Sampling Methods

Sampling and laboratory methods used for this study are detailed in Tagliaferri and others (2021). Groundwater samples in Hempstead Harbor were collected in November 2018. Wet conditions were defined as more than 0.25 inch of precipitation in 24 hours or 0.50 inch in 48 hours. Dry conditions were defined as less than 0.25 inch of precipitation in 72 hours. Rainfall totaled 0.53 inch on June 4 and 5, 2018, designating samples collected on June 6, 2018, as summer wet samples (more than 0.50 inch within 48 hours). Winter wet samples were also collected on April 15 and 16, 2019, with 0.49 inch of precipitation falling on April 13, 0.01 inch falling on April 14, and 0.61 inch falling on April 15, 2019. Rainfall values were obtained from the Hempstead Harbor Club rain gage (fig. 1) identified as site number KNYGLENC2 (Weather Underground, 2020).

From June 2018 to April 2019, 14 sites were sampled within Hempstead Harbor, yielding 41 surface-water samples (2 of which were replicates), 3 groundwater samples (1 of which was a replicate), and 2 sediment samples collected concurrently with surface-water samples at 2 sites (fig. 2). Of the surface-water sites, six were considered to be source sites (waters flowing directly into Hempstead Harbor from the land through culverts and groundwater discharge), and six were considered to be receptor sites (the harbor itself). Eighteen surface-water samples were collected under wet conditions, and 21, under dry conditions; 15 surface-water samples were collected in the summer, and 24, in the winter.

Laboratory Methods

Laboratory methods used for this study are detailed in Fisher and others (2020). Samples for fecal coliforms were analyzed at Suffolk County Department of Health Services Public Environmental Health Laboratory (PEHL). Additional information was either observed or compiled to facilitate interpretation of data. Precipitation measurements from the Hempstead Harbor Club rain gage, identified as site number KNYGLENC2 (Weather Underground, 2020), were used to determine dry and wet weather conditions. Quality assurance for samples collected at site KNYGLENC2 is unavailable, and values were not used in any quantitative analysis. Relative tide stage as noted by USGS field personnel and verified based on available tidal predictions (National Oceanic and Atmospheric Administration, undated), was documented along with the sample information in the National Water Information System (NWIS; U.S. Geological Survey, 2020a) database. Land-cover and sewer district geographic information system (GIS) coverages (Nassau County, 2015; Suffolk County, 2020; U.S. Geological Survey, 2020b) were used for data interpretation. These data were used to increase confidence in the geographical source of the water and in the likely transport mechanisms of fecal contamination to the Hempstead Harbor embayment.

Microbial Source Tracking

Throughout Hempstead Harbor, human (HF183) and canine Bacteroides (BacCan) markers were detected in roughly half of the surface-water samples from source and receptor sites. Ruminant Bacteroides markers (Rum2Bac) were detected in only two samples—one source and one receptor sample, each in dry weather and during the summer months—suggesting little contribution from the resident deer population. All but 1 of the 16 detections for the waterfowl Helicobacter marker (GFD) were in the winter months, suggesting a seasonal flux in avian fecal contamination. For HF183, a greater percentage of samples collected during dry weather (53 percent) had detections above the reporting limit compared with samples collected during wet weather (46 percent; table 1). This may indicate a constant source of fecal contamination entering Hempstead Harbor that is unrelated to weather events. Of the BacCan results detected above the reporting limit, 32 percent were in winter samples, and 70 percent were in summer samples (table 1), which may reflect the use of public space for dog walking in warmer months. As with HF183, BacCan detections during wet and dry weather were similar (38 percent and 47 percent, respectively), suggesting canine influence is unrelated to weather events.

Host Specific Markers

Reporting limits for the three predominant MST markers (HF183, BacCan, and GFD) detected in Hempstead Harbor varied slightly for samples because of analytical processes, such as dilutions required because of matrix interferences. Understanding occurrence of the predominant MST markers in Hempstead Harbor can help prioritize locations for additional investigation and infrastructure improvement and provide a baseline for host contributions when prioritizing sites of concern for assessing FC and after total maximum daily load implementation.

HF183

The HF183 human Bacteroides marker was detected at least once in surface-water samples at each receptor site, which is indicative of the continuous inputs from potential inland sources (including the Glen Cove Water Pollution Control Plant and illicit connections to the stormwater drain system). Concentrations in samples collected from the same site were generally higher in dry weather than wet weather (fig. 3). In general, concentrations of HF183 were lower in receptor samples than in source site samples, except for the Mouth of West Pond site. Crescent Beach, near West Pond, has been shown to contain high concentrations of bacteria with failing septic implicated (New York State Department of Environmental Conservation, undated); therefore, discharge from the Mouth of West Pond site would be more representative of a source to the DEC designated shellfish area of Hempstead Harbor.

Samples from three of the six surface-water source sites in Hempstead Harbor had detections of HF183 above the reporting limit. Samples collected in winter at the Culvert at Glenwood Road site and at the end of the marina dock in Glen Cove Creek had concentrations of HF183 that were orders of magnitude higher than most receptor samples. Concentrations of HF183 in the two samples collected at the Culvert at Glenwood Road site during winter months were 130,000 and 82,000 copies/100 mL compared with 500 copies/100 mL in the summer wet sample and corresponded with fecal coliform concentrations of 1,700 and 9,200 MPN/100 mL in the winter and 800 MPN/100 mL in the summer. The continuous flow observed at the Culvert at Glenwood Road source site, regardless of weather conditions, is suggestive of groundwater infiltration and discharges from illicit septic connections to the MS4 as observed in nearby Alley Creek in Queens, New York (New York City Parks Natural Resources Group, 2015).

Samples from Glen Cove Creek also contained high concentrations of HF183 in both wet (54,000 copies/100 mL) and dry (140,000 copies/100 mL) winter samples. Concentrations of fecal coliform were 9,200 and 170 MPN/100 mL in wet and dry winter samples, respectively (table 1.2; U.S. Geological Survey, 2020a). Given the timing (winter) of sample collection, it is unlikely that illicit boat discharges were responsible for the HF183 detections in these samples because most of the boats had been removed from the dock slips several months earlier. There were no HF183 markers detected in samples collected at the end of the marina dock in Glen Cove Creek during the summer.

BacCan

The canine Bacteroides marker (BacCan) was detected at least once at five of the six source sites and at least once at four of the five receptor sites with available MST data (fig. 4), likely owing to runoff contributions from pet waste and to mixing from the Glen Cove Water Pollution Control Plant effluent. Detections in receptor samples were generally from dry-weather samples, whereas detections in source samples were a more even mix of wet- and dry-weather samples.

Concentrations of BacCan were highest at the Culvert at Glenwood Road site (82,000 copies/100 mL) during wet weather in the winter season and coincided with high concentrations of HF183 (82,000 copies/100 mL) detected in the same sample. All three samples from Glen Cove Creek had concentrations above the reporting limit for BacCan, reflecting the multiple inputs to the creek. Detections of the BacCan marker in source samples from the Pond Culvert at Hempstead Harbor South may reflect the greater number of sidewalks and the dog park adjacent to the shore.

Fecal coliform concentrations in samples from the Outfall at Skillman Street site were consistently high, particularly during dry weather, and correspond with elevated BacCan concentrations. Fecal coliform concentrations in samples from the Pond Culvert at Hempstead Harbor South site (which discharges water from Roslyn Pond) were higher in the wet-weather sample that had low BacCan concentrations. This is possibly a result of dilution by stormwater from connected storm drains carrying greater fecal coliform loads; however, more samples are required to make this conclusion.

GFD

The waterfowl Helicobacter (GFD) marker was detected at least once at five out of the six source sites and at least once at four out of the five receptor sites with available MST data, primarily in the winter, with only one detection during the summer months at the Mouth of West Pond site (fig. 5). No GFD markers were detected in surface-water samples collected at the Hempstead Harbor near Morgan Park site or at the Culvert at Glenwood Road site. Further, the only host-specific marker present in the two samples collected from the Culvert at Tappen Beach site was GFD, despite Scudders Pond (which discharges to Hempstead Harbor through this culvert) being surrounded by homes with OWDS.

In general, receptor samples with detections of the GFD marker corresponded with low fecal coliform concentrations (table 1.2; U.S. Geological Survey, 2020a) relative to detections for HF183 (fig. 3) and BacCan (fig. 4). This was also the case with source samples and is particularly noticeable in samples with no other host-specific marker present. For example, fecal coliform and GFD concentrations in the dry sample from the Culvert at Tappen Beach site were less than 18 MPN/100 mL and 19,000 copies/100 mL, respectively, and in the dry sample from the Spillway at Skillman Street site were less than 18 MPN/100 mL and 2,500 copies/100 mL, respectively. An exception is a wet weather sample at the Outfall at Skillman Street site that had a fecal coliform concentration of 490 MPN/100 mL and a GFD concentration of 2,900 copies/100 mL.

Fecal Coliform Bacteria

Concentrations of fecal coliform in samples from surface-water source sites ranged from below the reporting limit in 12 percent of samples to greater than 16,000 MPN/100 mL (the maximum reporting limit for the dilution protocol) in two samples, one from the Outfall at Skillman Street site (under dry-weather conditions) and the other from Spillway at Skillman Street site (under wet-weather conditions). Concentrations for samples collected at the receptor sites ranged from less than the reporting limit in 41 percent of samples to 300 MPN/100 mL in a single sample at the Hempstead Beach Park North site under wet-weather conditions (table 1.2; U.S. Geological Survey, 2020a). Only three of the receptor samples, the sample from the Hempstead Beach Park North site, one at the Mott Cove site (140 MPN/100 mL), and one at the Mouth of West Pond site (61 MPN/100 mL), were above 49 MPN/100 mL, the 90-percentile concentration determined by the DEC for Hempstead Harbor as part of the National Shellfish Sanitation Program closure criteria (U.S. Food and Drug Administration, 2019). The two sites (Outfall at Skillman Street and Spillway at Skillman Street) with the highest fecal coliform concentrations (greater than 16,000 MPN/100 mL) did not have the highest concentration or highest frequency of detection for any of the Bacteroides and Helicobacter markers.

The relation between fecal coliform and marker detection is not straightforward. Fecal coliform and specific host-associated markers may decay at different rates and be present in different abundances. Host-associated markers were shown to decay more rapidly than FC in (Rogers and others, 2011; Drozd and others, 2013). In general, high fecal coliform concentrations also did not relate to weather condition or season.

Isotopes

Isotopes of nitrogen (δ¹⁵N) and oxygen (δ¹⁸O) in nitrate in samples from source sites indicated instances of contributions from wastewater-affected groundwater being diluted during wet weather (fig. 6). Most source samples exhibited signatures of groundwater influenced by anthropogenic sources (Burns and others, 2009). Of the three samples collected at the Culvert at Glenwood Road site, results of isotope dilution were mixed. In the summer wet weather sample, HF183 was relatively low and corresponded with the lowest observed fecal coliform concentration (800 MPN/100 mL) and the highest observed nitrate plus nitrite concentration (8.12 mg/L as N; δ¹⁸O = 1.60 ppt); whereas, in winter wet weather samples, nitrate plus nitrite concentration was lower and diluted (2.55 mg/L as N; δ¹⁸O = 4.19 ppt) by stormwater that carried higher fecal coliform concentrations (9,200 MPN/100 mL; table 1.2; U.S. Geological Survey, 2020a). This points to a combination of MS4 and septic influences that may be occurring from groundwater infiltration and illicit connections to the storm drain system, correlating with higher nitrate. A similar groundwater signature for nitrate and δ¹⁸O was identified at the Culvert at Tappen Beach site (which drains Scudders Pond); however, no HF183 was detected, suggesting that the OWDSs from homes upgradient and surrounding the pond are not contributing to elevated fecal coliform concentrations under dry or wet weather conditions.

Nonpoint Sources of Fecal Coliform Bacteria

Major potential nonpoint contributors of freshwater to Hempstead Harbor include groundwater seepage and direct stormwater runoff not conveyed by the MS4. Groundwater was collected from two sites, the Beach south of Glen Cove Creek and Tappen Beach sites. Fecal coliform was not detected in either groundwater sample (table 1.2; U.S. Geological Survey, 2020a). At the Tappen Beach site, which is downgradient from an unsewered residential area and the septic system of a seasonal restaurant, HF183 was detected in shallow groundwater (roughly 8 feet below land surface) at an estimated concentration of 400 copies/100 mL. The lack of positive detections for FC in groundwater throughout the study area, along with positive HF183 and high nitrate values (U.S. Geological Survey, 2020a) implies septic influence and strongly suggests that direct groundwater seepage to Hempstead Harbor is an unlikely contributor to the concentrations of fecal coliform observed in receptor samples.

Another nonpoint source that can potentially contribute to fecal contamination in water samples is sediment during resuspension (O’Mullan and others, 2019). To better understand these contributions, sediment samples were collected at the Culvert at Glenwood Road site and below the Spillway at Skillman Street site, with estimated fecal coliform concentrations of 260 and 510 colony forming units per gram dry weight, respectively (table 1.2; U.S. Geological Survey, 2020a). Additionally, human and waterfowl MST markers were also detected in these samples, with concentrations at the Culvert at Glenwood Road site of 830 copies/100 mL of HF183 and 49,000 copies/100 mL of GFD and at the Spillway at Skillman Street site of 76,000 copies/100 mL GFD. Recent research (Hassard and others, 2016; O’Mullan and others, 2019) shows that resuspension of sediment along natural shorelines can cause a spike in fecal coliform concentrations. Therefore, it is likely that during conditions that result in high wave energy, sediment can act as a source of FC when resuspended in the water column as a result of tidal movement, storms, or boat activity.

Culvert at Glenwood Road—Class 1

The samples collected at the Culvert at Glenwood Road site had some of the highest concentrations of both human and canine markers in the Hempstead Harbor watershed and elevated concentrations of FC in wet and dry weather (800 to 9,200 MPN/100 mL). The area is unsewered, with roughly 30 percent of the land use unspecified. Nitrate and isotope data point to diluted groundwater for the winter, wet-weather sample, suggesting runoff contributions that correlate with high human and canine markers and FC. The culvert flows continuously and is suspected of contributing high FC concentrations related to illicit sewer hookups to the storm drain system, which is influenced by groundwater infiltration under all conditions and stormwater that flows into the storm drain system from adjacent roadways during wet-weather conditions. The sediment at this location, which is exposed and can be in direct contact with the culvert flow, also may act as a source of FC when resuspended during tidal exchange or storms. The culvert is roughly parallel with the sandbar that acts as the regulatory divide between Hempstead Harbor North and South. Contributions from the culvert affect both Hempstead Harbor North and South through daily tidal exchange. Although stormwater is a big contributor at this location, there are other potential factors (for example, illicit connections of wastewater to drainage conveyances) that contribute to the consistent delivery of FC.

Outfall at Skillman Street—Class 1

The Outfall at Skillman Street site is adjacent to the new stormwater spillway (Spillway at Skillman Street site) and directs upland surface waters and some groundwater into the harbor. The population density in the surrounding area is high; 80 percent of the area is sewered, and 30 percent of the land use is residential, with the rest being a mix of commercial and industrial uses. Nitrate and isotope data point to diluted groundwater for the winter wet sample, suggesting upland runoff contributions are funneled through the system. The outfall is a consistent source of high FC that is influenced by humans, canines, and ruminants (deer). Human and canine markers were only detected in the dry, summer sample. Samples from this site had higher fecal coliform concentrations during dry weather (5,400 and greater than 16,000 MPN/100 mL) compared with wet weather (490 MPN/100 mL). This site is likely affected by upland surface-water drainage and infiltration of waste that appear to be diluted by stormwater runoff during wet weather events.

Glen Cove Creek—Class 1

Glen Cove Creek site is downstream from the Glen Cove Water Pollution Control Plant outfall and adjacent to a marina, with several stormwater outfalls along both sides of the creek. The creek was found to be a consistent source of FCs. Human, canine, and waterfowl markers were all detected; however, no human marker was present in the summer, dry-weather sample, collected during the summer boating season. The human marker was detected during both dry and wet weather at similar concentrations for two winter samples. More discrete and focused sampling would be needed throughout the year to confirm any influence from the marina. Although both winter samples showed comparable concentrations of HF183, the wet sample contributed orders of magnitude higher fecal coliform concentrations than the dry sample. This suggests that, although the human marker may be present and potentially originating from the Glen Cove Water Pollution Control Plant during dry weather, stormwater is a more substantial contributor of FC.

Pond Culvert at Hempstead Harbor South—Class 2

The Pond Culvert at Hempstead Harbor South site drains ponds and stormwater from the adjacent roadway. The area is 91 percent sewered and has a mix of land use, with 35 percent residential and 14 percent commercial. Two of the three samples had fecal coliform concentrations above 49 MPN/100 mL, which corresponded to a dry summer (500 MPN/100 mL) and wet winter (790 MPN/100 mL) sample (table 1.2; U.S. Geological Survey, 2020a). The HF183 marker was not detected in the samples collected at this site; however, BacCan was detected in both samples with elevated fecal coliform, suggesting canine influence regardless of season or condition. At this site, contributions of bacteria from pond drainage may be as substantial as contributions from stormwater.

Spillway at Skillman Street—Class 3

The Spillway at Skillman Street site (adjacent to the Outfall at Skillman Street site) drains stormwater from the roadway and (current [2019]) construction area. Fecal coliform concentrations were greater than 16,000 MPN/100 mL in the wet-weather sample and coincided with detections of waterfowl and canine markers. However, fecal coliform and canine markers were not detected in the dry weather sample. The human marker was not detected in either sample. Stormwater is the most substantial source of fecal contamination at the spillway.

Culvert at Tappen Beach—Class 4

The Culvert at Tappen Beach site drains upland ponds and possibly stormwater from the adjacent roadway. The area is unsewered, and land use is 39 percent residential and 27 percent recreational (mostly at Tappen Beach). FC concentrations under all conditions were relatively low. The waterfowl marker was the only marker present and was detected in both samples. This location is not a major contributor of fecal coliform to the harbor overall, though slightly higher concentrations in wet weather compared to dry suggest some stormwater contributions.