Assessing impacts of cemeteries on water quality in an urban headwater watershed with mixed human‐built infrastructure

Cemeteries are understudied integral components to urban watersheds, which provide ecosystem services but can also export nutrients, trace elements, and other contaminants to nearby water bodies. In this study, we focus on Meadowbrook Creek, an urban headwater stream in Syracuse, New York (USA), which has shown significant nitrate contributions from a local cemetery. We collected biweekly surface water samples over the course of 1 year from 2022 to 2023 for analysis of major and trace elemental concentrations including Na, Ca, Mg, K, F, Cl, sulfate, and nitrate. Here, we aim to assess the impact of various human infrastructures on urban stream water quality with a particular focus on the cemetery and nitrate. A comparison between the new dataset in this study and previously reported water chemistry data in Meadowbrook in 2012 suggests a decade‐long impact of road salting and the cemetery on water quality particularly with respect to Na, Cl, and nitrate. Sulfate, Mg, Ca, and K are likely mainly geogenic. Stable nitrogen isotope data, the usage of concrete or steel vaults in the cemetery in the past 50 years, and the lack of correlation between nitrate and fluoride concentrations in stream water argue against burial decay products being a major source of nitrate to the stream. Instead, other nitrate sources that exist in the cemetery such as, fertilizer, decaying plant material, and wastewater, are more viable dominant nitrate sources. In addition, nitrate loading calculations indicate that the groundwater‐connected reach, including the cemetery, acts as an annual net sink for nitrate despite the seasonally varying sink‐source patterns.

from cemeteries is under-explored, partially due to their cultural and religious significance.
Unlike natural streams, urban streams generally have highly compromised or no riparian zones compared with streams in unimpacted natural areas, which if present, would facilitate the uptake and degradation of some pollutants (Ledford et al., 2016;Meyer et al., 2005).
The presence of excessive pollutants and the absence of riparian zones make urban streams particularly vulnerable to anthropogenic pollution, including nutrients, major ions, and trace elements (Kaushal et al., 2008).Increased nutrient (e.g., nitrate) and major ion (e.g., sulfate, chloride, and sodium) inputs to urban streams are commonly attributable to road salt, fertilizers, sewage wastewater, and atmospheric deposition (Cañedo-Argüelles et al., 2013;Divers et al., 2014).Nitrate is of particular concern for water management due to its role as a limiting nutrient in many ecosystems.Increased nitrate loads can lead to algae blooms, which can be toxic and reduce dissolved oxygen, straining aquatic habitats.Chloride, another analyte of concern, which is dominant in road salt, can pose harm to urban infrastructure, aquatic life, and human health when present in water at elevated levels (Pieper et al., 2018;US EPA, 2022).Urban green spaces, including cemeteries, can help alleviate some water quality issues associated with urban infrastructure by increasing groundwater infiltration and serving as a riparian buffer.
Only a limited number of previous studies assessed the environmental impacts of urban cemeteries on water quality.For example, Brennan et al. (2018) found higher nutrient and trace metal concentrations in groundwater directly downgradient a cemetery in Lansing, Michigan, (USA).In addition, Lautz et al. (2020) examined concentration change in major ions such as nitrate, sulfate, sodium, and chloride upstream and downstream of St. Mary's Catholic Cemetery along Meadowbrook Creek, an urban creek in Syracuse, New York (USA), over the course of 1 year from 2012 to 2013.High frequency nitrate loading upstream and downstream of the cemetery from 2017 to 2019 was also previously estimated (Beltran et al., 2021).Lautz et al. (2020) observed an increase of 20-40 kg NO 3 À /day in stream water flowing through the cemetery in the months of June-September, based on which Lautz et al. hypothesized  up to 25%-50% of such nitrate increase could plausibly derive from burial decay products.However, the release of burial decay leachate to the stream has not been confirmed.Leaky urban infrastructure and fertilizers were also suggested as additional sources of nitrate in the Meadowbrook watershed overall, and nitrate concentrations in Meadowbrook are also controlled by seasonal and in-stream processes such as plant uptake, denitrification, and organic matter decomposition (Lautz et al., 2020;Ledford et al., 2017).
In this study, we focus on Meadowbrook Creek to assess the environmental impact of various urban infrastructure and activities (e.g., St. Mary's Catholic Cemetery and road salting), and geogenic sources on urban stream water quality with a particular focus on the role of the cemetery on nitrate.Meadowbrook Creek was selected due to the availability of rich historic water quality data at the site.
Here, we assess the decadal change and seasonality of solute leachate released from the cemetery and other sources.We also evaluate the control of hydrologic connectivity and processes on nitrate loading in an urban stream.We hypothesize that the cemetery continues to be a seasonally varying, important source of nitrate to Meadowbrook Creek with an increase in cemetery-originated nitrate loading in the last decade due to further decay and continual addition of burial products.We also hypothesize that sodium and chloride concentrations have increased due to the continued use of road salt within the watershed.

| STUDY SITE
Meadowbrook Creek is an urban headwater stream located in Syracuse, New York (USA).The stream is sourced from a stormwater pond and flows east before discharging into Butternut Creek in Dewitt, New York.The stream is approximately 5.5 km long with a drainage area of 11.3 km 2 (Figure 1).Flow is perennial and it does not freeze over.Land use in Meadowbrook Creek watershed is primarily residential (62% of the drainage area) with the rest consisting of schools/ religious institutions (10%), vacant land (9%), and cemeteries (9%) (Lautz et al., 2020).
The stream is much more urbanized in the upstream reach than in the downstream reach.Based on Ledford and Lautz (2015), the upstream watershed has 30% low impervious surface cover (ISC), >50% moderate ISC and 9% high ISC.The downstream watershed has 46% low ISC, 49% moderate ISC and 4% high ISC.Low, moderate and high ISC correspond to <20%, 20%-49% and >50% ISC.Much of the stream flows through the middle of Meadowbrook Drive, a heavily trafficked residential road.The upper 3.5 km of the stream is situated in this heavily urbanized area and flows through an artificially straightened and concrete-lined channel, which will be referred to as the 'disconnected reach' following Ledford and Lautz (2015) due to its limited connection to groundwater.In addition, this upstream portion is characterized by the absence of a riparian zone or floodplain, increased impervious surface area relative to the downstream portion, numerous culverts, and low baseflows.The lower 2 km of the stream, once again outlined by Ledford and Lautz (2015), has less artificial alteration and will be referred to as the 'connected reach' due to the strong connection to groundwater, that is, groundwater and stream water are free to interact through permeable surfaces in the stream channel, also known as the hyporheic zone.The downstream groundwater-connected reach is historically a gaining stream in the summer (June-October) and a losing stream in the winter (December-April).May and November function as transition months, where groundwater inputs and stream outputs to groundwater are approximately equal (Beltran et al., 2021;Lautz et al., 2020;Slosson et al., 2021).Meadowbrook watershed is underlain by the Cobleskill and Akron Members, Bertie Group, Camillus Shale, and the Syracuse and Rondout Formation (Belak, 1980;Muller, 1964).These bedrock units contain dolomite, gypsum, celestite, and halite, acting as geogenic sources of calcium, magnesium, sulfate, sodium, and chloride solutes in drainage waters of the basin.

| Stream water sampling and geochemical analysis
In this study, longitudinal synoptic stream sampling at 12 sites along the stream (Figure 1) was conducted approximately every 2 weeks for 1 year, starting in February 2022.Three sampling campaigns were Error in the ICP-OES result is determined as the relative standard deviation of three replicate measures, which vary from <1% to 10%.
Below detection limit (BDL) measurements were reported at one-half the detection limit level for any statistical analyses.The summary of these geochemical results is listed in Table 2.
Following the year-long sampling campaigns, two more sets of water samples from downstream (sites 1, 2), within (sites 3, 4, 6), and upstream (sites 8, 12) of the cemetery were collected on 29 June 2023 and 3 July 2023 for nitrate isotopes using the denitrifier method (Weigand et al., 2016;Yu & Elliott, 2018)  Two stream gauges measuring high-frequency (every 15 min) specific conductance, temperature, and water stage were installed along Meadowbrook in 2017 (Slosson et al., 2021) approximately 500 m upstream and downstream of the cemetery, respectively (Figure 1).
Rating curves were established by Slosson et al. (2021) to convert water stage to discharge.These rating curves were not recalibrated in this study, which might introduce some errors.However, these monitoring sites have been frequently visited and maintained to clear the stream channel.Historical data were downloaded from the literature (Beltran, 2021;Ledford & Lautz, 2023).
The comparisons in median concentrations were based on Wilcoxon Rank-Sum tests at the significance level of p = 0.05.

| Stream nitrate load modeling
Nitrate load modeling was completed at the previously mentioned upstream and downstream gauging sites using the United States Geological Survey (USGS) LOADEST model (Runkel et al., 2004) in R via  3.The lower the AIC score and the higher the rsquared value, the better the model performance.Models with bias percentage above 25% or below À25% are generally not recommended.
For both sites selected for nitrate load estimation modeling (upstream and downstream of the cemetery), Model 2 (Runkel et al., 2004) reported the lowest AIC, highest r-squared, and lowest bias percentage.Model 2 is outlined by the formula: where a 0 , a 1 and a 2 are coefficients to be optimized in the model fitting and development procedures.Optimal coefficients are shown in Table 3. Q is discharge corrected for collinearity by subtracting the center of ln(streamflow) from ln(streamflow).Models that accounted for other variables such as time, specific conductance, and temperature were developed and evaluated, but not used due to poorer performance.Model residuals were normally distributed and all coefficients were statistically significant at the confidence level of p = 0.05 for both models.All modeled nitrate concentrations for both upstream and downstream models were within the range of grab sample nitrate concentrations collected from the sites (BDL-8.29 mg/L for upstream and 0.68-6.52mg/L for downstream).The storm events sampled occurred on the highest discharge days over the sampling period for both the downstream and upstream sites (12 November and 20 January, respectively).The difference in nitrate loads between sites upstream and downstream of the cemetery was calculated to determine estimated nitrate loading from the cemetery from 1 February 2022 to 31 January 2023.

| Varied surface water-groundwater interactions under baseflow condition
Meadowbrook discharge varied seasonally in 2022-2023.In summer (mid-May through mid-October), the connected reach between the two gauges acted as a gaining stream with the downstream gauge recording higher baseflows than the upstream gauge (Figure 2a), while the connected reach was a losing stream from mid-October through mid-May.
Such discharge patterns in the connected reach are likely regulated by precipitation, evapotranspiration (ET) and surface watergroundwater interaction through the hyporheic zone.The seasonality in discharge is hypothesized to be primarily driven by the variability in  EPA, 2018EPA, , 2022;;Zambrano & Stoner, 1998).When calculating mean and median values, BDL measurements were replaced with half the detection limit.BDL represents below detection limit.precipitation and ET throughout the year.High ET in the summer (Figure 2b) likely depletes the storm pond (consistent with field observations) via evaporation and reduces the amount of water flowing into the pond, subsequently depleting the disconnected reach.Drier soils due to higher ET in the summer could also reduce flows as drier soils soak up more precipitation, resulting in less runoff to the channel.In the winter, lower ET rates lead to higher pond levels and thus higher discharge within the disconnected reach.Wetter (or frozen) soils in fall and winter would lead to more runoff to the channel.In the connected reach, the variability in the groundwater  lower median Na and Cl concentrations than in 2012 (Figure 5).
This corroborates with the findings from the literature (Kaushal et al., 2005(Kaushal et al., , 2023;;Kelly et al., 2008;Ledford et al., 2016) that while seasonal application of road salt can impact salt-related instantaneous peak analyte concentrations, legacy buildup of road salt has a lasting impact on the watershed as it does in much of the northeastern USA.
C-Q plots for Na and Cl show that as the flows increase in the winter at the upstream site, where road density-and therefore road saltingis higher, road salt will undergo greater flushing from the watershed into the stream, increasing concentration.At the downstream site, concentrations remain stable at higher flows, likely due to the lower density of roads and the tendency to store road salt in the stream banks (Ledford et al., 2016).This flushing phenomenon does not happen in the summer as there is limited road salt within the watershed to flush out at baseflow.The flat slopes occurring in Figure 6 indicate that there is a consistent supply of Na and Cl in the groundwater in the gaining season, contributing to baseflows.This is likely from road salt being used within the watershed over the course of decades (Kaushal et al., 2005;Kelly et al., 2008).including geogenic sources, fertilizers, and/or road salt (Ahmad et al., 2022;Granato, 1996).

| Nitrate in Meadowbrook watershed with a focus on St. Mary's cemetery
Nitrate median concentrations are statistically significantly higher in the winter (compared to summer) ( p < 0.01, W = 7014, n = 155) and increasing within the cemetery consistent with previous years (Lautz et al., 2020;Ledford et al., 2017).Higher nitrate in the winter might be due to lower denitrification rates and less plant uptake.Nitrate concentration increases from upstream to within/downstream of the cemetery (Figure 4), suggest that the cemetery is releasing nitrate to Meadowbrook Creek, which is also consistent with previous research (e.g., Lautz et al., 2020).Nitrate within the watershed and within the cemetery has increased since 2012 (p < 0.01, W = 49 387, n = 314).
One change in cemetery management relative to 2012 is that fertilizer is now used on cemetery grounds.This was not the case in prior years (Lautz et al., 2020).The C-Q plot of nitrate exhibits flushing patterns at the upstream site in the losing season.Less nitrate is taken up instream in the winter, allowing nitrate to accumulate in the surface water as higher flows flush nitrate out of the watershed.Nitrate in the upstream site could be sourced from fertilizers, wastewater, organic inputs such as lawn trimmings, or wet deposition.This C-Q pattern switches at the downstream site in the losing season to dilution at higher baseflows.Gaining season slopes are not statistically significantly different from zero, consistent with previous studies indicating that nitrate in urban watersheds follows negative or insignificant trends in the C-Q plot (Balerna et al., 2021;Duncan et al., 2017).
In particular, Duncan et al. (2017) showed that annual precipitation can lead to either positive or negative nitrate C-Q relationships in small (<15 km 2 ) urban watersheds, with overall multi-year trends being balanced out.Balerna et al. (2021) did not offer an explanation but showed that small urban watersheds over the course of multiple years had insignificant nitrate C-Q relationships.
In Figure 7, nitrate δ 15 N and δ 18 O values of water samples collected from downstream (sites 1, 2), within (sites 3, 4, 6), and upstream (sites 8, 12) of the cemetery are plotted along with typical isotopic ratios of five common nitrate sources in urban streams: atmosphere-derived nitrate, ammonium-based fertilizer, nitrate-based fertilizer, soil nitrate, and human/animal waste.The typical δ 15 N value of human bone is generally greater than 10‰ (O'Connell et al., 2012).
δ 15 N values of all collected Meadowbrook water samples are no more than 7.6‰, likely excluding burial decay products as a major nitrate source (Figure 7).δ 15 N and δ 18 O values of nitrate in Meadowbrook increase and decrease downstream, respectively.This suggests a mixing of at least two endmembers, with one characterized by high δ 18 O and low δ 15 N and the other with low δ 18 O and high δ 15 N. Meadowbrook Creek originates from a stormwater retention pond, which allows ample time for surface water to equilibrate with the atmosphere.In addition, δ 15 N (and δ 18 O) values of all samples are within or higher (and within or lower) than that of nitrate fertilizer.Thus, atmosphere-derived nitrate and/or nitrate fertilizer are deemed likely to be the high δ 18 O/low δ 15 N endmember.As water flows downstream, nitrate derived from other sources, for example, soil nitrate, more fertilizers, and human/animal waste, are released into Meadowbrook, which brings down nitrate δ 18 O and elevates nitrate δ 15 N values.At Meadowbrook, baseflow water samples at two stream sites (i.e., at the end of disconnected reach and the end of connected  while most samples that are further upstream from the two sites vis-  2023) reported that a cemetery with vaults showed no burial decay product impact on nearby groundwater quality, including nitrate concentrations.In addition, fluoride (F), which accumulates in human teeth and bones, can be released into the environment as part of the burial decay products (Dent et al., 2004).F concentration in Meadowbrook Creek either declines in 2012 or stays constant in 2022 as it flows through the cemetery in the gaining season, arguing against the presence of major impacts of burial decay products on Meadowbrook water quality.
Third, the isotopic composition of Meadowbrook nitrate shows that the nitrate is more likely to be dominantly sourced from atmospheric nitrate, fertilizers, and wastewater.
Nitrate is unlikely to be sourced from burials prior to the 1970s based on the estimated residence time of groundwater and nitrogen and the isotopic signature of our data.Van Meter et al. (2017) showed that nitrogen older than 50 years accounts for approximately 33% and 5% of the nitrogen at the basin outlet for the Mississippi (3 000 000 km 2 ) and Susquehanna (70 000 km 2 ) rivers.Smaller rivers have faster biogeochemical cycling of nitrate (Basu et al., 2011), suggesting that nitrate residence time within small watersheds like Meadowbrook is shorter than 50 years.This suggests that very little, if any, nitrogen in Meadowbrook is the result of pre-1970s burials.It is also important to note that soil organic nitrogen released from pre-1970s might have accumulated in a 'protected pool' of soil organic nitrogen that has a residence time on the order of 1000 years (Van Meter et al., 2017).However, St. Mary's has only been operational since the early 1900s, likely eliminating this 'protected pool' as a source of nitrate.In addition, our nitrogen and oxygen isotope data points to limited denitrification signals.If long water (and thus nitrate) travel time was expected, we would expect to see coupled enrichment of nitrogen and oxygen isotopes due to subsurface denitrification (Yu et al., 2023), that is, N and O isotopes will be enriched following a 1:1 to 1:0.5 line, but that is not the case here (Figure 7).Nitrate loading is driven primarily by patterns of surfacegroundwater interaction, as loading corresponds to when the stream is a losing and gaining stream within the cemetery reach and is shown in Figure 8, where we predominantly see the stream is losing both discharge and nitrate load or gaining both discharge and nitrate load on a given day.However, an exception to this general pattern is that Meadowbrook might experience a net gain of nitrate loads while also having a net loss of discharge throughout the cemetery, particularly in fall and spring (Figure 8).The fall and spring phenomena of a net loss in discharge but a net gain in nitrate loads are indicative of high nitrate concentrations in the groundwater within the cemetery.This reach likely experiences dynamic hyporheic exchange with a small net loss of stream water to groundwater during this time.However, the highconcentration cemetery groundwater carries enough nitrate relative August 2020.We thank the editor for handling this paper as well as two reviewers for their insightful and thorough reviews.

CONFLICT OF INTEREST STATEMENT
The authors declare no competing financial interest.
St. Mary's Catholic Cemetery has an area of 0.85 km 2 , through which over 500 m of Meadowbrook Creek flows.Slopes within the cemetery angle down towards Meadowbrook and range between 5% and 15%.Most burials within the last 10 years occur in sections outlined as 'Active Burials' (blue rectangle in Figure 1).The cemetery has been operational since the early 1900s and currently receives approximately 425 burials per year.Burials within the last 50 years at St. Mary's use concrete or steel vaults.Fertilizers are applied primarily at active burial sections 'a couple times per year', and elsewhere in the cemetery where needed.No irrigation is used and lawn clippings are not collected after mowing (T.Smith, personal communication, 9 May 2023).

T A B L E 1
Site numbers, upstream distance from where Meadowbrook Creek flows into Butternut Creek, whether the stream channel is strongly connected to groundwater, and position of the site relative to St. Mary's Cemetery.package 'rloadest'.The natural logarithm of both discharge and concentration measurements were calculated and combined to develop a model that estimates nitrate loads based on high-resolution discharge data from 1 February 2022 to 31 January 2023 to encapsulate the entirety of the field sampling in this study.This model estimates loads using the Adjusted Maximum Likelihood Estimation (AMLE) and can include a variety of variables including seasonality, temperature, and specific conductance, but primarily uses discharge.The load models chosen for these analyses were selected based on three factors: the Akaike Information Criterion (AIC) score, r-squared value, and bias percentage.Each of these values is shown in Table

4. 2 |
Sources of Na, Ca, Mg, K, F, Cl, and sulfate in Meadowbrook watershed where various human-built infrastructure overlaps Number of samples (total and above detection limit) and basic statistics of each geochemical analyte are listed in Table 2 along with corresponding water quality standards.Water quality standards were based primarily on aquatic life standards.If no aquatic life standards were available, drinking water standards were listed.Nitrate concentrations, sources, and loads in Meadowbrook are discussed in depth in Section 4.3.A Spearman correlation analysis was conducted to examine if any pairs of analytes share common sources (Figure3).Spearman correlation analysis was chosen over Pearson correlation because Spearman analysis can handle censored data (e.g., BDL values).Dissolved sulfate in Meadowbrook is mainly derived from gypsum in the underlying bedrock units.K, Ca, and Mg concentrations all correlate with sulfate concentration suggesting a common geogenic source for these analytes.Unlike Mg and K, Ca and sulfate both exhibit a strong seasonality, which might reflect seasonally varied hydrological controls on the release and transport of different analytes into Meadowbrook Creek (Figure4).Wilcoxon rank sum tests show that annual median values of Ca, sulfate, and K are statistically significantly higher in 2012 than in 2022 for the entire stream ( p < 0.01, W = 24 044, 29 281 and 20 883, respectively, and n = 314) (Figure5).This might be due to the dilution effect as evidenced by the lower precipitation in 2012 (90 cm) than in 2022 (113 cm)(Menne, Durre, Korzeniewski, et al., 2012; Menne, Durre, Vose, et al., 2012).Concentrationdischarge (C-Q) relationships were assessed at baseflow for the eight major water ions collected from 2017 to 2023(Beltran et al., 2021 and this study) to further assess the solute source as well as the hydrological controls on solute release into Meadowbrook.Baseflow for data collected during 2022-2023 is defined as samples collected outside of the previously identified storm samples.Baseflow for data collected in 2017-2019 was determined as the lower 90th percentile as individual storm events were not identified in the data.This Daily averaged discharge values upstream and downstream of the cemetery reach (a) and daily actual evapotranspiration (ET) and eight-day moving average of precipitation (b).Actual ET data are sourced from eight-day satellite-based remote sensing data(Running et al., 2017).Precipitation data(Menne, Durre, Korzeniewski, et al., 2012; Menne, Durre, Vose, et al., 2012)  is smoothed using an 8-day moving average to match.resulted in similar values of discharge to the 2022-2023 baseflows.Such relationships were evaluated based on site (upstream vs. downstream of the cemetery in the connected reach) and season (whether the connected reach was gaining or losing).C-Q relationships are shown in Figure6.Ca, Mg, and sulfate all show dilution patterns-decreased concentration at higher discharge.This again supports a common geogenic source for these three analytes.Unlike these three analytes, K shows dilution at both sites during the gaining season, and no change at either site during the losing season, likely reflecting different flow paths, contributing source material and/or hyporheic zone dynamics such as plant uptake, where K and other analytes are hypothesized to differ in riparian zones in part due to vegetation(Ledesma et al., 2013).Na and Cl concentrations are correlated showing statistically significantly higher values in winter (losing season) than in summer (gaining season) ( p < 0.01, W = 1873 and 2388, respectively, and n = 155) (Figure4).This pattern suggests a common source, i.e., road salt application that is prevailing in the Meadowbrook watershed in wintertime, which is consistent with whatLedford and Lautz (2015) observed in Meadowbrook in 2012.The connected reach likely releases stored road salt throughout the gaining season (i.e., summer) that has accumulated during the losing season.This phenomenon was proposed byLedford and Lautz (2015) for data in 2012-2013 and is still observed in 2022-2023.Cl routinely exhibits concentrations higher than aquatic life guidelines of 260 mg/L, warranting better urban water management to safeguard aquatic species from elevated chloride levels in urban water bodies.Na and Cl concentrations display no statistically significant difference in median between 2012 and 2022 (p = 0.197 and 0.233, W = 40 506 and 40 902, respectively, and n = 314).Predominantly sourced in the watershed as road salt, Na and Cl had higher instantaneous peaks in 2012 potentially due to the higher snowfall amounts than in 2022 (272 cm compared to 165 cm,Menne, Durre, Korzeniewski, et al., 2012; Menne, Durre, Vose,   et al., 2012)  likely resulting in more road salt being used.However, these higher instantaneous peaks might also be due to the low temporal resolution of sampling resulting in 2022-2023 samples that missed the highest peaks in 2022-2023.It is worth noting that the much lower snowfall in 2022 did not yield statistically significantly different Correlation matrix and scatterplots for all analytes.Spearman correlation coefficient is shown in each box."***" indicates a significance level at 0.001, "**" at 0.01, "*" at 0.05, and "." at 0.1.
Fluoride (F) shows a strong seasonality, reporting statistically significantly higher values in the gaining season (p < 0.01, W = 13 526, n = 155).We do not observe significant spatial variation within and downstream of the cemetery compared to upstream of the cemetery ( p = 0.276, W = 36 048, n = 165, 71, respectively) (Figure4).Unlike this study,Lautz et al. (2020) showed that F concentrations significantly decreased within the cemetery in 2012 and hypothesized that longitudinal fluoride variation along Meadowbrook is due to a lack of leaking sewer and water infrastructure within the cemetery.In both 2012 and 2022, F concentration in Meadowbrook Creek failed to increase through the cemetery, likely excluding the cemetery (including burial decay products) as a primary source of F in Meadowbrook (see Section 4.3 for details).Fluoride data exhibits C-Q patterns with a slope not significantly different from zero, except for upstream during the losing season where it undergoes slight dilution at higher baseflows.Such slopes demonstrate that fluoride is not diluted at higher flows.This indicates a source that varies with the flow, potentially Boxplots showing the concentration distribution of Ca, Mg, sulfate, K, nitrate, F, Cl, and Na of samples collected from eight sites within the connected reach in Meadowbrook, with green and orange boxes representing losing and gaining seasons, respectively.Vertical black lines indicate the upstream and downstream boundaries of the cemetery.Horizontal black lines denote aquatic life water quality standards.
Boxplots showing the concentration distribution of (a) nitrate, (b) fluoride, (c) sulfate, and (d) chloride within Meadowbrook, with hollow (left) and filled (right) black boxes representing the 2012 and 2022 sample years, respectively.Vertical black lines indicate the upstream and downstream boundaries of the cemetery, and the dashed green lines indicate the boundary between the upstream disconnected reach and the downstream connected reach.Magnesium, calcium, and potassium follow the same patterns as sulfate.Sodium follows the same patterns as chloride.reach) were analyzed for nitrogen and oxygen isotopes of nitrate in 2012-2013 (Ledford et al., 2017).Based on the isotope data, Ledford et al. (2017) concluded that baseflow nitrate in Meadowbrook Creek might be mainly sourced from the mixing of wastewater and groundwater.They also acknowledged that such baseflow samples might not reflect the contribution from sources related to overland flow and storm events (e.g., fertilizer).Many water samples collected in 2023 present similar nitrate isotope data as those in 2012-2013, U R E Concentration-discharge plots for sulfate, Ca, Mg, K, Cl, Na, nitrate, and F at the baseflow condition.Regression lines with 95% confidence intervals are shown for four scenarios: upstream or downstream of the cemetery reach during the gaining or losing season.
U R E 7 δ 18 O and δ 15 N values of samples taken on 29 June 2023 and 2 July 2023 in Meadowbrook.Coloured rectangles represent typical isotope signatures from various sources(Kendall, 1998).The black arrow represents the general sample location, pointing in the direction from upstream to downstream.The general pattern is that upstream samples are higher in δ 18 O and lower in δ 15 N, which decreases and increases, respectively, as moving downstream.The vertical black dash line represents typical human bone δ 15 N values(O'Connell et al., 2012).
to the stream water to offset the nitrate losses that are associated with a net loss of discharge and result in increased nitrate loads during these time periods.Jimenez-Fernandez et al. (2022) describes a headwater stream undergoing various hyporheic exchange patterns including both net gaining from groundwater input and net losing to groundwater input throughout the 500-meter reach studied.Meadowbrook likely undergoes a similar pattern throughout the cemetery and future research should identify more specific areas of the gaining and losing stream.Another possibility is that varying residence times within the hyporheic zone may lead to spatial and temporal changes in nitrification and denitrification patterns (e.g.,Zarnetske et al., 2011).Although the cemetery reach can be a seasonal source of nitrate, it acts as an annual nitrate sink.Practices such as reducing fertilizer usage, cleaning up lawn clippings, and phytoremediation via increasing near-riparian vegetation may help turn the reach into a year-round sink.5 | CONCLUSIONSMeadowbrook Creek in Syracuse, New York (USA) provides a unique opportunity to examine anthropogenic impact and hydrological control on solute release and contributions to an urban stream in a setting with various types of human-built infrastructure (e.g., cemetery, residential, paved roads).Our bi-weekly longitudinal stream sampling and continuous discharge measurements focused in and around St. Mary's cemetery over a year demonstrated highly variable spatiotemporal hydrogeochemical patterns representing contributions from a wide variety of solute sources, primarily including the cemetery and road salting.In particular, geogenic solutes such as calcium and potassium undergo various spatial and temporal trends and may be driven by the variability in precipitation and seasonal surface water-groundwater interaction.Decadal trends from 2012 to 2022 indicate that sodium and chloride sourced from maintained use of road salt continue to be a detriment to aquatic life, although the cemetery reach has been shown it can help attenuate the problem by benefiting from the ecosystem functions provided by its riparian zones and connection to groundwater relative to more urbanized channels.Previous studies suggested that up to 25%-50% of the observed nitrate increase in Meadowbrook Creek along St. Mary's cemetery could plausibly derive from burial decay products.However, we suggest that burial decay products are unlikely to be a significant source for nitrate (and other analytes) in Meadowbrook due to the sealed burial vaults and other prolific nitrate sources within the watershed such as fertilizer, decaying plant matter, and wastewater, along with nitrogen isotopic analysis.Nitrate loading calculation indicates that the cemetery acts as an annual net sink for nitrate within the watershed, providing an ecosystem service for the urban stream.ACKNOWLEDGEMENTSThis material is based upon work supported by the Syracuse University Small Equipment Grant SEG-2686-2021 to T.W. J.K. is supported by The Syracuse Office of Undergraduate Research & Creative Engagement (The SOURCE) fund.We thank Dr. Chris Junium (Syracuse University) for access and help with field equipment, Beibei E (Syracuse University) for assistance with field work, St. Mary's Catholic Cemetery and Dewitt residents for stream access, and Dr. Laura Lautz for establishing and maintaining Meadowbrook sites prior to Analytes examined and their method detection limits, number of samples ADL, minimum, maximum, median, and mean values, water quality guidelines, and water quality guideline type.Guideline types correspond with EPA Chronic Aquatic Life Standards (AL), New York State Aquatic Life (NYS AL), New York State Drinking Water (NYS DW) and EPA Drinking Water Secondary Guidelines (DW Secondary), (US T A B L E 2 ited in 2012-2013 show much higher δ 18 O/lower δ 15 N values approaching the nitrate fertilizer zone.In particular, in 2012-2013, δ 15 N values ranged from 1.9‰-12.6‰,slightly higher than the range of 0‰-7.6‰found in 2023.δ 18 O values in 2012-2013 ranged from À9.0‰-8.0‰,much lower than the 1.7‰-21.6‰found in 2023. Daily nitrate loads estimated with LOADEST models for upstream and downstream gauges of the cemetery.Coloured dots are used to indicate if the nitrate load (denoted by 'N') and discharge (denoted by 'Q') are gaining ('+') or losing ('À') between the two sites.