Site Selection

Information about sample sites is provided in table 1. Sites were selected, in consultation with the NYCDEP and NYSDEC, to include water and sediment sampling points adjacent to outfalls of specific concern related to the Tallman Island WRRF drainage area, along with freshwater sources tributary to Alley Creek. The outfalls are designated by the prefix “TI” (for Tallman Island) followed by an identifying number and are shown on figure 2. Outfall TI–025 (fig. 3A) is associated with the Alley Creek CSO Retention Facility, which became operational in 2011 and is designed to hold 5,000,000 gal of CSO that is then pumped to the Tallman Island WRRF for treatment (AECOM USA, 2014).

Oakland Lake (fig. 3B) flow to Alley Creek is approximately 2,500,000 gallons per day (gal/d) via outfall TI–008 (AECOM USA; 2014; fig. 3C). Oakland Lake is connected to the sewer system through a weir chamber to provide water-level control; the weir chamber is connected to a storm line, which is connected to a CSO line that then discharges from outfall TI–008 to Alley Creek (fig. 3C). With the construction of the Alley Creek CSO Retention Facility, CSO events are now rare at outfall TI–008 (AECOM USA; 2014). However, TI–008 had been receiving sewage from a parallel sanitary sewer leaking through a compromised joint (fig. 2) at approximately 400 feet (ft) upgradient of the outfall until it was repaired between April 27 and May 11, 2021. Both Oakland Lake and TI–008 were sampled as part of this study. About 500 ft downstream of TI–008 is TI–007, an emergency discharge outfall for a sewage pump station; TI–007 typically does not flow, even during wet-weather conditions, and the sediment sample for this study was collected on the opposite shoreline (fig. 3D).

Outfall TI–024 is an above-ground outfall that had been flowing continuously in recent years during both dry and wet weather (fig. 3E); but as of 2019, flow discharging from the outfall structure during dry weather had ceased. As of 2021, there is a hole in the top of the 7-ft-diameter sewer pipe that is exposed in the streambed immediately adjacent to the outfall from which water is continuously upwelling. Additional sampling was conducted at the hole location and is discussed later in this report.

Groundwater well Q277 (fig. 3F) is an artesian well with a total depth of 144 ft with flow that enters Alley Creek near TI–024 (USGS, 2021b). The well was originally part of a network of groundwater wells developed for public supply in Queens County and operated until the 1980s when the pumps were decommissioned and the structures removed. The flow reflects groundwater hydraulic pressure of the Magothy aquifer and discharges continuously with flow reaching Alley Creek near TI–024. Data from previous water-quality sampling conducted by the USGS in the 1990s and 2000s (USGS, 2021b) were reviewed for this study to determine the extent of urban land-use influence on the flowing groundwater.

Four sites in Little Neck Bay were selected to correspond with sampling locations used by the NYCDEP for either their Harbor Survey or the Sentinel Monitoring Program (SMP; table 1). LNB site 1 is at the mouth of Alley Creek and is near SMP station S1, and LNB site 2 is near SMP station S64 along the western shore of Little Neck Bay near Bayside Marina. LNB site 3 is near the Harbor Survey site LN1 in the middle of Little Neck Bay, and LNB site 4 is near SMP station S2 north of the mouth of a small creek and the discharge point for the Belgrave Sewage Treatment Plant (table 1; fig. 2).

Sample Collection and Processing

Procedures for collecting, processing, and cleaning samples for surface water, groundwater, and sediment were done according to the USGS National Field Manual for the Collection of Water-Quality Data (NFM; USGS, variously dated). Surface-water samples were either point or grab (dip), depending on accessibility (table 1). Field measurements of physicochemical constituents (water temperature, specific conductance, dissolved oxygen, pH, and turbidity) were collected at each site by using a YSI EXO multiparameter sonde (YSI, Inc.; Yellow Springs, Ohio). Water samples (point) at TI–025 and TI–008 were collected from an elevated perch directly above the outfall by using a peristaltic pump with sterilized Cole-Palmer Masterflex (Cole-Parmer: Vernon Hills, Illinois) pump head tubing connected to polytetrafluoroethylene (PTFE) intake tubing and were discharged into an 8-liter (L) PTFE churn splitter. Grab samples were collected at TI–024 and Oakland Lake by directly filling sample bottles roughly 0.5 ft below the water surface (TI–024) or from the flow before reaching the weir chamber (Oakland Lake). Cleaning of sampling equipment followed procedures outlined in the NFM (Wilde, 2004; Myers and others, 2007), except for sanitizing the PTFE churn, which was done in a bleach bath followed by thorough rinsing with sodium thiosulfate and deionized water. Field blank samples were collected to ensure cleaning and sterilizing procedures were adequate. Sample collection order of unfiltered water collected by spigot from the churn splitter was as follows: two high-density polyethylene 1-L bottles for MST, one 1-L glass wide-mouth jar for TSS, and two high-density polyethylene 500-milliliter (mL) bottles—one each for enterococci and fecal coliform. A 20-mL aliquot of sample water, for analysis of pharmaceutical compounds, retrieved directly from the churn splitter (TI–025 and TI–008) or from flow (TI–024 and Oakland Lake) via syringe, was passed through a 0.7-micrometer (µm) syringe filter into a 20-mL amber glass vial, placed on ice in the field, and frozen to a temperature of −20 degrees Celsius upon return to the office. Selected frozen dry-weather samples were later submitted for analysis of pharmaceutical compounds three times throughout the study. Criteria for submitting a sample for the pharmaceuticals analysis included detection of the human and (or) canine MST markers; all wet-weather and groundwater samples were submitted for analysis of pharmaceutical compounds; in all, 23 samples were submitted.

Bed-sediment samples were collected from the shoreline at three (TI–024, TI–008, and TI–025) of the four routine water-quality sampling sites plus an additional site (TI–007, table 1) during low tide from exposed shoreline by using a disposable sterile spatula to a depth of 2 centimeters (cm) from the surface in each area. Sediment was collected in 500-mL baked (at 450 degrees Celsius) amber glass jars filled with sediment to three-quarters full. Care was taken to only collect from points that did not appear disturbed by human or animal traffic. Sediment was partitioned for fecal coliform, enterococci, and MST-marker analyses.

In May 2021, samples in Little Neck Bay were collected during low (May 3) and high (May 10) tides by boat. Samples collected on May 10 also reflect wet-weather conditions because greater than (>) 0.1 in. of precipitation was recorded on May 9 (Weather Underground, 2021a, b, c). Grab samples were collected at each site by filling sample bottles directly at a depth of 0.5 ft below the water surface. Measurements of water temperature, specific conductance, and turbidity using a multiparameter sonde were also taken. Water samples were collected for fecal coliform, enterococci, TSS, and MST-marker analyses.

Thermal images were collected at TI–024 and the surrounding area by using an FLIR E8 handheld infrared camera (Teledyne FLIR Systems, Inc.; Wilsonville, Oregon) to determine if a difference in water temperature between areas of upwelling and surrounding surface waters was noticeable. Images were taken on May 3 and 10, 2021, corresponding to the depth-profile sample-collection events in the sewer-pipe hole at the streambed adjacent to the TI–024 outfall. Multiple temperature and specific conductance measurements were made as the multiparameter sonde was lowered incrementally into the hole to a maximum depth of 7.5 ft. Samples for TSS, fecal coliform, and enterococci were collected by using a peristaltic pump and sterile tubing at select depths corresponding to multiparameter measurements. Thermal images and temperature data, along with depth-profile data, are available in the USGS data release that accompanies this report (Fisher and others, 2021).

Laboratory Analyses

Samples collected throughout the study were analyzed at the USGS Ohio Water Microbiology Laboratory (OWML), the USGS National Water Quality Laboratory (NWQL), the NYCDEP Wards Island Process Control Laboratory (WIPCL), and the NYCDEP Newtown Creek Microbiology Laboratory (NCMBL). NYCDEP laboratories are accredited by the New York State Environmental Laboratory Accreditation Program (ELAP; https://www.wadsworth.org/regulatory/elap, Wadsworth, N.Y.) for TSS, fecal coliform, and enterococci in nonpotable water. The OWML analyzed water samples for MST markers and sediment samples for both MST markers and FIB, and the NWQL analyzed samples for pharmaceutical compounds.

Sediment Resuspension Assessment

A laboratory experiment was designed to replicate tidal activity in Alley Creek using sediment collected in the sewers and on the shoreline, as well as water collected from Oakland Lake. These sediment samples were assessed to understand the relation between sediment resuspension and FIB in the water column. A set of samples was collected at low tide along the combined sewer pipe connecting Oakland Lake to TI–008 (fig. 2) by using a disposable sterile spatula at points upgradient and downgradient of the observed sewage leak. Sediment deposited along the walls of the combined sewer pipe was above water during low tide but had been deposited as result of tidal exchange moving sediment from Alley Creek via the TI–008 outfall. Two water samples (1 L each) from Oakland Lake, one upgradient and one downgradient of the observed leak in the sewer pipe, were also collected. The combined sewer sediment samples were collected and processed according to methods described in chapter A7 of the NFM (Myers and others, 2007). Samples were shipped on ice overnight to the OWML.

For the sediment resuspension experiment, varying amounts of sediment were added to 250-mL Erlenmeyer flasks containing either 150 mL of sterile phosphate buffer or Oakland Lake water. Flasks prepared were as follows: (1) four buffer flasks with 1.8, 5.0, 10.2, or 30.3 g of combined sewer sediment added, (2) one Oakland Lake flask containing 10.3 grams of combined sewer sediment, (3) one negative control flask containing only 150 mL of buffer, and (4) one matrix control flask containing only 150 mL of Oakland Lake water (appendix 2; table 2.1). All seven flasks were secured onto a Burrell Model 95 wrist-action shaker (Burrell Scientific, Pittsburgh, Pennsylvania) and allowed to rest for 60 minutes. Before shaking started, a 12-mL sample from each flask was pipetted from each of the seven flasks to represent a time step of 0 minutes. The set was shaken at approximately 50 oscillations per minute, and samples were collected from each flask that was amended with sediment after 5, 30, and 240 minutes of shaking. For the negative control and matrix control, samples were collected at 240 minutes. All samples were refrigerated immediately after collection. Within 1 hour of collection, samples were analyzed by EPA Method 1600 (EPA, 2009) and Standard Method 9222D (Standard Methods Committee, 2017) to determine enterococci and fecal coliform concentrations, respectively. Methods and results from the sediment resuspension assessment can also be found in Fisher and others (2021).

Quality Assurance

Quality-control samples were collected to ensure sampling procedures were consistent and did not result in outside contamination of the environmental samples. A review of the quality-control data confirmed that neither sampling procedures nor laboratory data qualifiers resulted in environmental data being disqualified from interpretation. Field blanks were collected by pumping deionized water (August 24, 2020) or sterile phosphate buffer solution (December 9, 2020, and June 8, 2021) through peristaltic pump tubing and into the PTFE churn splitter; thus, the field blanks represent TI–025 and TI–008 sample procedures. Replicate sample sets were collected sequentially, where replicate sets from sites TI–025 and TI–008 were sequential fills of the PTFE churn splitter (following deionized water and native water rinses) and replicate sets from sites TI–024 and Little Neck Bay (LNB site 1; table 1) were sequential grab samples. Variability among paired replicates with at least one detected value above the reporting limit was compared by using relative percent difference for physicochemical constituents and absolute value log-10 difference (AVLD) for microbiological constituents (Francy and others, 2011; Tagliaferri and others, 2021a). The number of quality-control samples varied based on analysis because samples for FIB and TSS were collected more regularly.

Three blank and three replicate samples were collected for FIB. Concentrations of fecal coliform in two of the three blank samples were above the reporting limit of 2 CFU/100 mL: 14 CFU/100 mL on August 24, 2020, and 4 CFU/100 mL on June 8, 2021. These values were at least 4 times lower than the lowest environmental sample collected; therefore, environmental results were not qualified. No detections of enterococci were found in the blanks. The AVLD of replicates coliform ranged from 0.01 to 0.14 for fecal coliform and from 0.03 to 0.26 for enterococci.

Two blank and five replicate samples were collected for TSS. Concentrations of TSS in both blank samples were below the reporting limit (less than [<] 2.5 mg/L). The relative difference among the four of five replicates with both values above the reporting limit ranged from 0.21 to 0.61 percent, with the highest values reflecting replicate values of 8.0 and 15 mg/L.

Two blank and four replicate samples were collected for the MST markers. Concentrations for all six MST markers were below reporting limits in both blank samples. The AVLD for the four replicates varied by marker: 0.00–0.08 (GenBac), 0.00–0.05 (HF183/BacR287), 0.10–0.19 (BacCan), 0.04–0.12 (CPQ_056), and 0.04–0.11 (CPQ_064); concentrations of GFD were below the reporting limit for all paired samples. Of the duplicate qPCR results run for each sample for internal laboratory quantification verification, 10 of the values did not match and were qualified; however, these data were not excluded from interpretation.

Pharmaceutical method results were evaluated for a range of concentrations and based on isotope dilution standard recovery, method reporting limits, and result qualifiers according to Furlong and others (2014). Isotope recovery for one compound (acetaminophen) in one sample was outside acceptable range (422 percent); however, the total pharmaceutical concentration for the corresponding sample is such that increasing the environmental detection of the acetaminophen concentration would result in less than a 4-percent difference in total pharmaceutical concentration. Relative percent differences for total pharmaceutical compound concentrations for replicate samples collected at TI–008 (313 and 563 ng/L; August 24, 2020) and TI–024 (1,211 and 1,189 ng/L; June 17, 2021) were 57 and 1.8 percent, respectively (table 3). The greater percent difference at TI–008 reflects the variability in the dynamic tidal system; substituting both replicate values for the environmental concentrations did not change the statistical significance when comparing HF183/BacR287 to total pharmaceutical concentrations. Therefore, no adjustments were made to the dataset used for this assessment.

Data Availability and Analysis

Microbiological and physicochemical data from the field and laboratories are publicly available. Data from the routine water samples, bed-sediment samples, and Little Neck Bay samples can be accessed on the USGS National Water Information System website (USGS, 2021b) by entering the USGS station identifier (USGS station ID) listed in table 1 and selecting “Site Information” and then “Site Number” on the web page. Data from the sediment resuspension experiment and the TI–024 intensive sampling are available in the USGS data release that accompanies this report (Fisher and others, 2021).

Precipitation data from Weather Underground (stations KNYNEWYO1105, KNYNEWYO1408, and KNYNEWYO234; Weather Underground, 2021a, b, c) were used to determine whether rainfall occurred within 48 hours prior to sampling. While quality assurance on the data provided by Weather Underground stations could not be evaluated, precipitation data from several events within the study period were compared and were found to be in agreement on the basis of whether the total precipitation exceeded the criterion for wet-weather conditions of 0.1 in. 24 hours prior to the time of sample collection. Wet-weather samples were collected on September 30, 2020, May 10, 2021, and June 14, 2021; all other samples were collected during dry weather. For May 10, 2021, and June 14, 2021, rainfall totals exceeded 0.1 in. by the time of sample collection (Weather Underground, 2021a, b, c).

Additional analyses included summarizing pharmaceutical compound data and performing statistical analyses on FIB and MST marker data. Results from the pharmaceutical method are presented as number of detections and total concentration of pharmaceutical compounds for relative comparison to MST markers and FIB. A nonparametric analysis using Spearman Rank correlation (Helsel and others, 2020) was conducted using SigmaPlot 14 (Inpixon, Palo Alto, California) to determine correlations (ρ) between fecal coliform, enterococci, MST markers, and TSS across sites and with respect to the tidal cycle, and to compare HF183/BacR287 and total pharmaceuticals concentrations.

Routine Water-Quality Monitoring

Results from samples collected along Alley Creek are presented by site and in the context of tidal stage and sampling conditions (dry versus wet weather) and differences in known and potential sources contributing to each outfall location. Fecal indicator bacteria concentrations were not consistently detected during wet-weather conditions or low tide and ranged from <2 to 10,700 CFU/100 mL for fecal coliform (fig. 4A) and from <2 to 13,100 CFU/100 mL for enterococci (fig. 4B). Concentrations of TSS were indicative of water clarity and ranged from <2.5 to 616 mg/L (fig. 4C) for this study. The general Bacteroides marker, GenBac, was detected in all surface-water samples and was strongly correlated with the human Bacteroides marker HF183/BacR287 (ρ=0.965, P<0.001, where P is significance value), across the entire dataset (fig. 5A). Human Bacteroides marker HF183/BacR287 was prevalent (27 of 28 samples) at the three outfall sites sampled along Alley Creek but was not found in any of the eight Oakland Lake or two Q277 samples. Both CrAssphage markers—CPQ_056 and CPQ_064—were detected in 38 of the 39 water samples (including Little Neck Bay) with HF183/BacR287 concentrations above the reporting limit, and good correlation of indicator performance between the methods was observed (ρ=0.875 and 0.855, respectively; P<0.001), which has also been shown in a different study (Stachler and others, 2018). Canine marker BacCan was also commonly detected, although less prevalent (20 of 28 samples) in samples from the three outfall sites and occasionally in Oakland Lake, and showed good correlation with HF183/BacR287 (ρ=0.821, P<0.001). Waterfowl marker GFD was detected in 8 of 28 water samples collected at the three outfall sites and Oakland Lake.

At TI–025, concentrations of MST markers, concentrations of FIB, and both total number of detections and concentrations of pharmaceutical compounds were generally higher in wet-weather samples than in dry-weather samples. Concentrations of HF183/BacR287 in dry-weather samples ranged from 540 to 17,000 copies/100 mL (n=8; median 5,800 copies/100 mL) and were measured at 3,400,000 and 31,000,000 copies/100 mL for the wet-weather samples collected during low tide (fig. 5B), suggesting influence from sewage sources. The BacCan marker was detected in four of eight dry-weather samples at TI–025 (December 2020 and June 2021 samples only) and with roughly two orders of magnitude lower concentrations relative to the wet-weather samples of 130,000 and 870,000 copies/100 mL (fig. 5C). Further, BacCan concentrations were greater in samples collected at low tide relative to at high tide. The waterfowl-associated GFD marker was detected in the dry-weather samples from August (3,700 and 2,500 copies/100 mL) and December (1,500 and 3,800 copies/100 mL) and not in either wet-weather sample at TI–025. Concentrations of TSS ranged from 6 mg/L to 616 mg/L; the greatest concentration of TSS corresponded to a high concentration of fecal coliform (745 CFU/100 mL) and was detected during wet weather (figs. 4A, C). Fecal coliform concentrations ranged from 8 to 230 CFU/100 mL in dry-weather samples and from 80 to >6,000 CFU/100 mL in wet-weather samples (fig. 4A). Enterococci concentrations ranged from <2 to 360 CFU/100 mL in dry-weather samples and from 36 to >6,000 CFU/100 mL in wet-weather samples (fig. 4B). The number of pharmaceuticals detected in the two dry-weather, low-tide samples were 9 (August 24, 2020) and 3 (June 17, 2021), with total concentrations of 237 and 71.3 ng/L, respectively (table 3). One wet-weather sample (September 30, 2020; low tide), with 25 pharmaceutical compounds detected and a total concentration of 3,907 ng/L, was also collected. Interestingly, concentrations of pharmaceutical compounds on December 14, 2020 (designated as dry weather for this study) were 885 ng/L with 17 compounds detected, whereas concentrations and number of compounds detected for June 14, 2021 (designated as a wet-weather sample), were 56.8 ng/L and 2, respectively. The lower concentrations and number of detected compounds on June 14 is likely because the retention facility did not release during the earlier part of the storm during sample collection.

At TI–008, concentrations of MST markers, concentrations of FIB, and both total number and concentrations of pharmaceutical compounds were generally higher in low-tide samples than in high-tide samples. Human marker HF183/BacR287 was detected in seven of the eight dry-weather samples (n=8, median 53,000 copies/100 mL; no wet-weather samples for MST markers were collected at this site). Of the six dry-weather samples collected between August 2020 and March 2021, concentrations of HF183/BacR287 were greater in low-tide samples (median 1,100,000 copies/100 mL) relative to high-tide samples (median 8,300 copies/100 mL), reflecting less dilution of the inputs from the leaking sanitary sewer parallel to the combined sewer between Oakland Lake and TI–008 (fig. 5B; fig. 2). Samples collected in June 2021 (HF183/BacR287 concentrations of 970 and <330 copies/100 mL) likely reflect repairs to the sanitary sewer conducted roughly 1 month prior to sample collection. Similar to what was observed at TI–025, the canine marker concentrations were also greater in samples collected at low tide than at high tide and likely reflect a combination of sewage and tidal dilution and would include pet waste inputs from Oakland Lake (fig. 5C). Only two of the eight samples had concentrations of the GFD marker above the reporting limit (1,900 and 9,700 copies/100 mL), both of which were collected in December and are possibly related to the fall migration of waterfowl (Matic and others, 2015). Concentrations of TSS ranged from 8 to 206 mg/L (fig. 4C). Fecal coliform and enterococci concentrations ranged from 8 to 12,000 CFU/100 mL (fig. 4A) and <2 to 13,100 CFU/100 mL (fig. 4B), respectively. On May 10, the highest observed concentrations of TSS (206 mg/L), enterococci (190 CFU/100 mL), and fecal coliform (800 CFU/100 mL) were from a sample collected during high tide following a rainfall event and possible CSO overflows the night before that resulted in sediment being resuspended throughout Alley Creek (fig. 4; samples for MST markers were not collected from Alley Creek outfalls on May 10). Four dry-weather samples—two during low tide (August 24, 2020, and June 17, 2021) and two during high tide (August 31 and December 14, 2020)—were analyzed for pharmaceutical compounds (table 3). The number of compounds detected in the three 2020 samples ranged from 8 to 16, whereas in June 2021, only 2 compounds were detected. Total concentrations of pharmaceutical compounds from 2020 samples ranged from 313 to 753 ng/L, whereas the concentration was 62.0 ng/L in the June sample (table 3), which is indicative of improved water quality emerging at TI–008; again, likely the result of stopping the sewage leak.

At TI–024, concentrations of MST markers and both total number and concentrations of pharmaceutical compounds were lower in wet-weather samples than in dry-weather samples, whereas concentrations of FIB were high. Concentrations of human MST markers were consistently high during dry-weather samples (HF183/BacR287 ranging from 190,000 to 4,900,000 copies/100 mL; median 950,000 copies/100 mL), which was similar to low-tide samples collected at TI–008 (fig. 5B). The two wet-weather samples appeared to be diluted with respect to human MST markers, with low marker concentrations for both September and June samples (570 and 32,000 copies/100 mL, respectively). Canine marker concentrations did not appear to be affected by precipitation, with the overall range from <460 to 1,700,000 copies/100 mL (median 16,500 copies/100 mL; fig. 5C). Of the 10 samples collected, the waterfowl marker GFD was only detected in one sample, from December 14, 2020, and was the highest concentration of the waterfowl marker detected in this study (21,000 copies/100 mL). Concentrations of TSS ranged from 2.5 to 31 mg/L and were not affected by wet weather for this study (fig. 4C). The ranges of fecal coliform and enterococci concentrations were greatest in the August 2020, December 2020, and June 2021 dry-weather samples, with maximum concentrations of 10,700 CFU/100 mL fecal coliform (August 24; fig. 4A) and 1,500 CFU/100 mL enterococci (August 31; fig. 4B). Concentrations of fecal coliform and enterococci in samples from March 2021 were both substantially lower, with concentrations of fecal coliform 7 times lower than the minimum concentration from the other three seasons. Results from the analysis of six dry-weather and two wet-weather samples for pharmaceutical compounds supports the finding based on high HF183/BacR287 marker concentrations that TI–024 is influenced by sewage (table 3). The number of compounds detected among the samples ranged from 6 to 12, and total concentrations ranged from 333 to 1,757 ng/L, inclusive of dry- and wet-weather samples.

At Oakland Lake, human MST markers were not detected in any of the eight samples (fig. 5B). The BacCan was detected in four of the samples, with concentrations ranging from 4,600 to 9,400 copies/100 mL (fig. 5C). Only one sample was positive for the waterfowl marker (1,500 copies/100 mL; December 9, 2020), and the detection coincided with one of the canine-marker detections and was possibly connected to the fall waterfowl migration (Matic and others, 2015). Fecal coliform and enterococci concentrations at Oakland Lake were typically low (especially relative to concentrations at TI–008), particularly for enterococci, which ranged from <2 to 100 CFU/100 mL (fig. 4A, B). Pharmaceuticals analysis from two Oakland Lake samples contained one (August 24, 2020) and two (December 9, 2020) compounds with total concentrations of 1.80 and 52.7 ng/L, respectively (table 3), which is consistent with little to no sewage influence.

Concentrations of all MST markers were below reporting limits in both groundwater samples collected at Q277, and very low concentrations of enterococci (6 CFU/100 mL) and fecal coliform (2 CFU/100 mL) were observed in one of the two sample sets collected. The total concentration of pharmaceutical compounds detected at Q277, from the two samples submitted, was the result of a single compound—carbamazapine—which was at a greater concentration than in any other sample for this study (USGS, 2021b); carbamazapine is regularly detected in groundwater influenced by urbanization and onsite wastewater disposal (Benotti and others, 2006; Bexfield and others, 2019).

Samples collected in Little Neck Bay represent open-water receptor sites as described in previous MST studies conducted in Nassau and Suffolk Counties (for example, Tagliaferri and others, 2021b). Human MST markers (HF183/BacR287, CPQ_056, and CPQ_064) were detected at all four sites (fig. 2) on May 3 and 10, 2021, and were at greatest concentration at LNB sites 1 (near the mouth of Alley Creek) and 4 (eastern part of the bay near Nassau County border; table 4). The BacCan marker was also detected in all samples, with the greatest concentrations again at LNB sites 1 and 4; at LNB site 1, concentrations were higher on May 10 (23,000 copies/100 mL), possibly because of precipitation the night before, than May 3 (4,200 copies/100 mL). The waterfowl marker was not detected in any of the Little Neck Bay samples, which was consistent with samples collected in Alley Creek and Oakland Lake in March and June 2021. Fecal coliform concentrations were greater on May 10 at three of the four sites, with the greatest concentration (600 CFU/100 mL) at LNB site 1. Enterococci concentrations were similar between dates, except at LNB site 4 on May 10, which was 73 CFU/100 mL, compared to 20 CFU/100 mL on May 3.

Sediment and Resuspension

The four sediment samples, collected in September 2020 at the routine monitoring sites (table 2), provided a snapshot of FIB and MST marker concentrations that are bound in the bottom material and tidally influenced combined sewer lines of the Alley Creek watershed. Fecal coliform concentrations ranged from 1,400 to 15,000 colony-forming units per gram dry weight (CFU/g dw) (median 2,950 CFU/g dw), and enterococci concentrations ranged from 1,300 to 8,500 CFU/g dw (median 1,950 CFU/g dw; USGS, 2021b). The sediment collected at TI–025, TI–008, and TI–007 contained greater amounts of organic materials, representative of the tidal marsh and consistent with information reported in the Alley Creek LTCP (AECOM USA, 2014, p. 2–43), compared to the sandier sample collected at TI–024. The GenBac marker was present in all four samples with concentrations ranging from 290,000 to 7,000,000 copies/100 mL (median 1,700,000 copies of genes per gram dry weight [copies/g dw]). The HF183/BacR287 marker was not detected in the sediment sample at TI–007 and ranged from 4,000 to 34,000 copies/g dw in the other three outfall samples (median 14,900 copies/g gw). The BacCan marker was not detected in any of the samples, which was consistent with many other samples collected in Nassau and Suffolk Counties from a similar study (Tagliaferri and others, 2021b). The GFD marker was only detected in one sample (site TI–025; 49,000 copies/100 mL).

Sediment that had been deposited by tidal fluctuations was collected from the exposed area of the pipe above the water line inside the combined sewer connecting Oakland Lake to TI–008 and was used to demonstrate the potential for resuspension of sediment ladened with fecal contaminants from damaged infrastructure (the leaking sanitary sewer line) to contribute fecal bacteria to Alley Creek. Laboratory-produced sediment resuspension results are presented in table 2.1 and available in Fisher and others (2021). Enterococci and fecal coliform concentrations in the combined sewer sediment were 25,000 and 54,000 CFU/g dw, respectively. These concentrations were substantially greater than concentrations in shoreline sediment samples, including at TI–008 (8,500 CFU/g dw enterococci and 15,000 CFU/g dw fecal coliform). Prior to oscillation of the sample to simulate tidal disruption (time step=0 minutes), concentrations in the supernatant liquid of fecal coliform ranged from 60 to 170 CFU/100 mL and of enterococci ranged from 67 to 290 CFU/100 mL for the varying amounts of sediment added to the buffer and Oakland Lake water samples (table 2.1). Generally, concentrations of both FIB increased in the water matrix over time during oscillation, though concentrations of FIB did not increase as a function of the mass of sediment added to the slurry. After 240 minutes of oscillation, concentrations of fecal coliform ranged from 1,500 to 6,300 CFU/100 mL and of enterococci ranged from 540 to 2,800 CFU/100 mL for the varying amounts of sediment added to the buffer and Oakland Lake water samples (table 2.1).

Intensive Sampling at TI–024

The focused sampling at TI–024 revealed variability of FIB concentrations and specific conductance with depth within damaged infrastructure that is inundated by shallow groundwater and flowing continuously from the hole in the storm-sewer pipe and streambed adjacent to outfall TI–024 structure (fig. 6). Results from the samples collected on May 3 and 10, 2021, at TI–024 are available in Fisher and others (2021) and were useful for understanding the extent of wastewater that may become entrained in the groundwater upwelling. With the first depth profile, specific conductance values changed substantially with depth, ranging from 612 to 8,300 microsiemens per centimeter (μS/cm) at 25 °C from surface to bottom (fig. 6A); whereas specific conductance values collected during the second profile (conducted following a large precipitation event) were more uniform throughout, ranging from 304 to 438 μS/cm at 25 °C (fig. 6B), suggesting flushing of the stratified water in the pipe with a large amount of runoff. Similarly, fecal coliform and enterococci concentrations were greater at depth on May 3 (fig. 6A) but were elevated throughout the entire depth on May 10, 2021 (fig. 6B).

Thermal images were also collected around TI–024 by using an FLIR E8 handheld infrared camera to determine if a difference in water temperature was noticeable at and around the hole in the underground pipe leading to TI–024 (fig. 7; Fisher and others, 2021). Thermal gradients around the elevated outfall (fig. 7A), at the hole in the pipe (fig. 7B), and from areas of groundwater seepage (figs. 7C, D, and E) near TI–024 were assessed and compared to thermistor readings at the water surface using a YSI EXO sonde with a temperature probe (Fisher and others, 2021). No differences in surface temperatures were determined at the point of upwelling and at nearby and downstream points that would suggest significant thermal contributions from a different source (for example, a sewage leak).

Temporal Variability

Results from the two wet-weather samples at TI–025 and TI–024 showed different patterns of MST markers despite elevated FIB concentrations during both events. Precipitation preceding sample collection on September 30, 2020 (Weather Underground, 2021a, b, c), was sufficient to exceed the capacity of the 5,000,000-gal retention facility (NYCDEP, 2020), resulting in a combined sewer overflow at TI–025—this overflow is evidenced by the HF183/BacR287 marker concentrations detected at TI–025 that were three orders of magnitude higher in the September 30, 2020, wet-weather sample than in the preceding dry-weather samples. In contrast, dilution of the sewage signature discharging from the hole in the damaged storm sewer adjacent to TI–024 appears to have occurred and is reflected in the relatively lower concentrations of human MST markers in wet-weather samples as compared to dry-weather samples; however, concentrations of total pharmaceutical compounds and FIBs did not decrease proportionally and were within the same range as the dry-weather samples. Therefore, it is possible that contributions from stormwater infrastructure (for example, a storm drain or grit chamber) containing high concentrations of FIB resulted in the observed elevated FIB concentrations that were otherwise persistent in samples collected between August and December 2020.

There was a substantial difference in FIB and HF183/BacR287 concentrations during the low-tide sample collection following repair to the sanitary sewer running parallel to the combined sewer between Oakland Lake and TI–008. Concentrations of fecal coliform and HF183/BacR287 in the June 17, 2021, sample were 12 CFU/100 mL and <330 copies/100 mL, respectively, compared to low-tide concentrations of fecal coliform that ranged from 2,000 to 12,000 CFU/100 mL, and HF183/BacR287 that ranged from 170,000 to 3,700,000 copies/100 mL in samples prior to June 17, 2021. According to the NYCDEP (Jorge Villacis, NYCDEP, written commun., August 2021), repairs at TI–008 were conducted between April 27 and May 11, 2021. Whereas the precise date the leak stopped contributing to the TI–008 discharge is unknown, it was likely before May 3, 2021, based on low-tide concentrations of FIB at TI–008 on that date (12 CFU/100 mL fecal coliform and 46 CFU/100 mL enterococci, which were 26- and 7-times decreases, respectively, from the lowest concentrations collected for the previous three low-tide samples). Furthermore, the decrease in human markers and number and concentration of pharmaceuticals at TI–008 at low tide in such a short period of time between repairs and the June 2021 samples indicates the utility of combined analyses as a means of verifying successful implementation of corrective actions to sewage infrastructure through time.

Seasonal variability of canine and waterfowl MST markers was observed at the three outfall sites and Oakland Lake, while human markers were not observed to vary seasonally but did vary based on the tide and precipitation. Canine marker BacCan was detected in each of the December 2020 and March 2021 samples collected but infrequently in the August 2020 sample; this difference may be related to hotter summer months resulting in fewer dog walkers (Hinojosa and others, 2020) or fewer rain events causing runoff of parkland to the lake. Waterfowl markers were detected in at least one of the December 2020 samples from each site; the only other sample with a waterfowl marker concentration above the detection limit was TI–025 in August 2020. As was observed in an adjacent embayment during a 2018–19 study (Tagliaferri and others, 2021b) and in other MST studies (Weiskel and others, 1996), migratory waterfowl present in the watershed in spring and autumn (Matic and others, 2015) can directly contribute to bacterial contamination in receiving waters.

Spatial Variability

For much of the study period, there was a clear distinction between the three outfall sample sites with respect to concentrations of FIB and MST markers. Concentrations of FIB were consistently high at TI–024 and generally higher than at TI–025 under dry-weather conditions (figs. 4A, B). Dry-weather samples collected at TI–024 exhibited a range of concentrations of HF183/BacR287 similar to low-tide samples at TI–008, where a known source of sewage was influencing the water at the outfall. Groundwater from well Q277 was not found to be a source of bacteria or MST markers to Alley Creek, which was not surprising since Q277 is sourced from the deeper Magothy aquifer. Based on previous studies of shallow groundwater known to be affected by upgradient onsite wastewater disposal, fecal contamination is not likely because of the sandy substrate acting as a filter and limiting transport in the groundwater (Fisher and others, 2020; Tagliaferri and others, 2021b).

Correlation between either fecal coliform or enterococci and TSS concentrations was not observed at TI–024 and Oakland Lake and was mixed with respect to tide at sites TI–025 and TI–008 (table 2.2) in dry-weather samples. Correlation of both fecal coliform and enterococci with TSS was strong during high tide (ρ=0.880, P<0.05) when data from TI–008 were evaluated independently and when data from both TI–008 and TI–025 were combined; however, there was no correlation during low tide. Concentrations of TSS were 6 and 86 mg/L at TI–024 and TI–025, respectively, in the September 30, 2020, wet-weather samples, where both enterococci and fecal coliform were both >6,000 CFU/100 mL at each site. Conversely, TSS concentration in the May 10, 2021, wet-weather sample at TI–008 was 616 mg/L, and concentrations of enterococci and fecal coliform were lower than from September wet-weather samples. The overall lack of correlation between FIB and TSS can be attributed to factors including influences from sewage at TI–008 (particularly, dilution at high tide) and potential sampling interferences (such as bed sediment in the peristaltic pump sampling line during low tide).

The co-occurrence of the canine-associated BacCan and human-associated HF183/BacR287 MST markers was prevalent in this study, with 80, 71, and 60 percent of samples containing both markers at sites TI–024, TI–008, and TI–025, respectively. While studies have documented detection of both markers in sewage samples, often attributed to the cohabitation of humans and dogs (Kildare and others, 2007; Silkie and Nelson, 2009), differences in BacCan to HF183/BacR287 ratios throughout the study area and patterns at each site can be used to parse sources. Results from TI–025 and TI–024 show a similar pattern with respect to dry- and wet-weather conditions and, for TI–025, tidal stage (figs. 5B and C), which suggests coinciding sewage influence. At TI–008, the samples collected prior to June 2021 were influenced by both Oakland Lake and sewage and reflect both human and canine markers in most samples. In the June 17, 2021, (low tide) sample at TI–008, following repair of the sanitary sewer line, HF183/BacR287 was not detected above the reporting limit, but BacCan was detected, suggesting sources other than sewage (such as Oakland Lake). In samples collected at Oakland Lake in 2014 as part of the Alley Creek LTCP, one of five sample sets tested positive for human markers and no canine markers were detected (AECOM USA, 2014). In this study, only canine and waterfowl markers were detected (fig. 5), suggesting that dog or waterfowl waste regularly enters the lake through runoff from nearby sidewalks and roadways or other upstream sources, but there is no association with sewage that would also contain human markers.

Results from Little Neck Bay samples collected in May 2021 allowed for comparison of MST markers between source (that is, outfall sites along Alley Creek) and receptor (that is, Little Neck Bay). Interestingly, all three human MST markers—HF183/BacR287 and both CrAssphage—were detected in all samples along with similar patterns of BacCan, indicating widespread influence of sewage throughout the bay, although the source may also be from treated sewage and could explain the higher concentrations of HF183/BacR287 at LNB site 4 relative to the other three LNB sites because the Belgrave Sewage Treatment Plant discharges effluent nearby (fig. 2). However, this was not the case in an adjacent embayment to the northeast of the study area, Hempstead Harbor, sampled in 2018–19, where roughly 60 percent of samples from five open-water receptor sites detected HF183 above the method reporting limit (Tagliaferri and others, 2021b). The concentrations of HF183/BacR287 in Little Neck Bay were within the range found in Alley Creek, regardless of tidal stage, indicating consistent presence of markers likely being contributed from the East River and other wastewater treatment facilities within the Little Neck Bay watershed. If Alley Creek were the only source, it would be expected that concentrations of MST markers would be considerably lower because of dilution and marker decay (which one study of marine waters determined is roughly 6 days; Mattioli and others, 2017). Unfortunately, current analytical methods cannot readily discern whether Bacteroides or crAssphage viral indicators were introduced to the waterways by raw or treated sewage, or by surface runoff.

Decision Support and Future Work

Findings from this study may help guide future sampling and management strategies of outfalls along Alley Creek and provide additional context for persistently high concentrations of FIB that have been observed during regulatory compliance monitoring. Concentrations of FIB determined during this study and previous NYCDEP monitoring at TI–008 and TI–024 led to investigations of the sewer system and upstream sources. Sealing leaks, correcting illicit connections, and rerouting sewage away from CSO outfalls to retention tanks has been shown to improve water quality (AECOM USA, 2014). However, evidence of past and current bacterial contamination at TI–008 and TI–024 was found. Factors such as bed-sediment resuspension during tidal fluxes, wind or storm events, and the potential for sampling bias (particularly during low tide) was shown to increase detection of FIB in water samples from the resuspension experiment, despite TSS and FIB not correlating statistically at TI–008 and TI–025 at low tide or at TI–024 and Oakland Lake. There was also evidence of upgradient combined sewer line infiltration of sewage prior to upwelling along the way to outfall TI–024.

The sanitary sewer running parallel to the combined sewer that was leaking through a compromised joint was repaired between April 27 and May 11, 2021, thereby ceasing the constant infiltration of sewage into Alley Creek from TI–008. The clear difference in FIB and MST marker concentration in water samples collected at TI–008 before (six samples; fecal coliform concentrations ranged from 2,000 to 12,000 CFU/100 mL [low-tide samples] and from 8 to 321 CFU/100 mL [high-tide samples]) and after (two samples; fecal coliform concentrations were 12 and 40 CFU/100 mL) sewer line repair, as well as the order of magnitude difference between storm-sewer sediment and sediment at the TI–008 outfall, indicates the continued dosing of sediment within the line has likely been mitigated. Further consideration to the accumulation of sediment along the combined sewer line may be warranted given the concentrations of fecal coliform and enterococci in the sewer sediment just upgradient of the former leak (54,000 CFU/g dw and 25,000 CFU/g dw, respectively; Fisher and others, 2021). According to the Alley Creek LTCP, sediment within larger sewer lines must be removed as it accumulates; however, the sediment buildup within the combined sewer between Oakland Lake and TI–008 is the result of tidal flux moving and depositing bed sediment upgradient. Although these may be considered natural conditions, bacteria could potentially continue to incubate in the sediment (and thereby continue to pose a threat to downstream water quality) given the limited exposure to sunlight/UV radiation (Myers and Juhl, 2020). Outfall redesign to reduce sediment transport into the sewer line and (or) seasonal cleaning of sediment accumulating along the walls of the sewer at low tide may reduce source material that is resuspended in the line during tidal shifts and wet-weather events. Continued monitoring of TI–008 for FIB and TSS by the NYCDEP Harbor Survey Program (their site AC2), along with the addition of human MST markers and documentation of tidal stage and antecedent weather conditions, may indicate whether further inspections of the combined sewer line and cleanings are necessary. Additional bed-sediment samples and a more intensive resuspension experiment could be conducted to better quantify contributions of FIB to Alley Creek during events such as tidal exchanges, wakes, and storms that result in resuspension.

Within the drainage area of Alley Creek (including the nontidal reaches where TI–024 is located), the NYCDEP has been working for over a decade to track down illicit connections from buildings to storm-sewer pipes (AECOM USA, 2014) and has conducted dye testing, all of which has led to mitigation through repairs and disconnections of improper hookups in the drainage area. Within the past few years, the focus has been on the shoreline of Alley Creek and infrastructure integrity immediately surrounding TI–024. Although constant flow from the aboveground outfall TI–024 ceased during dry weather, upwelling can be observed adjacent to the outfall from a hole in the top of the storm-sewer pipe leading to it. It is possible that this flow is the result of hydraulic pressure causing groundwater to infiltrate into the damaged pipes at an upgradient location and then to emerge from the hole. Concentrations of FIB, HF183/BacR287, and pharmaceutical compounds from samples collected at TI–024 were consistently high, strongly suggesting that a persistent source of sewage is influencing the water upwelling at the outfall. This influence is further evidenced by the elevated concentrations of FIB throughout the depth profile observed during both sampling events, including after a storm that effectively flushed the pipe (thereby removing the density gradient). Since tidal influence is not regularly observed, sediment resuspension does not appear to be a major factor. Routine sample collection along a vertical gradient (that is, depth profiles) at the adjacent hole before and after each stage of future repairs upgradient (for example, cleaning inline grit chambers) could provide data needed to track progress and relative changes in FIB concentrations (and therefore, loads to Alley Creek) as investigations by the NYCDEP into probable causes continue. Following repairs, incorporating MST markers into the sample analysis along with FIB and TSS should improve monitoring of remedial work and for possible recurrence of sewage contamination.

Given the reliance on field-collected samples in determining compliance, care should be taken during sample collection, particularly during low tide in areas with shallow depth, to not introduce bias by disturbing bottom material or bed sediment. Routine documentation of water clarity, sampling depth, weather conditions, and tidal stage during the sample-collection process can provide necessary metadata for interpretation of data. Although not directly correlated at all sites and under all conditions during this study, collecting samples for TSS in addition to FIB may also be useful for understanding fluxes in solid particulates related to sediment and wastewater contributions.

The multiple lines of evidence approach used for this study (FIB analysis coupled with analyses for TSS, MST, and pharmaceutical compounds) proved useful for understanding the complex nature of fecal contamination inputs to Alley Creek. The bacterial and viral MST marker data, along with FIB data, were helpful in differentiating host sources (human, canine, and waterfowl) and watershed sources (outfalls, groundwater, and Oakland Lake) from receptor sites (Little Neck Bay). The CrAssphage analytical method used in this study showed a very good correlation between the two viral-MST human markers (CPQ_056 and CPQ_064) and the bacterial-MST human marker (HF183/BacR287) and that all were equally effective in detecting human influence. The viral CrAssphage marker may offer greater sensitivity and persist longer than Bacteroides and thus be more effective when human host sources are lower, or when samples become diluted (Stachler and others, 2018). Total concentrations of pharmaceutical compounds and HF183/BacR287 correlated marginally (ρ=0.585, P=0.005; n=21) and allowed a differentiation of samples with recent sewage inputs from ambient groundwater conditions.

Initially, additional sampling locations were sought along Alley Creek to assess alternative sites; however, thick, muddy bed sediment and an incised channel limited access by foot or boat and made sampling during low tide impractical. Additional samples collected along Alley Creek downgradient and upgradient from outfalls TI–024, TI–008, and TI–025 could serve to improve the dataset, especially at sites slated for remediation or repairs. Furthermore, verification by field crews that the model for discharge at TI–025 is accurate, or that lower precipitation amounts can cause overflow, may be warranted to better understand conditions observed on June 14, 2021. A field-deployed camera focused on the outfall at TI–025 collecting images at routine intervals and an in situ microbiological analyzer could also serve to confirm proper functioning at TI–025 and could also help evaluate remediation or wastewater system upgrades that are planned or carried out in the future.