A review of available historical PFAS data within the studied watersheds was completed (tables 1, 1.3–1.5). Database queries were limited to publicly available PFAS data in surface and groundwater samples collected within the Passaic River Basin. Analytical methodology, when reported, was obtained through queries or provided with requested information, and not a search criterion. Querying the USGS Water Data for the Nation website (USGS, 2026) resulted in 71 PFAS sample results for 21 surface-water sites and 13 groundwater sites in the study area watersheds (table 1.3). Wastewater treatment plant (WWTP) data from the Passaic River Basin and submitted voluntarily to the NJDEP were obtained through an Open Public Records Act electronic request (https://www.nj.gov/opra/). The request resulted in data from one WWTP in the basin and included data for one influent site, one effluent site, and one effluent surface-water site (table 1.4). The NJDEP New Jersey Pollutant Discharge Elimination System (NJPDES) permitting program results were obtained by querying the NJDEP DataMiner database, and the search was limited to calendar years 2020–24. Results for two discharges to groundwater and one general remediation site were obtained (table 1.5; NJDEP, 2024a). Results of these queries and requests are detailed in table 1, and PFOA and PFOS results are summarized in the “Results” section.

The Passaic River Basin is in northern New Jersey and southern New York and includes more than 2,400 square kilometers (km²) and a population greater than 2 million people according to the U.S. Census Bureau (2020). Predominant land cover within the basin is about 44 percent developed (including Developed Low, Medium, and High Intensity), and 18 percent was classified as “Developed, Open Space,” 39 percent forest (including Deciduous Forest, Evergreen Forest, Mixed Forest, Shrub/Scrub, and Herbaceous), and 12 percent wetlands (including Woody Wetlands and Emergent Herbaceous Wetlands) (Dewitz, 2023). Within the Passaic River Basin, sampling sites for this study fall within four 10-digit hydrologic unit code (HUC10) scale watersheds: Passaic River (750.96 km²), Wanaque River (274.44 km²), Ramapo River (417.03 km²), and Pompton River (289.32 km²) (USGS, 2025).

Four geospatial scales were used in this study and are shown in figure 1:

A total of 37 sampling sites (table 2) were selected within the nontidal Passaic River Basin in collaboration with the NJDWSC and included 5 sites in the upper portion of the Wanaque River watershed above the Wanaque Reservoir, 9 sites in the Ramapo River watershed, 8 sites in the Pompton River watershed, and 15 sites in the Passaic River watershed (fig. 2). The sampling sites were comprised of tributaries and main-channel river locations. Selection of sites downstream from known point sources, sites at the confluence of major tributaries, and sites near drinking-water pump stations was prioritized (fig. 2). Additional considerations in site selection included ease of access to the location for sampling, safety, and the ability to collect a streamflow measurement or to leverage an existing USGS streamgage.

Surface-water-quality samples were collected at 37 sites in the Wanaque, Ramapo, Pompton, and Passaic River watersheds in January, March, July, and September 2025 as close to base-flow conditions as possible. The USGS defines “base flow” as, “sustained, low flow in a stream” (USGS, 2014). To determine when sampling should occur, weather radar and USGS hydrographs within the study area were monitored for periods of no rain and sustained lower flow. In addition to base-flow sampling, two additional sampling events targeted higher stream flows. In July 2025, 27 sites were sampled after a rain event when the rivers were just past peak streamflow; this sampling took place a week after a regularly scheduled base-flow sampling event. In December 2025, samples were collected from nine main-stem river sites during the rising limb of the hydrograph during a rain event. Samples were only collected at sites that could be safely accessed for sample collection and (or) streamflow measurements under existing weather conditions. Sampling event information is summarized in table 3. Refer to table 1.1 and Romanok and others (2026) for more sampling details.

Water-quality samples were collected directly into the sample containers (grab samples) from the center of flow using standard methods: while wading or from a bridge as conditions permitted (USGS, 2018). A clean stainless-steel weighted-bottle sampler attached to a clean length of polyethylene rope was used for sample collection from a bridge. Surface-water samples were collected for PFAS and total organic carbon (TOC) during all six sampling events. At the request of the cooperator, opportunistic samples were collected for trace elements during three base-flow samplings (March, July, and September) and for 1,4-dioxane during the March base-flow sampling event (table 4; Romanok and others, 2026).

For PFAS analysis, whole-water samples were collected into two pre-cleaned 500-milliliter (mL) high-density polyethylene (HDPE) screw-cap bottles after three rinses with stream water, placed in a cooler on ice in the field, then transported to the New Jersey Water Science Center where they were stored frozen at −20 degrees Celsius (°C) until shipment to Eurofins Cleveland Laboratory (Barberton, Ohio). Samples for TOC were collected in 125-mL pre-cleaned, baked, amber-glass bottles and preserved with 2 mL of 1:1 OmniTrace hydrochloric acid (samples were not field rinsed). TOC samples were chilled on ice in the field until returned to the New Jersey Water Science Center and then refrigerated until their transport to the NJDWSC laboratory, Wanaque, New Jersey, where analysis was performed. Surface-water samples for trace elements were collected in a 500-mL HDPE screw-cap bottle, preserved with 2 mL of 7.7 normality nitric acid, placed on ice in the field until returned to the New Jersey Water Science Center and then refrigerated until their transport to the NJDWSC laboratory for analysis. Samples for 1,4-dioxane were collected in 500-mL pre-cleaned, baked, amber-glass bottles prepared with sodium sulfite and sodium bisulfate preservative, placed on ice in the field until returned to the New Jersey Water Science Center and then refrigerated until being shipped on ice to ALS Environmental Laboratories (Middletown, Pennsylvania) where analysis was performed. Field water-quality parameters were recorded in situ during the sampling events using a YSI EXO2 multi-parameter sonde. This included water temperature, dissolved oxygen, pH, specific conductivity, and turbidity. During the July rain event only, turbidity was alternatively measured at 13 sites with a Hach benchtop turbidimeter (nephelometric turbidity ratio units) from a grab sample collected in the stream.

Streamflow at the time of sample collection was determined using one of several methods. If the sampling site was at an active USGS continuous discharge streamgage, the approved computed streamflow from the streamgage at the time of the sample collection was used. If the sampling site was at a USGS station without a streamgage, a discrete discharge measurement was made using an acoustic Doppler current profiler or acoustic Doppler velocimeter to determine the streamflow at the time of sample collection. If a discrete measurement of discharge at an ungaged site was not made, the streamflow associated with the sample was calculated by adjusting the calculated discharge from an active USGS continuous records streamgage to the sampling site by the ratio of the drainage area of the streamgage to the drainage area of the sampling site. This method has an inherent increase in uncertainty because it relies on several assumptions, for example, that the watershed is contributing a relatively consistent combined base flow and runoff at each location and does not include any streamflow regulation, alteration, or notable storage effects in the reach. Estimated qualifiers were applied to certain discharge values because of the effect of backwater on the ability to calculate or measure discharge. The streamflow measurements described above apply methods of measuring and calculating discharge as described in Rantz (1982) and Turnipseed and Sauer (2010). All streamflow results are described in table 1.2 and in Romanok and others (2026).

For this study, PFAS samples were analyzed at the Eurofins Environmental Testing Laboratory using EPA Method 1633 (EPA, 2024a). The entire sample was extracted onto a pre-conditioned WAX solid-phase-extraction (SPE) cartridge using a pre-cleaned vacuum manifold at 5 milliliters per minute. The SPE cartridge was dried and eluted with 5 mL of 1 percent methanolic ammonium hydroxide. After extraction, 25 microliters of concentrated acetic acid were added to each sample extract. The extract was vortexed before 10 milligrams of carbon were added. Finally, the extraction tube was mixed, vortexed again for 30 seconds, then centrifuged at 2,800 rotations per minute for 10 minutes. All sample extracts were analyzed by liquid chromatography tandem mass spectrometry in multiple reaction monitoring mode. Method detection limits (MDL) ranged from 0.38 to 6.8 ng/L (Romanok and others, 2026, table 3). Values above the MDL, but below the reporting levels, are reported by the laboratory as estimated with an “E” code. These data were considered “detections,” and the value was included in all summary statistics. For a full list of compounds analyzed for this study, refer to Romanok and others (2026).

TOC analysis was performed within 10 days of sample collection at the NJDWSC laboratory following Standard Method 5310B (Lipps and others, 2017). After field preservation, the samples were homogenized. The acidified samples were sparged with high-purity air to eliminate ion chromatography components before measuring the total carbon (TC) concentration, which is defined as non-purgeable organic carbon; the measurement obtained is also generally called TOC. TOC was measured by injecting the sample into a TC combustion tube filled with a platinum oxidation catalyst and heated using a 680 °C furnace. Carrier gas, high-purity air, was added at a controlled flow rate of 150 milliliters per minute. After the sample’s introduction to the combustion tube, the TC component in the sample decomposed, producing carbon dioxide. After decomposition, the carrier gas containing the combustion product flowed through the combustion tube. An ion chromatography reaction vessel was cooled and dried via a dehumidifier, and sodium hydroxide pellets were added to eliminate bacterial growth as needed. The sample then went through a halogen scrubber into a sample cell set in a non-dispersive infrared gas analyzer where carbon dioxide is detected. The non-dispersive infrared gas analyzer output a detection signal that generated a peak, the area of which was calculated via a data processor.

Trace element samples were collected in a 500-mL HDPE bottle and were analyzed at the NJDWSC laboratory by EPA Method 200.8 (Creed and others, 1994) using a Perkin Elmer NexION 1000 Series Inductively Coupled Plasma-Mass Spectrometry instrument. After field acidification, samples were held for 24 hours and a pH of less than 2 was verified before analysis. Samples were then digested using nitric and hydrochloric acid. The sample was converted to an aerosol by a nebulizer, then introduced into a plasma via a peristaltic pump tube. In the plasma, the energy transfer processes caused desolvation, atomization, and ionization. Perkin Elmer Syngistix software (Perkin Elmer, 2026) was used to process and report results.

Samples for mercury analysis were collected in a 500-mL HDPE bottle and were acidified in the field to a pH less than 2. After 24 hours, the pH was checked and verified to have remained at less than 2. Mercury analysis followed EPA Method 245.1 (O’Dell and others, 1994) using a Perkin Elmer FIMS 100 atomic absorption instrument with an autosampler at the NJDWSC laboratory. A portion of each sample was digested in a 50-mL Digitube using a block digester at 95 °C for 2 hours and then aspirated into a mixing cell. Tin dichloride was then added to generate mercury as cold vapor. The vapor was passed on to a Cold Vapor Atomic Absorption analyzer fitted with an electrically heated 10-centimeter cell, then absorption was measured. Perkin Elmer Syngistix software (Perkin Elmer, 2026) was used for the processing and reporting of results.

1,4-dioxane was analyzed using EPA Method 522 (Munch and Grimmet, 2008) at the ALS Environmental Laboratory. Samples were extracted by SPE and eluted with dichloromethane before a tetrahydrofuran-d8 reference standard was added. The samples were subsequently dried and analyzed using gas chromatography and mass spectrometry with an MDL of 0.023 micrograms per liter (µg/L) and a reporting limit of 0.07 µg/L. Values above the reporting limit and below the MDL were reported as estimated with a “J” code.

Quality-assurance and quality-control samples were collected and reviewed for each sampling event and included field blanks, laboratory blanks and spikes, and (or) stable isotope surrogates. Field personnel collected the following quality-control blanks:

Field blank collection procedures matched the procedures used for the collection of environmental samples but used organic-free blank water. One low-level detection of TOC (0.504 mg/L) was observed in the field blank collected during the September sampling event at USGS-01384000 (Site 3), which was just above the reporting limit of 0.500 mg/L and not within the range of detected TOC concentrations (1.35 to 9.93 mg/L). One low-level detection of 6:2 fluorotelomer sulfonate (6:2 FTS) was reported during the January sampling event at USGS-01379200 (Site 24); the concentration was above the MDL but below the reporting limit and was reported as estimated. However, 6:2 FTS was not detected in any environmental sample during that sampling event, so no values required censoring. Trace element analysis of the field blanks detected low levels of mercury in three field blanks collected in September. Environmental detections at or below twice the associated field blank concentrations were censored to non-detect. This required censoring to non-detect for all mercury detections (18 total, all from September) for the dataset.

Laboratory quality assurance included isotope dilution standards for PFAS. The median isotope dilution standard percentage recovery for all sampling rounds was 92.2 percent (ranging from 8.98 to 170 percent, 25th and 75th percentiles are 88.6 and 96 percent, respectively). Refer to table 4b in Romanok and others (2026) for the sampling event summary table. Other laboratory quality assurance performed included method blanks (no detections for all sampling rounds) and laboratory spike samples (all recovery percentages were within the recovery percentage limits as defined by the laboratory). The USGS reviewed these data and no issues were identified.

For this study, landscape metrics, including land cover, catchment size, and the number of catchments in the upstream watershed and potential PFAS sources, were obtained to help characterize areas that could be contributing PFAS to the environment. Data were analyzed for the broader Passaic River Basin to characterize the influence of areas upstream from the sampling area watersheds. Land-cover data were downloaded from the National Land Cover Database (NLCD) 2021 (Dewitz, 2023), and potential PFAS sources were downloaded from (1) EPA Enforcement and Compliance History Online (ECHO) PFAS Analytics Tools database (EPA, 2025c), (2) NJPDES (NJDEP, 2025c), and (3) the [New York] State Pollutant Discharge Elimination System (SPDES; New York State Department of Environmental Conservation, 2025). For more information on the potential source datasets used in the study, as well as the processing steps, refer to appendix 2.

Landscape metrics were evaluated at both the local catchment and the upstream watershed scales using NHD (EPA, 2012). Local catchment size for the sampling sites ranged from 0.0117 to 21.5631 km², and upstream watershed size ranged from 3.476 to 2,109.229 km². Local catchments for the sampling sites were identified using a spatial join in ArcGIS Pro (version 3.5.3; Esri, 2025). Upstream watersheds for the sampling sites were identified utilizing the “nhdplusTools” (version 1.4.2; Blodgett and Johnson, 2022) package of the R Statistical Software (version 4.5.1; R Core Team, 2021), which uses NHD flowline data to identify all contributing NHD catchments. These contributing catchments, including the local catchment, were merged to create one upstream watershed per sampling site. The resulting local catchment and upstream watershed boundaries were used to summarize the land-cover categories provided in the 2021 NLCD 30-meter raster dataset. The NLCD was used to ensure consistency between state lines because the Passaic River Basin falls in New Jersey and New York. The “Zonal Histogram” tool was used to create these summaries in ArcGIS Pro. Land-cover category descriptions can be found in table 5.

Potential PFAS sources were also summarized at the local catchment and upstream watershed scales for each sampling site. Before calculating the summaries, potential PFAS source records were assessed to identify and remove duplicates within and between datasets to ensure that PFAS sources were not counted more than once. Where multiple industrial categories were identified at a single facility in the ECHO database, a single point, or “potential source,” was retained, without regard to category. State data downloads (NJPDES and SPDES) were parsed only to include permit types relevant to PFAS (table 2.1). Because a single facility within the state datasets may hold more than one permit type owing to varying discharge methods, each permit was counted as an individual source. Once each dataset was checked for duplicates and cleaned, the ECHO data were compared to the state data. Where a permit location existed in both the ECHO and the NJPDES or SPDES datasets, the state record was kept owing to the higher level of detail provided, and the ECHO record was removed. Counts were executed using the “Summarize Within” tool in ArcGIS Pro. More information about this process can be found in appendix 2, with details about permits described in table 2.1.

To understand water reuse broadly, the previously developed ACCWW model (Barber and others, 2019, 2025; Faunce and others, 2023a) was applied to the Passaic River Basin to quantify the percentage of streamflow in a flowline that was comprised of municipal and industrial WWTP effluent during various times of the year. The ACCWW percentage is equal to the sum of all upstream wastewater inputs divided by the streamflow of an individual flowline multiplied by one hundred. Because wastewater effluent is considered a likely source of PFAS to the environment (Barber and others, 2025), the ACCWW percentage can be used to identify potential “hot spots” of contamination for future study or to inform management actions. Code developed from algorithms to analyze location and point-discharge data from WWTPs (EPA, 2024b), travel time, and upstream and downstream connections from NHD flowline data were used to derive mean-monthly (mean-August or mean-April) and mean-annual ACCWW percentages (Faunce and others, 2023a,b; Barber and others, 2025). The ACCWW percentage was then calculated as the percentage of accumulated (upstream) wastewater to total streamflow (Faunce and others, 2023a). Mean-monthly (12 values per flowline) and mean-annual (one value per flowline) streamflow were obtained from the NHD Enhanced Unit Runoff Method and reflect normalized data from 1971 to 2000 (EPA, 2012). More information about the input data sources and the ACCWW model archive is available in Romanok and others (2026).

Cumulative (∑) PFAS concentrations were calculated for each sample by summing the concentrations of all detected compounds. Summary statistics (median and ranges) were calculated for PFOA, PFOS, and ∑PFAS concentrations in each watershed by month under base-flow conditions and for each rain event. All non-detects for PFOS (2.7 percent, or 5 of 181 samples) were replaced with zero before applying summary statistics and linear models; for PFOA and ∑PFAS, there were no non-detects. Statistical analyses were performed using R Statistical Software (version 4.5.2; R Core Team, 2021) and the “stats” (R Core Team, 2013), “emmeans” (Lenth, 2025), and “corrplot” (Wei and Simko, 2024) packages. Spearman’s rank correlation was used to evaluate potential correlations among water-quality parameters, potential sources, mean-annual ACCWW percentage, and land-cover metrics with median ∑PFAS concentrations and median concentrations of PFOA and PFOS. Median concentrations of PFOA, PFOS, and ∑PFAS were calculated using only base-flow sampling events. The linear association between median PFOA, PFOS, and ∑PFAS concentrations and significant geospatial predictor variables identified by the correlation analysis was graphed in SigmaPlot (version 16; Grafiti LLC, 2026) to assess the strength and linearity of the relationship. For graphing purposes, concentrations were log-transformed to address normality concerns.

Linear models were used to determine differences in mean concentrations of PFOA and PFOS (1) between the four watersheds at the HUC10 level, (2) between two seasonal groupings (winter and spring [represented by January and March] against summer and fall [represented by July and September]) under base-flow conditions, (3) between the July base-flow and rain-event samples across all watersheds, (4) between the July and December rain events at the eight sites sampled during both events, and (5) between the rain-event and base-flow samples across all sites. Estimated marginal means were calculated and pairwise comparisons for each of the five groups listed above were performed, including a Bonferroni correction (p<0.05), to identify significant differences.

PFAS yields were calculated to normalize each site to the watershed drainage area at both the local catchment and upstream watershed scale. Yields were calculated from measured PFAS concentrations, drainage area, and instantaneous discharge, reported in Romanok and others (2026), by converting flow to liters per day and concentrations to grams using the following equation:Y = (C × Q × 0.00245) ÷ DA,(1)whereY

= yield in grams per day per square kilometer (g/d/km²),

A review of all the groundwater and surface water data collected within the study area indicated that, with few exceptions, concentrations of PFOA were greater than PFOS (table 1.3; USGS, 2026). Where PFOS was greater, the concentrations were within the same magnitude. In five instances, PFAS samples collected by the USGS from 2020 to 2023 were at or near sites sampled in 2025 for this study (table 6). Differences in median concentrations of PFOA and PFOS for samples collected in 2020 to 2023, and those collected in 2025 at all five sites, were within ±3 ng/L of one another (table 6).

Through the Open Public Records Act request, we obtained municipal wastewater influent and effluent sample data from 2021 to 2024 for the Pompton River (table 1.4). PFOA was detected in 10 of the 12 influent samples, and the median concentration was 15.0 ng/L. PFOS was detected in all but 3 samples, and the median concentration was 10.0 ng/L. The samples with no detections had elevated reporting limits (greater than 16 ng/L), which could explain the non-detects. Similarly, PFOA was detected in all 13 reported effluent samples, and the median concentration was 16.8 ng/L; PFOS was detected in all but one of the 13 samples, and the median was 6.80 ng/L. These results are consistent with concentrations of PFOA and PFOS detected in samples collected at Sites 20 and 21, both downstream from a WWTP. Median concentrations of PFOA at Sites 20 and 21 were 9.20 and 9.25 ng/L, respectively, and for PFOS, 6.20 and 6.05 ng/L, respectively.

Thirteen groundwater samples within this study area were collected by the USGS in 2022 (table 1.3; USGS, 2026). PFOA was detected in 85 percent of the samples. The concentrations ranged from below the reporting limit to 54.0 ng/L, the median concentration was 15.5 ng/L, and all detections exceeded 4 ng/L. PFOS was detected in 69 percent of the samples with concentrations ranging from below the reporting limit to 26.0 ng/L, the concentration median was 4.60 ng/L, and 53 percent of the samples exceeded 4 ng/L.

Querying the NJPDES database identified five permitted dischargers that were monitoring and reporting PFAS in effluent (table 1.5). Based on this query, at least one WWTP discharges effluent to the groundwater in the watershed. The median PFOA concentration in samples reported from 2023 and 2024 was 13.0 ng/L, and the median concentration of PFOS was 6 ng/L. One additional NJPDES permittee was noted as discharging to groundwater in the watershed from 2022 to 2024, with corresponding median sample concentrations of 9.70 ng/L and 5.20 ng/L for PFOA and PFOS, respectively. Two permitted surface-water dischargers had PFAS data available. The listed concentrations of PFOA and PFOS were low or not detected (table 1.5).

Streamflow was measured or computed along with 175 samples collected in this study (table 1.2). There was no streamflow for USGS-01383905 (Site 4) because the sample was collected near a boat launch and was not conducive to measuring. At two small tributary sites, USGS-0138752050 (Site 7) in September 2025 and Site 3 in January 2025, there was no flow. During the January sampling, the presence of ice affected streamflow measurements and computations, and resulted in several estimated streamflow values; these values are noted in table 1.2. For USGS-01389490 (Site 36), PVWC provided estimated streamflows for the diversion channel where samples were collected (Romanok and others, 2026). Sampling during January, March, July, and September aimed to be as close to base flow as possible.

Historical streamflow data were compared with the reported streamflow from this study for 11 sites where long-term mean-monthly streamflow data were available (table 7). Streamflow from this study ranged from 16 to 97 percent of the mean-monthly streamflow for the period of record listed, and from 31 to 159 percent of the monthly-mean streamflow for 2025. Median streamflow was highest in the Passaic River watershed, followed by (in descending order) the Pompton, Ramapo, and Wanaque River watersheds. Generally, streamflow in January (median, 95.0 ft³/s) and March (median, 242 ft³/s) was greater than that in July (median, 62.0 ft³/s) and September (median, 23.5 ft³/s).

In July, there were two sampling events, one under base-flow conditions and one after an estimated 1 to 4 inches of precipitation in the Passaic River Basin. The USGS computed streamflow for 26 of the 27 sites sampled during the July rain event, approximately 8 to 12 hours after the event. The computed streamflow values were 2 to 192 times higher than those reported during the base-flow sampling (median, 9 times higher) (table 1.2). In December, nine main-stem river sites were sampled as streamflow increased to help characterize PFAS concentrations from active overland runoff (fig. 3). Measured streamflow during the December rain-event sampling ranged from 55 percent less than the median base-flow samples (from January, March, July, and September) to 99 percent greater than the median base-flow samples (table 1.2).

Surface-water-quality samples were collected for PFAS and ancillary constituents under base-flow and rain-event conditions. Forty PFAS, 17 trace elements, TOC, 1,4-dioxane, and water-quality parameters (pH, specific conductance, water temperature, dissolved oxygen, and turbidity) were analyzed in samples collected from January to December 2025. All results are reported in Romanok and others (2026) and summarized below.

Water-quality parameters, including TOC, varied spatially and seasonally, and are briefly summarized in this section. Refer to Romanok and others (2026) for the full dataset. Median specific conductance measurements in all watersheds ranged from 31.5 µS/cm in the Ramapo River watershed to 943 µS/cm in the Passaic River watershed. Specific conductance was highest during July and September (527 and 616 µS/cm, respectively) and lowest in January and March (453 and 402 µS/cm, respectively). The highest median specific conductance was observed during the December rain event (636 µS/cm), and the lowest was during the summer rain event (278 µS/cm). Median concentrations of TOC in all watersheds ranged from 1.96 mg/L in the Ramapo River watershed to 8.22 mg/L in the Passaic River watershed. The highest median TOC concentrations were observed during July (base flow and rain event) and September (4.67, 4.66, and 4.26 mg/L, respectively). The lowest TOC concentrations were observed during the December rain event and the base-flow event in January and March (3.50, 3.22, and 3.07 mg/L, respectively). Median dissolved oxygen measurements ranged from 6.90 mg/L in the Passaic River watershed to 15.0 mg/L in the Ramapo River watershed. The highest dissolved oxygen concentrations were in January, March, and the December rain event (13.6, 12.1, and 12.0 mg/L, respectively), and the lowest were in September, July during the rain event, and in July during base-flow (7.80, 7.10, and 6.95, respectively).

Samples during the March base-flow event were analyzed for 1,4-dioxane, which was detected above the MDL (23 ng/L) in 32 percent of the samples. Overall, concentrations ranged from 29 (USGS-01381940 [Site 34]; fig. 2) to 250 ng/L (Site 3), with a median of 105 ng/L. 1,4-dioxane was detected in samples from 3 of 5 sites in the Wanaque River watershed (Sites 1, 3, 5; range, 68–250 ng/L), from 3 of 9 sites in the Ramapo River watershed (Sites 6, 7, 9; range, 130–230 ng/L), and from 6 of 15 sites in the Passaic River watershed (Sites 32–37; range, 29–170 ng/L). There were no detections of 1,4-dioxane in the Pompton River watershed. The results for all 17 constituents collected opportunistically for trace element analysis are in Romanok and others (2026).

Land cover and potential PFAS sources were summarized at the NHD local catchment and upstream watershed scales for each sampling site (Romanok and others, 2026; Williams and others, 2026). HUC10 summaries were also calculated (table 10) and showed that the Wanaque River watershed is the most forested among the HUC10 watersheds and that the Passaic River watershed is the most developed. “Developed, Open Space” (defined in table 5) was identified as a relatively prominent land cover category in several HUC10 watersheds. The Pompton River watershed had the fewest potential sources of PFAS (57), and the Passaic River watershed had the most (858). USGS-01389890 (Site 37; fig. 2), on the Passaic River, the most downstream sampling site in this study, has a local catchment size of 21.6 km² that contains 55 ECHO and 20 NJPDES sources and is approximately 96-percent developed (Developed Low, Medium, and High Intensity, and Developed Open Space). The upstream watershed is 2,109 km², is 38-percent developed, and contains 470 ECHO, 470 NJPDES, and 214 SPDES sources (fig. 4).

Mean-monthly and mean-annual ACCWW percentages were calculated for all flowlines (Romanok and others, 2026). Mean-monthly streamflow data were generally the lowest in August (EPA, 2012), so mean-August streamflow data were used in this study to demonstrate when wastewater inputs might make up the largest percentage of the streamflow, resulting in the greatest mean-monthly ACCWW percentages, similar to other studies (Faunce and others, 2023a; Barber and others, 2025). A de facto wastewater reuse of 1 percent in streams has been associated previously with higher concentrations of contaminants and can be considered an acceptable threshold for assessing potential risks to drinking-water resources (Weisman and others, 2021). This report describes only flowlines with mean-annual ACCWW percentages that are greater than or equal to 1 percent.

In the Wanaque River watershed, there were 254 total flowlines, 24 (9 percent) of which had ACCWW percentages greater than or equal to 1 percent. The greatest mean-annual (4.06 percent) and mean-August (13.6 percent) ACCWW within the Wanaque River watershed was calculated for a flowline downstream from a WWTP. The streamflow in this segment ranged from 23.8 ft³/s (mean-August) to 224 ft³/s (mean-April). The ACCWW for the flowline at the outlet of the Wanaque River watershed was calculated as 11.4 percent during mean-August conditions and 3.29 percent during mean-annual conditions.

In the Ramapo River watershed, there were 350 total flowlines, 81 (23 percent) of which had ACCWW percentages greater than or equal to 1 percent. A flowline classified as a “CanalDitch” had the largest mean-annual ACCWW of 26.7 percent, and a separate flowline classified as a lake/pond had the highest mean-August ACCWW in the HUC10 (43.4 percent). Both segments are adjacent to WWTP facilities. Site 13, on the main-stem Ramapo River near a drinking-water intake, was on a flowline with a mean-annual ACCWW of 3.15 percent. Just downstream, at the outlet of the Ramapo River watershed, ACCWW values were 3.19 percent (mean-annual) and 8.62 percent (mean-August).

In the Pompton River watershed, 21 out of the 204 flowlines (10 percent) had a mean-annual ACCWW percentage greater than or equal to 1 percent. The highest mean-annual ACCWW was 28.2 percent at a flowline classified as both a “LakePond” and a headwater (where mean-annual streamflow was 0.39 ft³/s) and was adjacent to a non-major publicly owned treatment works facility. The same flowline had the highest mean-August ACCWW of all segments in the watershed, at 32.6 percent (mean-August streamflow is 0.34 ft³/s). The segment immediately downstream had the next highest average mean-August ACCWW at 15.1 percent, potentially because of the dilution of treated effluent entering from the headwater segment. Where the Ramapo and Pequannock Rivers converge into the Pompton River (near USGS-01388500 [Site 16]; fig. 2), the mean-annual ACCWW was 2.72 percent and the mean-August ACCWW was 7.81 percent. The ACCWW conditions at USGS-405412074161601 (Site 20; fig. 2) on the main-stem Pompton River, near a drinking-water intake, were calculated as 2.52 percent (total streamflow of 598 ft³/s) and 7.50 percent (total streamflow of 201 ft³/s) for mean-annual and mean-August, respectively. Downstream, USGS-0138900503 (Site 21; fig. 2), at the outlet of the Pompton River watershed, just upstream from the confluence with the Passaic River, was on a flowline where the mean-annual ACCWW was 11.3 percent (total streamflow of 734 ft³/s) and the mean-August ACCWW was 21.9 percent (total streamflow of 378 ft³/s).

In the Passaic River watershed, 123 of the 349 flowlines (35 percent) within the watershed had a mean-annual ACCWW greater than or equal to 1 percent. The highest ACCWW percentages in the broader Passaic River Basin (geospatial study area) were within the Passaic River watershed at two headwater flowlines between USGS-01379550 (Site 30; fig. 2) and USGS-01381900 (Site 33; fig. 2), each adjacent to a WWTP. The highest mean-annual ACCWW was 77.7 percent (total streamflow, 8.70 ft³/s), and the second highest was 54.1 percent (total streamflow, 2.80 ft³/s). The maximum ACCWW percentage at each segment was in August under low-flow conditions (86.5 and 66.9 percent, respectively). One of the downstream-most sites in the study, USGS-01389490 (Site 36; fig. 2), was near a drinking-water intake on the main-stem Passaic River and on a flowline where mean-annual ACCWW was 8.63 percent (total streamflow, 1,280 ft³/s) and mean-August ACCWW was 18.5 percent (total streamflow, 597 ft³/s).

An interactive map application was developed by Williams and others (2026) as part of this study to support further exploration of the data (PFAS results, landscape characteristics, and ACCWW percentages) at the various analysis scales (local catchment, upstream watershed, sampling study area, and geospatial study area) presented in this report. The application can be accessed at geonarrative.usgs.gov/northjerseypfas/.

As expected, and similar to other studies (Nguyen and others, 2022; Castellani and others, 2025; Gorski, 2025), concentrations of PFAS varied seasonally, likely driven by streamflow. Under base-flow conditions, concentrations of PFOA were similar between samples collected in January and March (winter and spring; p=0.839) and those collected in July and September (summer and fall; p=1.000), with higher concentrations observed in the summer and fall compared to in the winter and spring when flows were higher (fig. 6). For PFOS, the highest concentrations were observed in July followed by September with no significant differences between January and March (fig. 6). In urban watersheds, under low-flow conditions in the summer and early fall, wastewater effluent can be a contributor to streamflow and, based on previous studies, an increased proportion of wastewater can lead to higher concentrations of PFAS and more frequent exceedances of the EPA MCL at drinking-water intakes (Barber and others, 2025). PFOA concentrations near the three water-treatment-plant intakes sampled as part of this study frequently exceeded the MCL despite seasonally higher flows in January, with concentrations ranging from 4.6 to 12.0 ng/L. However, PFOS concentrations tended to be below or only slightly above the MCL in January and March (2.5−5.5 ng/L), followed by a marked increase as flows decreased in July and September (5.9−9.1 ng/L). Shallow groundwater is another source of PFAS to surface-water systems (McMahon and others, 2022; Tokranov and others, 2024). Based on historical USGS data, PFOA and PFOS were detected in 85 and 69 percent, respectively, of the 13 groundwater samples collected from the study area in 2020–23. PFOA concentrations in groundwater ranged from not detected to 54.0 ng/L (median, 15.5 ng/L), and PFOS concentrations ranged from not detected to 26 ng/L (median, 4.6 ng/L), indicating that groundwater could be a source of PFAS to the study area, particularly in the summer and fall during lower flow conditions.

Stormwater is often discharged into receiving waterbodies largely without treatment and thus has the potential to transport PFAS to streams within urban areas during rain events (Masoner and others, 2019). Because stormwater quality varies by rain-event and catchment, it can be difficult to determine the importance of stormwater runoff as a source of PFAS to the receiving water bodies. However, studies have shown increases in both the concentration and number of unique PFAS, particularly long-chain PFAS, detected during rain-events (Müller and others, 2011; Kali and others, 2025). In July, the number of individual compounds detected and concentrations of PFOA, PFOS, and ∑PFAS basinwide were similar between wet and dry events sampled within a week of each other, indicating that stormwater may not be a primary contributor of PFAS to the Passaic River Basin, particularly in the summer. At the site level, the only exception was USGS-01386000 (Site 5), where concentrations of PFOA and PFOS increased from 5.6 to 14 ng/L and 2.9 to 5.3 ng/L, respectively. Further, when comparing all base-flow events with the two rain events, concentrations of PFOA were similar (p=0.5919). PFOS concentrations were higher (p=0.0056) during the rain events compared to base-flow events (fig. 7). Rain-event-driven PFAS dynamics can be complicated by sources (point versus non-point) and the hydrology of the watershed. More information temporally and spatially could further our understanding of PFAS contributions from non-point sources in the study area.

Eight sites were sampled during both the July and December rain events: three sites each in the Ramapo and Pompton River watersheds, and two sites in the Passaic River watershed (table 1.1). No differences between events across all watersheds were observed for the number of PFAS detected (p=0.109) and concentrations of PFOS (p=0.0707). However, despite some site-level variations, concentrations of ∑PFAS (p=0.012) and PFOA (p=0.0055) across all watersheds were higher during the July rain event compared to the December rain event (fig. 8). The average streamflow during the July rain event was higher than what was observed in December (table 1.2), which could explain the higher concentrations observed during the July rain event. In the Passaic River Basin, PFAS occurrence and distribution varied seasonally and were complicated by varying flow regimes, stormwater quality, potential shallow groundwater interactions, and the loading of WWTP effluent into local stream reaches. Although data on PFAS in WWTP effluent in New Jersey is limited, based on other studies, PFAS loading from individual WWTP effluent can be highly variable and tends to be affected by sources of wastewater, treatment, and seasonality (Tavasoli and others, 2021; Thompson and others, 2022).

To provide a more direct comparison of relative PFAS loads among sites with different drainage areas and streamflow, yields were calculated for the local catchment and the upstream watershed (tables 1.6, 1.7). Throughout the Passaic River Basin, PFAS yields associated with individual sites calculated at the local catchment scale were driven mainly by PFOA and PFOS, which comprised anywhere from 17.5 to 100 percent (median, 25.4 percent) and 0 to 56.6 percent (median, 13.8 percent), respectively, of the total yields. The PFAS yields varied by site, season, and watershed, with the lowest yields calculated for the Wanaque River watershed (table 11). Except for Site 20 on the Pompton River, directly downstream from a WWTP and within a very small local catchment (0.0189 km²), yields were similar among the Ramapo, Pompton, and Passaic River watersheds (table 11). Yields at Site 20 ranged from 99 to 1,098 g/d/km² for PFOA and from 67 to 829 g/d/km² for PFOS, with the highest yields observed during the July rain event. Wanaque South pump station is between Sites 20 and 21. Yields decreased substantially at Site 21 and ranged from 10 to 17 g/d/km² for PFOA and 6.5 to 12.4 g/d/km² for PFOS.

Under base-flow conditions, the yields were generally highest in March and lowest in September (table 11). However, under elevated flow conditions caused by rain events, yields did increase as expected. Under base-flow conditions at all sites except Site 24, tributary PFAS yields tended to be lower (table 11) and contributed a limited mass of PFAS to the main-stem rivers, owing to their characteristically lower flows and the fact that the tributary sampling sites generally had fewer potential sources in the local and upstream catchments (fig. 9). Rain and higher-flow events have the potential to mobilize PFAS, and in mixed urban and industrial catchments, diffuse non-point sources such as stormwater can be as important as point sources. A more detailed understanding of yields within watersheds could better characterize PFAS dynamics which are complicated by hydrologic events and source types (diffuse versus point sources). Further, water-treatment plants are considered pivotal control points for PFAS, but inconsistencies with removal capabilities and precursor formation complicate any planned reduction strategies (Kim and others, 2024).

Calculating yields in the Wanaque River watershed is complicated by lakes and diversions. Seasonal yields were calculated for Sites 1, 2, and 5, which represent the primary inputs to the Wanaque Reservoir (a major drinking-water source). Overall, yields were highest in March, followed by the July rain event; PFOA and PFOS site-level yields ranged from 0.009 to 3.24 g/d/km² and 0.00 to 1.32 g/d/km², respectively. Similarly, in the Ramapo, Pompton, and Passaic River watersheds, yields were highest during the July rain event, followed by March. PFOA and PFOS yield ranges were as follows: in the Ramapo River watershed, from less than 0.0001 to 13.96 g/d/km² and 0.00 to 13.6 g/d/km², respectively; in the Pompton River watershed, from 0.009 to 1,098 g/d/km² and 0.003 to 829 g/d/km², respectively; and in the Passaic River watershed, from 0.0001 to 23.1 g/d/km² and less than 0.0001 to 9.69 g/d/km², respectively. The highest site-level yields were at Sites 20 and 21 in the Pompton River watershed, followed by Sites 33, 34, and 35 in the Passaic River watershed. In the Passaic River watershed, an increase in streamflow was observed between USGS-01379580 (Site 32; fig. 2) and Site 33 that resulted from surface-water inputs from the Rockaway River, which flows into the Passaic River. PFOA yields increased substantially from 0.064 to 0.498 g/d/km² at Site 32 to 1.47 to 9.35 g/d/km² at Site 33 (fig. 9) and then again at USGS-0138900501 (Site 35; fig. 2) from 5.18 to 23.1 g/d/km², at the confluence of the Pompton and the Passaic Rivers (fig. 10). The Rockaway River and its major tributary, the Whippany River, were not sampled during this study; however, based on the potential PFAS sources (fig. 4) and historical USGS results, they may contribute PFAS into the Passaic River. PFOA and PFOS samples collected on the Whippany River by the USGS from 2020 to 2023 ranged from 7.1 to 12.0 ng/L and 8.3 to 17.0 ng/L, respectively (USGS, 2026). Future sampling within the Rockaway and Whippany Rivers could likely help explain increased observed PFAS yields after Site 32 (fig. 9).

Urban watersheds often exhibit higher values of specific conductance, turbidity, and TOC compared to non-urban (forested or agricultural) watersheds, but the degree varies by region, land-cover intensity, and season. Under base-flow conditions, median ∑PFAS concentrations and the number of PFAS detected were positively related to specific conductance, turbidity, and TOC (tables 1.8, 1.9). Similar relationships were observed with median PFOA and, to a lesser extent, median PFOS concentrations, particularly for specific conductance and TOC (tables 1.8, 1.9). Specific conductance and turbidity are often higher in urbanized watersheds owing to increased salinization, runoff of sediments, and mobilization of dissolved solids (EPA, 2017; Baker and others, 2019; Miguel-Chinchilla and others, 2019), and TOC tends to be elevated in urban settings. However, relationships between TOC and land cover can be complex because of wastewater effluent and increased flows (Elias and others, 2013; Smith and Kaushal, 2015).

To understand potential predictors of PFAS contamination in the Passaic River Basin, ∑PFAS concentrations and concentrations of PFOA and PFOS were compared to landscape metrics, including land-cover categories and presumptive sources, using a correlation matrix and linear regression to assess the strength of the relationship. Correlation analysis revealed several potential predictors of PFAS concentrations (table 12). These relationships varied by scale (local catchment versus upstream watershed). At the local scale (table 12), predictors that were positively associated (p<0.05) with PFOA concentrations included catchment size, mean-annual ACCWW percentage, number of sources (ECHO and NJPDES), percentage of Open Water, and percentage of developed land, including Developed Open Space, and Developed Low, Medium, and High Intensity land-cover categories. The percentage of Deciduous Forest was negatively correlated with PFOA concentrations (table 1.8). For PFOS, percentage of developed land, including Developed Open Space, and Developed Low, Medium, and High Intensity land-cover categories, was positively associated with median concentrations, and the percentage of Deciduous and Mixed Forest, as well as the percentage of Pasture/Hay, were negatively associated with median concentrations (tables 12, 1.8). In the upstream watershed (table 12), median PFOA concentrations were positively related to the number of ECHO and NJPDES sources, percentage of developed land, including Developed Open Space, Low, Medium, and High Intensity land-cover categories, percentage of Pasture/Hay and Cultivated Crops, and percentage of Woody Wetlands. PFOA concentrations were negatively related to the percentage of Open Water, Grassland/Herbaceous, Deciduous and Evergreen Forest, and Shrub/Scrub (tables 12, 1.9). Median PFOS concentrations were positively related to the number of catchments, watershed size, number of sources (ECHO, NJPDES, and SPDES), and the percentage of developed land, including Developed Open Space, Low, Medium, and High Intensity land-cover categories, Emergent Herbaceous Wetlands, and Barren Land. They were negatively related to the percentage of Deciduous Forest (tables 12, 1.9).

Results of the linear regressions indicate that ∑PFAS and PFOA concentrations increased with increasing percentages of total developed land (sum of Developed Low, Medium, and High Intensity), Developed Open Space, and mean-annual ACCWW, and decreased with increasing percentages of Deciduous Forest at both the local catchment (fig. 11; table 1.10) and upstream watershed scales (table 1.11). PFOS concentrations, on the other hand, increased with increasing developed land at the local catchment scale and decreased as the percentage of Deciduous Forest increased at both the local and upstream scale (tables 1.10, 1.11). Many of the other predictors showed a significant relationship with PFAS concentrations (for example, ECHO and NJPDES sources and Pasture/Hay and Cultivate Crops), but the relationships were relatively weak (tables 1.10, 1.11).

Across the United States, urban land cover is a predominant driver of PFAS occurrence in a variety of matrices and has been identified as one of the top predictor variables in previous modeling efforts (McMahon and others, 2022; Smalling and others, 2023; Tokranov and others, 2024). Further, point-source municipal and industrial WWTP effluent discharges have been related to increased PFAS in surface waters (Schultz and others, 2006; Hu and others, 2016; Barber and others, 2025), and factors such as wastewater treatment technique, population size served, and proximity to WWTPs have been linked to concentrations (Podder and others, 2021), which could explain the positive relationships observed between PFAS concentrations and mean-annual ACCWW percentage. Barber and others (2025) reported a similar positive relationship between ∑PFAS concentrations in surface waters and the ACCWW percentage calculated for the Potomac River watershed. It is important to note that municipal WWTPs are considered “reservoirs” or “passive receivers” of PFAS because they do not manufacture it and cannot control its presence in the influent that arrives from homes and businesses (Mehan and Norris, 2026; Liu and others, 2026). Although agricultural land cover was limited in the study area (less than 2.5 percent; table 10), concentrations were only marginally correlated with Pasture/Hay and Cultivated Crops in the upstream watersheds, and the relationships were relatively weak (tables 1.10, 1.11). In agricultural areas, the occurrence of PFAS has been linked to land application of biosolids (Sepulvado and others, 2011; Munoz and others, 2022) and field application of pesticides (Donley and others, 2024). Thus, stronger relationships in more heavily agricultural areas are expected, as was observed in Pennsylvania surface waters (Breitmeyer and others, 2023). Presumptive sources in a watershed are often correlated with urban land cover (Smalling and others, 2023), but knowing the specific locations and types of sources helps identify areas most vulnerable to PFAS contamination. Calculating mean-annual and mean-monthly ACCWW percentages (fig. 12; table 1.12) for each flowline in a watershed can help identify and compare areas where PFAS concentrations could be elevated during various times of the year, such as during August when streamflow is generally lower, and help prioritize future sampling and mitigation efforts.

In urbanized watersheds like the Passaic River Basin, PFAS are widespread in surface waters used as drinking-water resources. Despite advanced techniques in the treatment process to minimize PFAS concentrations in treated drinking water, information on PFAS concentrations in source water can support water purveyors in managing their systems and preparing for enforcement of Federal drinking-water regulations in the coming years. Addressing water reuse and associated introduction of PFAS into drinking-water sources requires knowledge of upstream presumptive sources and changes in concentrations across time and space. Data collected across four large watersheds in the Passaic River Basin, including near two drinking-water intakes and upstream from the Wanaque Reservoir, can help water purveyors make decisions regarding effective treatment and mitigation strategies and work with the NJDEP to reduce PFAS contamination of drinking-water supplies.

Understanding seasonal differences in PFAS concentrations and yields and in water chemistry (for example, TOC) can inform decisions on when to pump water from existing surface-water intakes as well as water treatment options. PFAS associated with elevated TOC can complicate treatment options (Tshangana and others, 2025), therefore understanding the complete water chemistry allows purveyors to best address treatment throughout the year. For example, PFOA and PFOS concentrations were lower during the winter and spring when flows were higher. Further mean-annual ACCWW percentages across the study area were low (maximum, 7 percent) compared to mean-August (maximum, 22 percent) when streamflow was low. Sites 13, 20, and 36 were sampled near drinking-water intakes, with mean-annual ACCWW percentages of 3.2, 2.5, and 8.6 percent, respectively, and mean-August ACCWW percentages of 8.6, 7.5, and 18.4 percent, respectively. Further, median groundwater concentrations in the study area were 15.5 ng/L and 4.6 ng/L for PFOA and PFOS, respectively, indicating shallow-subsurface groundwater could also be an important contributor to surface waters, particularly in the summer and fall when streamflow is low. Additional information could inform decisions to change pumping rates or drinking water sources. For example, water purveyors may be able to minimize the degree of treatment and mitigation needed to meet Federal drinking-water standards by pumping less from surface waters, relying primarily on reservoirs, when PFAS concentrations are highest (the summer and fall), thereby reducing intake concentrations. Concentrations of PFAS in the Wanaque River watershed, upstream from the Wanaque Reservoir, were low most of the year compared to the other watersheds. The Wanaque River watershed is affected by fewer sources, lower urbanization, and more forested land cover. During the July rain event, concentrations of PFOA and PFOS increased compared to those observed under base-flow conditions, which indicates non-point sources like overland flow are also likely contributors of PFAS in the Wanaque River watershed. Because non-point sources were not the primary focus of the current study, continued sampling of the reservoir and its major tributaries under different flow conditions could help accurately assess PFAS dynamics within the Wanaque Reservoir.

The ACCWW model applied to the Passaic River Basin is adaptable to other watersheds, can be scaled up or down to address a variety of water-quality issues. For example, the ACCWW model could be used to identify potential “hot spots” of PFAS and other wastewater-derived contaminants in unsampled watersheds, allowing resource managers to target funds to specific areas and potentially eliminate the need for broad spatial and temporal reconnaissance studies. Water purveyors could also use this type of information to determine when to pump water from certain stream reaches, potentially identify areas in need of mitigation, or inform potential locations of new surface-water intakes. Monitoring PFAS concentrations in wastewater and industrial effluent could help improve the ACCWW model and allow for improved prediction of PFAS concentrations in unsampled stream reaches throughout the state, as described in other watersheds (Barber and others, 2025). To understand the occurrence, distribution, and potential loading of PFAS into local surface waters used as drinking-water resources, studies could be designed in collaboration with water purveyors to address their needs and inform potential strategies for reducing PFAS to help meet new Federal regulations.

The current study provided preliminary information needed to understand overall occurrence, distribution, and loading into local surface waters over 1 year of sampling under various seasons and flow regimes, but it has its limitations. First, shallow groundwater, which is considered a source of PFAS (Tokranov and others, 2024), was not assessed in the study and could be a contributor in some reaches, particularly during low-flow conditions in the summer. Based on limited data from groundwater samples collected in 2022, PFOA and PFOS were observed in more than 65 percent of the samples, with 84 and 53 percent of the samples, respectively, exceeding the EPA MCL of 4 ng/L (USGS, 2026). This indicates that shallow groundwater could be a contributing source of PFAS to surface waters, particularly during lower-flow conditions, and more data on groundwater PFAS contributions could help managers more adequately assess PFAS dynamics in the region. Oyen and Ophori (2025)’s chloride particle-tracking study within the Passaic River watershed indicated the probability of groundwater-to-surface-water interaction. They also highlighted a study by Newell and others (2022) that described how the application of road salt could potentially prolong the occurrence of PFAS contamination owing to changes in solubility, potentially affecting PFAS dynamics in the winter. Second, atmospheric deposition is a known source of PFAS, particularly in forested areas like the Wanaque River watershed where anthropogenic sources are more limited. Often, atmospheric concentrations of PFOA and PFOS exceed 4 ng/L (Cousins and others, 2022) and could be affecting smaller, forested stream reaches with low ACCWW percentages. Third, there is currently little data on PFAS concentrations in wastewater and industrial discharges available, making it difficult to pinpoint specific sources that could be a major contributor of PFAS to a stream reach. To model potential occurrence and concentrations in unsampled reaches, more accurate data on PFAS concentrations from all sources with a discharge permit are needed. Fourth, the Rockaway and Whippany Rivers were not sampled as part of this study but could be a source of PFAS loading to the Passaic River watershed based on the geospatial data compiled, the increase in yields on the Passaic River downstream from the confluence with these rivers, and the limited USGS data available. Lastly, event sampling was limited across time and space in this study. Although some conclusions can be made regarding presumptive sources of PFAS to the basin, how non-point sources, such as combined sewer overflows and urban runoff, affect PFAS occurrence and distribution during rain events, and if those effects vary seasonally, is not fully understood. Despite the above limitations, the current study provides foundational information on PFAS in surface waters to help resource managers make decisions and identify next steps as they prepare to meet PFAS regulations for drinking water.