Introduction

In Curtice and Williston, Ohio, persistent elevated concentrations of the bacteria Escherichia coli (E. coli) in Cedar and Crane Creeks, respectively, have led to the Ohio Environmental Protection Agency (Ohio EPA) listing these two water bodies as impaired for primary contact recreation (Tetra Tech, Inc., 2012). Additionally, these two creeks are frequently the subject of public complaints due to strong odors of sewage in certain locations (Kris Barnswell, Ohio EPA, written commun., March 2024). In 2016, the Ohio EPA conducted a public health nuisance investigation in accordance with rule 3745–1–04(F) of the Ohio Administrative Code (Ohio EPA, 2024b) on Cedar Creek and Crane Creek in Lucas and Ottawa Counties, Ohio. Nuisance conditions were verified on Cedar Creek downstream from Wildacre Road in Curtice, Ohio (fig. 1; Kris Barnswell, Ohio EPA, written commun., March 2024).

In 2017, the Ohio EPA conducted a water-quality survey that included Cedar and Crane Creeks (Ohio EPA, 2023). This survey collected chemistry, biology, sediment, and fecal indicator bacteria data in the watersheds. Evaluations of these streams included sampling at two locations upstream from where the 2016 nuisance investigation occurred. Sampling downstream from the nuisance investigation area was limited to biological sampling only because of the lacustuary zone (mixing of stream and Lake Erie water). Results from this survey indicated that the nuisance conditions could be caused by the presence of sewage in Cedar and Crane Creeks (Ohio EPA, 2023). There is also documented fecal bacteria concentrations (E. coli) exceeding regulatory standards under both dry and wet weather conditions (Ohio EPA, 2023). The sewage odors and E. coli survey results indicated human sources, potentially from unsewered areas, as one potential cause of these conditions. However, as suggested in the Ohio EPA water-quality survey, other sources may be contributing to the nuisance conditions and elevated fecal bacteria, such as wastewater treatment plant effluent violations, livestock waste, wildlife waste, combined sewer overflows, urban runoff, and biosolid application (Ohio EPA, 2023). An examination of the sources of fecal contamination within the Cedar and Crane Creek watersheds could help water-resource managers identify primary fecal sources, and water-resource managers may use the information to improve focus areas for remediation efforts.

Exposure to microbiological pathogens from human or animal feces poses a major human health risk. Water-quality criteria and standards to protect the public from exposure to microbiological contaminants are based on fecal indicator bacteria such as E. coli, but these bacteria are common to all warm-blooded animals, and thus, do not provide information about the source of the contamination. Microbial source tracking (MST) refers to a group of methods and strategies that can be used to help characterize the dominant sources of fecal contamination in a water body. (Santo Domingo and others, 2007). The detection of fecal source-associated genetic markers in the aquatic environment from MST can help reveal the timing and area of fecal contamination in a watershed.

To better understand the sources of fecal contamination in Cedar and Crane Creeks, the U.S. Geological Survey (USGS), in cooperation with the Ohio EPA, monitored 12 sites on Cedar and Crane Creeks during the recreational season of 2021 (May through September). A total of 118 water samples were evaluated for the magnitude of E. coli and source-associated MST markers. All samples were evaluated for MST markers associated with human and canine sources and a subset of samples were evaluated for MST markers associated with swine, ruminant, cattle, horse, waterfowl, and poultry sources. Samples were also collected and analyzed for nutrients including ammonia, total Kjeldahl nitrogen, nitrate plus nitrite, nitrite, total phosphorus, and phosphate. Nutrient and E. coli samples were collected and analyzed by Ohio EPA. At the time of publication, nutrient and E. coli data were not publicly available from the Ohio EPA.

Study Area and Site Descriptions

Cedar and Crane Creeks are tributaries to Lake Erie that respectively flow through the towns of Curtice in Lucas County and just north of Williston in Ottawa County, Ohio (fig. 1). Drainage areas for Cedar and Crane Creeks are about 58 and 56 square miles, respectively. (Tetra Tech, Inc., 2012). These waterways have a beneficial use designation of warmwater habitat for aquatic life (Tetra Tech, Inc., 2012). The creeks are used for agricultural and industrial water supplies and are assigned as primary contact recreation by the Ohio EPA (waters that, during the recreation season, are suitable for one or more full-body contact recreation activities such as, but not limited to, wading, swimming, boating, water skiing, canoeing, kayaking, and scuba diving; Tetra Tech, Inc., 2012). Based on 2006 data for Cedar and Crane Creeks, along with several other surrounding Lake Erie tributaries, land use in these watersheds is dominated by cultivated cropland (64 percent), low-intensity development (10 percent), and emergent herbaceous wetlands (8 percent; Tetra Tech, Inc., 2012).

Sampling was done at 12 sites (table 1) as part of this study: 6 Cedar Creek sites and 6 Crane Creek sites. The study area (fig. 1) includes Cedar Creek from river mile 9.59 (Billman Road) to river mile 5.56 (Decant Road) and Crane Creek from river mile 10.98 (Reiman Road) to river mile 7.45 (Opfer-Lentz Road). A river mile is the curvilinear distance, in miles, measured upstream from the mouth of the stream along the path of the stream channel.

Methods

This section describes the general procedures for environmental and quality-control sample collection and laboratory analytical methods. Data availability and analysis also are described.

Results

Microbial source tracking and E. coli results from samples collected along Cedar and Crane Creeks are presented here in the context of site and sampling conditions (dry or wet weather). Based on the distribution analysis of the rainfall data (National Oceanic and Atmospheric Administration, 2025), classification of sampling events resulted in 4 of the 10 sampling dates being classified as wet weather and 6 of the 10 sampling dates being classified as dry weather (table 2).

In three sampling events, samples were tested for all eight investigated sources of fecal contamination: swine, ruminants, cattle, horses, waterfowl, poultry, humans, and canines (table 2). Among these three sampling events, one of them was during wet-weather conditions. Fecal contamination from swine, ruminants, cattle, horses, waterfowl, and poultry was not detected above the assay LoD in any of the samples from these three dates (U.S. Geological Survey, 2025). In these three sampling events, fecal contamination from humans and canines was detected, and the two MST markers associated with those sources (HF183/BacR287 and BacCan) were analyzed in all samples collected in this study.

All data for the human and canine MST markers from each site are presented in boxplots (fig. 2); and the same data are categorized into dry- and wet-weather conditions for each stream (fig. 3).

The human and canine MST markers were prevalent in samples collected for this study (figs. 2 and 3; summary data presented in table 4). The human marker was detected in 114 of 118 samples and percent detection among sites ranged from 90 to 100 percent. The canine marker was detected in 112 of 118 samples and percent detection among sites ranged from 62 to 100 percent. Among the 114 detections of the human marker, 32 were qualified as estimates based on concentrations being below the LoQ (28 percent); whereas among the 112 detections of the canine marker, 58 were qualified as estimates based on concentrations being below the LoQ (52 percent).

Elevated E. coli concentrations were persistent in samples collected for this study (figs. 2 and 3; table 4). For all samples collected, E. coli concentrations ranged from 75 to 20,000 MPN/100 mL. Median concentrations of E. coli at each site ranged from 880 to 2,000 MPN/100 mL. Results of Kendall rank correlation coefficient analysis indicate that within each creek, log concentrations of E. coli positively correlate with log concentrations of the human marker (Cedar Creek Kendall rank correlation is 0.48, and significant with a p value of less than 0.0001; Crane Creek Kendall rank correlation is 0.3, and significant with a p value of 0.0009).

Results of the Kruskal-Wallis test indicate that the distribution of concentrations of the human marker, the canine marker, and E. coli are not statistically significantly different among sites in Cedar Creek. However, results from the same test in Crane Creek indicate that, for the human marker, the distribution of concentrations of the human marker are significantly different. Results of the “Wilcoxon Each Pair” test indicate that at the Martin Williston Road ditch effluent site (Crane Creek site 4), human marker concentrations are significantly higher than at other sites (p value 0.0002). The median human marker concentration at Crane Creek site 4 was 120,000 copies/100 mL; median concentrations at the other five sites on Crane Creek ranged from 2,300 to 16,000 copies/100 mL. Additionally, results from the “Wilcoxon Each Pair” test indicate that the distribution of human marker concentrations at Crane Creek at Curtice E W Rd. (Crane Creek site 5) are significantly higher than at Crane Creek at Genoa-Clay Rd. (Crane Creek site 2). Results of the Kruskal-Wallis test for differences in the distribution of concentrations of the canine marker (p value 0.25) and E. coli (p value 0.95) are not significantly different for sites in Crane Creek.

Results of the Peto-Prentice test (Helsel, 2005) indicate that human marker, canine marker, and E. coli concentrations are significantly greater in both Cedar and Crane Creeks when collected during wet-weather than when collected during dry-weather conditions (fig. 3). Median concentrations at each site during wet-weather and during dry-weather conditions are listed in table 4.

Discussion

Following the analysis of results from the three sampling events (May 4, June 9, and June 30, 2021) that evaluated eight MST markers from various animal sources, it was found that only human and canine markers were detected above the limit of detection. Consequently, the remainder of the study concentrated solely on these two markers (refer to table 2 for sampling dates and evaluated MST sources). The absence of MST markers associated with swine, ruminants, cattle, horses, waterfowl, and poultry in all of the 36 samples collected during the three sampling events does not prove that fecal contamination from these animals is absent, but it does indicate that these sources are unlikely to be the primary contributors of the elevated E. coli concentrations and nuisance conditions observed in these watersheds. Therefore, efforts to reduce fecal contamination from swine, ruminants, cattle, horses, waterfowl, and poultry may not lead to a decrease of E. coli concentrations or nuisance conditions in the Cedar and Crane Creek watersheds.

Human fecal contamination was prevalent at all sites sampled in this study and was greater during wet-weather conditions than when compared to dry-weather conditions. This finding, along with persistently elevated E. coli concentrations, corroborated prior evidence of nuisance conditions based on observations of sewage odors and high E. coli concentrations (Kris Barnswell, Ohio EPA, written commun., March 2024; Tetra Tech, Inc., 2012; and Ohio EPA, 2023). In Cedar Creek, although human fecal contamination was detected at all sites, differences in concentrations among the sites were not significant. In contrast, at the Martin Williston Road ditch effluent site on Crane Creek (site 4), the concentration of human fecal contamination was statistically significantly higher than at other Crane Creek sites. This finding indicates that while contamination was present at all sites, only the Martin Williston Road ditch drain pipe site on Crane Creek exhibited a substantial difference in contamination levels compared to the other sites. The high concentration at the Martin Williston Road ditch indicates that this site is likely a source of human fecal contamination to Crane Creek at its outfall. Aside from the Martin Williston Road ditch, the results at all other study sites indicate that human fecal contamination was nearly uniform for the reaches of these two streams during this evaluation. Elevated concentrations of the human MST marker at the most upstream sites sampled in Cedar and Crane Creeks indicate that human fecal sources may exist upstream from the study sites. Previous monitoring and these study results show that there are sources of human feces contaminating Cedar Creek and Crane Creek near Curtice.

Concentrations of the human MST marker from the Martin Williston Road ditch samples were at magnitudes that are typically seen in raw human waste sample types such as sewage or septic wastewater. Ahmed and others (2016) summarized current MST research on human fecal contamination in environmental waters that included concentration data for the HF183 marker (the predecessor MST marker from which HF183/BacR287 is based) in various types of raw human waste. The mean concentrations of the human marker in those sample types, obtained from 18 studies reporting on raw sewage and 7 on septic wastewater, were 17,000,000 and 8,000,000 copies/100 mL, respectively.

The canine and human markers were commonly detected together in this study. The canine marker was detected in 95 percent of the samples where the human marker was detected. Co-occurrence of these two markers is not unexpected, given the cohabitation of dogs and humans. Additionally, BacCan targets the same region of highly conserved gene found at high concentrations in a number of hosts (Green and others, 2014b). Previous studies have reported BacCan occurrence in human sources (known as cross-reactivity), including in human septage (Boehm and others, 2013; Schriewer and others, 2013), stool (Kildare and others, 2007; Tambalo and others, 2012a; Schriewer and others, 2013; Nshimyimana and others, 2017), and sewage (Kildare and others, 2007; Schriewer and others, 2013; Nshimyimana and others, 2017). The cross-reactivity of BacCan in human sources does not rule out canine feces as contributing to overall fecal contamination in the watershed. However, the occurrence of BacCan in human sources calls into question the predominance of canine fecal contamination, especially given the co-occurrence with human marker detections. Additionally, the frequency of low-level detections of the canine marker (52 percent below the limit of quantification) implies that canine sources of fecal contamination may be less prevalent than human sources.

E. coli is used as an indicator to determine attainment or nonattainment of recreational uses as codified in rule 3745–1–07 of the Ohio Administrative Code (Ohio EPA, 2024a). Water quality for primary contact recreation must meet a 90-day geometric mean (126 MPN/100 mL) and a statistical threshold not to be exceeded by more than 10 percent of samples taken during any 90-day period (410 MPN/100 mL). Among the 118 samples collected, 107 samples (91 percent) had E. coli concentrations exceeding 410 MPN/100 mL. Median E. coli concentrations for each site ranged from 2.1 to 4.9 times the statistical threshold of 410 MPN/100 mL.

Summary

Cedar and Crane Creeks, near the towns of Curtice in Lucas County and Williston in Ottawa County), Ohio, have been the topic of community complaints due to sewage-like odor problems and have also been listed by the Ohio Environmental Protection Agency (Ohio EPA) as impaired for recreational use because of elevated Escherichia coli (E. coli) concentrations. Because these creeks were impaired for beneficial use and presented nuisance conditions, Cedar and Crane Creeks were evaluated for potential sources of fecal contamination in the watersheds. In cooperation with the Ohio EPA, the U.S. Geological Survey conducted a study from May to September 2021 using microbial source tracking (MST) methods to characterize host fecal contamination sources. Samples were collected during 10 sampling events from 12 sites: 6 sites along Cedar Creek and 6 sites along Crane Creek. All samples were analyzed for E. coli concentrations and MST markers associated with human and canine fecal sources. A subset of samples was analyzed for swine, ruminant, cattle, horse, waterfowl, and poultry fecal sources.

Results from this study showed persistently elevated concentrations of E. coli and persistent detection of both the human- and canine-associated MST markers (HF183/BacR287 and BacCan, respectively), regardless of site or sampling conditions. There were no detections of MST markers associated with the other sources evaluated (swine, ruminant, cattle, horse, waterfowl, or poultry). Among all samples, 91 percent had E. coli concentrations exceeding the statistical threshold of 410 counts per 100 milliliters, indicating that all sites evaluated were impaired by fecal contamination and were unsuitable for recreational uses. The human marker was detected in 114 of 118 samples, and the canine marker was detected in 112 of 118 samples. Concentrations of E. coli, the human marker, and the canine marker were statistically significantly higher during wet-weather in comparison to dry-weather conditions.

This study provides support that human fecal contamination is likely the dominant source of elevated E. coli concentrations in both the Cedar Creek and Crane Creek. In addition to the prevalence of the human marker both spatially and temporally in the study area, concentrations in the Martin Williston Road ditch were at magnitudes that are typically seen in raw human waste sample types such as sewage or septic wastewater. The E. coli concentrations are significantly correlated with human marker concentrations, further supporting the potential predominance of human fecal contamination among potential sources. Although the canine marker was also persistently detected, co-occurrence issues with human sources confounds the characterization of canine waste as a dominant fecal source.

There were no statistically significant differences among MST marker or E. coli concentrations at sites in either Cedar or Crane Creek with one exception. The Martin Williston Road ditch site along Crane Creek had significantly higher human marker concentrations than the other Crane Creek sites and could be a source of human fecal contamination. The near uniformity of human fecal contamination at the other sites evaluated in this study indicates that sources may exist upstream from evaluated sites or at other locations within these reaches of Cedar and Crane Creeks. With additional data collected from more sites in the watersheds, it may be possible to more clearly identify the entry points of human fecal contamination. This study, taken in concert with previous monitoring, suggests that while further data collection might determine the location of additional human fecal contamination sources upstream, there are likely existing human fecal contamination sources to Cedar and Crane Creeks near Curtice, Ohio.