (S-2), MG 2203 (S-3), and MG 668 (GKM), or collected directly from sample ports for the flowing artesian well MG 2204 (D-3) and for the active extraction well MG 2201 (S-1). Samples have been collected quarterly (in March or April, June, September or October, and December) from well MG 2201 (S-1) since 2008 when pumping for remediation began in this well. For the monitoring wells, samples were collected semiannually (in June and December) (U.S. Environmental Protection Agency, 2023).

The highest concentrations of PCE and TCE were found in well MG 2201 (S-1) (fig. 2). Concentrations of PCE in well MG 2201 (S-1) were 450–650 µg/L from September 1999 to June 2002, then increased to 2,200–8,300 µg/L from December 2002 to December 2004 and declined abruptly after pumping from well MG 668 (GKM) ceased in February 2005 (fig. 35A). Concentrations gradually increased from February 2005 until June 2014, then began a general decreasing trend to December 2017, increasing again from 2018 through April 2020, followed by generally lower levels (less than 350 µg/L) through June 2022, apart from a value of 998 µg/L in December 2021. The reasons for the reported high concentrations in 2002–04 and a low concentration (28.4 µg/L) in December 2020 are not known. TCE concentrations in well MG 2201 (S-1) followed the same general pattern (fig. 36A).

The large decreases in PCE and TCE concentrations in water from extraction well MG 2201 (S-1) from December 2004 to June 2005 were probably related to the shutdown of extraction well MG 668 (GKM), but differences in precipitation during the period may also have had an effect. The drought of 2001–02 was followed by a wet period in 2003–04, as indicated by water levels in nearby long-term observation well MG 917 (figs. 2 and 37).The elevated PCE concentrations during the wet years, could have been caused by flushing from increased groundwater recharge, and there is generally an inverse relation between PCE concentrations in well MG 2201 (S-1) and depth to water in observation well MG 917 (fig. 37A). The gradual increase in PCE concentrations from 2007 to 2012 followed a period of slightly greater than normal precipitation (National Oceanic and Atmospheric Administration, 2024a) that resulted in higher recharge and groundwater levels (fig. 37). PCE concentrations generally decreased in 2015–17 during a period of lower groundwater levels in well MG 917 and below average precipitation; concentrations increased in 2018–20 during a period of higher groundwater levels in well MG 917 and above average precipitation (fig. 37). These patterns indicate precipitation and related recharge may have affected contaminant concentrations in groundwater. Other factors that may affect reported concentrations are changes in sampling methodology for well MG 2201 (S-1) before and after pumping began in this well in September 2008; samples from extraction well MG 2201 (S-1) have been collected under operational pumping conditions since September 2008, but low-flow methods used to sample other monitoring wells since 2009 (AECOM, 2017) may have been used to sample MG 2201 (S-1) before September 2008.

Concentrations of PCE and TCE for the other monitoring wells are shown in figures 35B and 36B. Concentrations of PCE in all wells have generally decreased since extraction well MG 2201 (S-1) began operation in September 2008. PCE concentrations were below 5 µg/L in samples from well MG 668 (GKM) except for samples in June of most years before well reconstruction in October 2017; after October 2017, PCE concentrations in samples from well MG 668 (GKM) did not show a consistent pattern, being above 5 µg/L in December 2019 and December 2020 samples. Factors causing the apparent sawtooth pattern of concentrations in samples from well MG 668 (GKM) are not known, but PCE concentrations generally are higher in June samples than in December samples each year until 2017, indicating factors affecting PCE concentrations may be seasonal. A temporal change in pattern appears after reconstruction of well MG 668 (GKM) and indicates some effects related to reconstruction of this well. The sawtooth pattern does not seem to relate to differences in precipitation or groundwater level. The opposite pattern is found in samples from wells MG 2202 (S-2) and MG 2203 (D-3): PCE concentrations were lower in June and higher in December during 2009–18. However, the differences in concentration between June and December samples were much smaller for these wells than for well MG 668 (GKM). TCE concentrations were less than 5 µg/L in all samples and had a similar sawtooth pattern over time. Concentrations were higher in June than in December for samples from wells MG 668 (GKM; until reconstruction in October 2017), MG 2202 (S-2), and MG 2204 (D-3); after reconstruction of well MG 668 (GKM) in October 2017, the temporal pattern in TCE concentrations in samples from this well changed similar to that observed for PCE.

The VOC cis-1,2-dichloroethene at relatively low levels (less than 10 µg/L) was found in samples from wells MG 668 (GKM), MG 2201 (S-1), and MG 2204 (D-3), and was the predominant form of dichloroethene (DCE) in the samples. The cis-1,2-DCE is probably an indication of biodegradation of VOCs in the groundwater. Guidance for evaluation of natural attenuation of chlorinated solvents notes that if the concentration of cis-1,2-DCE is greater than 80 percent of total DCE, it was probably formed by dehalogenation of TCE or PCE (Minnesota Pollution Control Agency, 2006).

Sampling method may have affected the reported concentrations of VOCs, especially in open boreholes containing multiple water-bearing fractures. This may introduce some uncertainty in interpretation of results or comparison of results for samples collected using different methods. Potential issues associated with low-flow sampling from the 6-in.- and 8-in.-diameter open holes include

• • groundwater samples could contain a substantial amount of water from borehole storage rather than water directly entering well from an active water-bearing fracture;

• • contaminant concentrations of the water in borehole storage could be altered by chemical reactions or biological activity; and

• • groundwater samples from the open holes may represent a mixture of concentrations in water from multiple water-bearing fractures and (or) from fractures nearest the sample collection depth within the well, such that sample depth is another variable that could affect results.

In the cases of wells MG 2201 (S-1) and MG 2204 (D-3), groundwater samples were collected from pumping or flowing wells, respectively, and thus explicitly represent a mixture of water from different water-bearing fractures. The PCE concentrations in groundwater samples from the 35-ft-deep MG 668 (GKM) after reconstruction in October 2017 (less than 10 µg/L) are much lower than the value of 330 µg/L reported for packer testing of the 0–28 ft interval during RI (CH2M-Hill, Inc., 1993a), despite apparently similar interval depths. Differences in PCE values after 2017 compared to the RI packer-test results could be partly related to (1) sampling methods because low-flow sampling was used after 2017 and the interval was pumped for three volumes before sampling for packer tests during the RI, (2) differences in interval depths and water-bearing zones in MG 668 (GKM) after reconstruction in October 2017 and the reported packer test interval of 0–28 ft in this well, and (or) (3) soil removals in 1998. The 2012 borehole video shows likely water-bearing fractures from about 29 to 33 ft bls, which would reduce the ability to obtain a packer seal at 28 ft bls. These fractures could be the same as those noted at about 22 ft in well MG 668 (GKM) in the RI report (CH2M-Hill, Inc., 1993a) but offset by the approximate depth from land surface to top of casing (about 5.9 ft bls), such that the interval from of 0–28 ft is below top of casing or about actually 0–34 ft bls. The 2018 borehole video shows that the water-bearing fractures from about 29 to 33 ft bls are less open than in the 2012 video and may have been partly obstructed by grouting during reconstruction.

MG 2220 Sampling, 2018

A groundwater sample was collected from well MG 2220 by the USGS on May 2, 2018, after purging more than three well volumes and was analyzed for 34 VOCs at the USGS National Water Quality Laboratory. Results of the analysis and equipment blank are shown in table 4. Two compounds, PCE and TCE, were detected at concentrations greater than their EPA MCLs of 5 µg/L for drinking water. The PCE and TCE concentrations in the sample collected from well MG 2220 on May 2, 2018, were 1,830 µg/L and 9.8 µg/L, respectively. The concentrations of PCE and TCE in the May 2018 sample from well MG 2220 were about four times and three times greater, respectively, than the PCE and TCE concentrations of 444 µg/L and 3 µg/L measured in a sample from extraction well MG 2201 (S-1) collected on December 15, 2017. Additional measurements of PCE and TCE concentrations in groundwater samples from well MG 2220 may be needed to determine if elevated concentrations are persistent, which would support consideration of use of this well for monitoring or as an extraction well as part of remediation.

Table 4.    

Field parameters and 34 volatile organic compounds analyzed in an equipment blank collected on April 30, 2018, at 3:00 pm and a water sample collected from well MG 2220 on May 2, 2018, at 2:00 pm.

[°C, degree Celsius; —, no data; µS/cm, microsiemens per centimeter; ft bls, feet below land surface; mg/L milligrams per liter; µg/L, micrograms per liter; <, less than]

Table 4.    Field parameters and 34 volatile organic compounds analyzed in an equipment blank collected on April 30, 2018, at 3:00 pm and a water sample collected from well MG 2220 on May 2, 2018, at 2:00 pm.

Parameter or volatile organic compound	Unit	Results
		Equipment blank	Well MG 2220
Field parameters
Water temperature	°C	—	15.4
Specific conductance	µS/cm at 25 °C	—	933
pH	standard units	—	6.1
Dissolved oxygen	mg/L	—	0.7
Depth to water	ft bls	—	23.97
Volatile organic compounds
Bromodichloromethane	µg/L	<0.1	<0.1
1,2-dichloroethane	µg/L	<0.2	<0.2
Bromoform	µg/L	<0.2	<0.2
Dibromochloromethane	µg/L	<0.2	<0.2
Trichloromethane	µg/L	<0.1	10.2
Toluene	µg/L	<0.1	<0.1
Benzene	µg/L	<0.1	<0.1
Chlorobenzene	µg/L	<0.1	<0.1
Ethylbenzene	µg/L	<0.1	<0.1
Dichloromethane	µg/L	<0.2	<0.2
Tetrachloroethylene (PCE)	µg/L	<0.1	21,830
Trichlorofluoromethane	µg/L	<0.2	<0.2
1,1-dichloroethane	µg/L	<0.1	<0.1
1,1-dichloroethylene	µg/L	<0.1	10.1
1,1,1-trichloroethane	µg/L	<0.1	10.4
1,2-dichlorobenzene	µg/L	<0.1	<0.1
1,2-dichloropropane	µg/L	<0.1	<0.1
Trans-1,2 dichloroethylene	µg/L	<0.1	10.2
1,3-dichlorobenzene	µg/L	<0.1	<0.1
1,4-dichlorobenzene	µg/L	<0.1	<0.1
Dichlorodifluoromethane	µg/L	<0.2	<0.2
Vinyl chloride	µg/L	<0.2	<0.2
Trichloroethylene (TCE)	µg/L	<0.1	29.8
Ethyl tert-butyl ether	µg/L	<0.1	<0.1
Tert-pentyl methyl ether	µg/L	<0.2	<0.2
Cis-1,2-dichloroethene	µg/L	<0.1	16.5
Styrene	µg/L	<0.1	<0.1
O-xylene	µg/L	<0.1	<0.1
1,1,2-trichlorotrifluoroethane	µg/L	<0.1	<0.1
Tert-butyl methyl ether	µg/L	<0.2	<0.2
Diethyl ether	µg/L	<0.2	<0.2
Diisopropyl ether	µg/L	<0.2	<0.2
M- and p-xylene	µg/L	<0.2	<0.2
Tetrachloromethane	µg/L	<0.2	<0.2

¹

Detections

²

Concentrations of detections exceeding a maximum contaminant level set by the U.S. Environmental Protection Agency (2018b).

The water sample from well MG 2220 was collected using a Grundfos RediFlo-2 pump after purging the well for 90 minutes at an average rate of 4.4 gal/min. The purging caused drawdown in well MG 2220 of about 1.7 ft after about 90 minutes (fig. 29C), from which a specific capacity of about 2.6 gallons per minute per foot of drawdown was calculated. Drawdown of about 1.1 ft was measured in nearby well MG 2201 (S-1) (figs. 2 and 29C), indicating interconnections between the wells. These data could potentially be used to estimate transmissivity of the aquifer, although the approach would need to account for fractured-rock properties.

Surface-Water Monitoring

Surface water has been sampled semiannually (in June and December) for VOCs since December 2011 at two sites (fig. 2) on the unnamed tributary to Skippack Creek as part of monitoring by PADEP. Of the three VOCs of concern, PCE, TCE, and cis-1,2-DCE, only PCE was measured above the reporting level of 0.5 µg/L (AECOM, 2012a, b; 2013; 2014; 2015; 2016; 2017; 2018; U.S. Environmental Protection Agency, 2023). PCE has been consistently found at the upstream site (SW-2) at concentrations ranging from 1.3 to 3.4 µg/L during December 2011 through December 2017 and then at generally lower and more variable concentrations since June 2018; however, the second highest reported value of 2.6 µg/L was for the December 2021 sample (fig. 38). Contamination in the shallow groundwater system could be discharging in the stream near SW-2 and upstream. At the downstream site (SW-1), PCE has only been detected three times: twice in estimated concentrations (less than the reporting limit of 0.5 µg/L) (December 2014, 0.44 µg/L; June 2020 0.12 µg/L), and once at 1.4 µg/L (December 2018) (AECOM, 2012a, b; 2013; 2014; 2015; 2016; 2017; 2018; U.S. Environmental Protection Agency, 2023).