Evaluation of Escherichia coli (E. coli)Concentrations inMill Creek, Newton Co., Arkansas2016-2017Office of Water QualityPlanning BranchWQ18-5-01

IntroductionPoint and non-point source water quality threats are well established in Mill Creek (Newton Co.),which is a tributary of the Buffalo National River. Maner and Mott (1991) provided the firstcomprehensive water quality study of Mill Creek watershed and found factors contributing towater quality degradation in the Buffalo River included elevated fecal coliform concentrationsand nutrient concentrations. Secondary and tertiary concerns of note which may cause degradedconditions in the Buffalo River were bank erosion, sedimentation. Lastly, Maner and Mott(1991) stated water quality impacts from Marble Falls wastewater treatment facility SewerImprovement District (SID) were insignificant due to minimal flow and low concentrations ofpollutants.Maner and Mott’s study hypothesized that elevated bacteria and nutrient concentrations werepossibly due to groundwater infiltration. Mott et al. (1999; 2000) tested an interbasin watertransfer hypothesis using dye studies to delineate groundwater recharge areas confirmed thatwaters geographically originating in the Crooked Creek basin enter Mill Creek basin throughconnection of karst features and springs. Mott et al. (1999) proposed that elevated bacteria andnitrogen concentrations in Mill Creek originate in the Crooked Creek watershed due to land usepractices between the two watersheds.Permitted FacilitiesConstructed in the 1960s to serve the Dogpatch U.S.A. theme park, Marble Falls SewerImprovement District (SID) (AR0034088) is the only NPDES permitted facility discharging toMill Creek. Currently, the Marble Falls SID serves approximately 30 accounts that include aprivate residences, post office, recreation center, restaurant, hotel, and conference center, whichreceive moderate tourism visitation during summer months.In 1992, Arkansas Department of Environmental Quality (ADEQ) entered into a ConsentAdministrative Order (CAO 92-007) requiring Marble Falls SID to conduct a 12-month study ofthe polishing pond and ensure that the pond met the Ten States Standards (Great Lakes 2012). In2001, ADEQ issued a Notice of Violation (NOV) (LIS No. 01-217) because the 1992 CAOrequirements were not completed. The NOV reinforced requirements for Marble Falls SID toconduct a 12-month study of the polishing pond and ensuring that the pond met the Ten StatesStandards with the additional requirement to submit monthly progress reports. ADEQ enteredinto a Consent Administrative Order (CAO 03-049) in 2003 with Marble Falls SID settingcertain monitoring requirements for plant effluent at the discharge point prior to entering thepond. Additionally, this CAO required a report by a professional engineer demonstrating that thepond meets the Ten States Standards.In 2009, Marble Falls SID suffered multiple equipment failures and coupled with an aginginfrastructure led to raw influent infiltrating groundwater and subsequently surface waters fornearly a year. Usrey (2010) reported an increase E. coli arithmetic mean from 53 colony forming

units (cfu)/100 mL upstream of Marble Falls SID to 150 cfu/100mL at the lowest Mill Creekmonitoring point in 2010. During 2010, Mill Creek at Pruitt (BUFT04) had 13 of 74 (18%)exceed 298cfu/100 mL. Late 2010 brought a consent decree and settlement between ADEQ andMarble Falls SID (Case No. CV 2010-10-1) and an agreement that Marble Falls SID wouldimplement phased infrastructure and facility remediation. However, contempt of the 2010consent decree was filed in 2015, and Marble Falls SID is still considered out of compliance withtheir NPDES permit and required to submit quarterly monitoring reports to ADEQ Office ofWater Quality Enforcement. As of 2017, Marble Falls SID has repaired damaged transfer lines,closed inoperable transfer lines, and replaced damaged equipment.Bacteria Water Quality StandardsE. coli is common indicator of fecal contamination of freshwater because it constitutes more than90% of bacteria found in warm blooded animal excrement (US EPA 1986). Indicators such asfecal coliform are used less frequently as indicators of contamination because poor correlationwith swimming related illness, whereas E. coli presence has a much stronger relationship (USEPA 1986). Many states, including Arkansas, utilized fecal coliform as the primary analyte untilthe early 2000s. It was not until 2004 that Arkansas switched to utilizing E. coli as the primaryindicator for monitoring recreational contact.Arkansas’s water quality standards are established by the Arkansas Pollution Control andEcology Commission Regulation No. 2 (APC&EC 2017). Regulation No. 2.507 (Bacteria)defines primary contact recreation season as May 1 to September 30 and establishes bacteriacriteria for primary and secondary contact seasons for Extraordinary Resource Waters (ERWs)and Ecological Sensitive Water (ESWs), lakes/reservoirs, and all other waters. Mill Creek isclassified as “all other water” and thus has an E. coli individual sample criteria of 410colonies/100mL primary contact season. Data quality requirements necessitate that eight (8)samples be collected from May 1 through September 30 and cannot exceed 410 colonies/100mLmore than 25% of the time (i.e. more than 2 of 8 samples). For all streams with watersheds lessthan ( ) 10mi2, primary contact shall not be designated for primary contact unless siteverification indicates that such use is attainable. Fecal coliform individual sample and geometricmean criteria of 400 and 200 cfu/100 mL, respectively, apply to ERWs, ESWs, lakes/reservoirs,and all other waters.In an effort to protect public health, NPS applies ERW criteria to waters within the NPSboundary even though not all waters within the NPS boundary are designated as an ERW. Forthe NPS monitoring program, if E. coli concentrations exceed an individual-sample value of 298colonies/100mL, then the NPS will increase sampling efforts to five (5) samples per month.During this high frequency sampling, if a 30-day geometric mean exceeds 126 colonies/100mL,then NPS staff will erect signage advising public caution, issue health advisories on theirwebsite, and contact Arkansas Department of Health and ADEQ (NPS Pers. Communication,Figure 1).

Impairment ConcernsIn 2015, the NPS submitted a letter and supporting data to ADEQ requesting Mill Creek beplaced on the 2016 impaired waterbodies list pursuant to Section 303(d) of the Clean Water Actfor violation of fecal coliform bacteria water quality standards (Cheri 2015). In this letter, NPSalso requested assistance from ADEQ in determining the source(s) of the fecal coliformcontamination so it could be reduced or eliminated. The NPS presented data for Mill Creeksampling station BUFT04 during summer 2015. Five samples were collected within a 30-dayperiod for calculation of geometric means (APC&EC Reg. 2.507). Summer 2015 E. coligeometric means ranged from 120 most probable number (MPN) /100mL in June to 392MPN/100mL in July.ADEQ received over 150 comments during the 2016 303(d) list public comment periodrequesting that Mill Creek and additional Buffalo River tributaries be added to the list ofimpaired waterbodies. Commenters were concerned that E. coli concentrations in Mill Creekexceeded the state water quality standard and were degrading water quality of the Buffalo River.ADEQ’s 2016 303(d) Responsiveness Summary noted that the NPS data received did not satisfy2016 Assessment Methodology temporal requirements and thus were unable to be used for CleanWater Act attainment decisions (ADEQ 2016). On July 19, 2017, US EPA approved Arkansas’s2010, 2012, 2014, and 2016 303(d)/305(b) Integrated Reports (EPA 2017).ObjectiveIn 2016, ADEQ Office of Water Quality (OWQ) initiated a two-year bacteria monitoring studyin the Mill Creek watershed to determine if the primary contact recreation designated use wasbeing supported per APC&EC Regulation 2.507.

Figure 1. Buffalo National Park bacteria advisory decision matrix (NPS, personal comm.).MethodsStudy AreaLocated in the Ozark Highlands Ecoregion of Arkansas, Mill Creek (watershed area 21.3mi2)originates in south central Boone County and flows south into Newton County for 5.2 river milesto its confluence with the Buffalo National River. Mill Creek is a perennial, 3rd order, high

gradient, Ozark stream with riffle, run, pool sequencing. Karst features are present throughoutthe watershed, including springs and subsurface drainage. Watershed land use is predominantlyforested (77%) with some pasture (18%). Developed land (5%) includes the abandoned themepark, Dogpatch USA (now Heritage USA), and the rural community of Marble Falls. Twoimpoundments on Mill Creek were constructed within the boundaries of theme park in the late1960s and served as trout fisheries until abandonment in the mid-1990s. The impoundments daminflow from two springs, commonly referred to as Upper Dogpatch Spring and Lower DogpatchSpring.Figure 2. Distribution of sample locations throughout the Mill Creek (Newton Co.) watershed.Bacteria Collection and In-situ Water QualityDuring the summers of 2016 and 2017, Mill Creek watershed was sampled for E. coli and in situwater quality parameters: dissolved oxygen (mg/L), temperature ( C), and specific conductance(µS/cm). Twelve collection sites were established throughout the watershed (Table 1; Figure 1).One tributary collection site (BUFT0406) was discontinued in 2017 due to ephemeral flow. Ninesampling events per year were distributed throughout the primary contact season, as defined in

APC&EC Reg. 2.507 (May 1-September 30). Sample container was triple rinsed with nativewater. Water samples were collected from mid-channel and a 100mL aliquot of ambient waterwas transferred to a sterile plastic container. Dechlorination tables (sodium thiosulfate) wereadded if treated effluent samples were collected. Containers were labeled with sampling stationidentifiers and immediately placed on ice until processing at ADEQ Environmental MediaLaboratories. E.coli processing and colony counts followed EPA Method 1603 (2014) andADEQ’s approved 2014 Quality Assurance Project Plan. Data were reported as cfu/100mL.In situ water quality parameters were measured with a YSI ProDSS, YSI 550A, YSI 100, or YSI556 meters. Meters were calibrated per manufacturer’s instructions and calibration values wererecorded. All in situ measurements were collected and recorded after decanting a 100mL aliquotof water from the container for E. coli analysis. In-situ readings were recorded on ADEQ’ssample requisition form.Table 1. Sampling site name, coordinates, and brief site description.ADEQ Site NameWHI0213WHI0211BUFT0402WHI0212OUTFALL 4-93.136073-93.127768DescriptionMill Creek north of DogpatchMill Creek directly below Dogpatch and Marble Falls relift 1Mill Creek at Spring Valley Rd.Unnamed tributary south of Outfall 001.Marble Falls SID permitted outfall.Harp Creek prior to merging with Mill Creek (Lower)Harp Creek below James Spring (Upper)Mill Creek at Mill Creek CampgroundMill Creek at Pruitt-Yardell Rd bridge upstream of spring outfallBuffalo River upstream of merging with Mill Creek (Pruitt)Buffalo River downstream of Mill Creek at Crow Hole

Table 2. ADEQ monitoring site, associated watershed size, and individual sample (IS) E. colicriteria, and results of primary contact site evaluation.ADEQ Site NameWHI0213WHI0211BUFT0402WHI0212Marble Falls SID 0304Watershed Size (mi cal Season IS E. coli Criteria410410410410410410298298Site Verification for Primary ContactPrimary contact unlikelyPrimary contact unlikely*YesPrimary contact unlikelyNA*YesPrimary contact unlikelyNA*YesNANANA*Regulation 2.507, “For the purpose of this regulation, all streams with watersheds less than 10mi 2 shall not bedesignated for primary contact unless and until site verification indicates that such use is attainable.”ResultsMean primary contact season water temperatures in 2016 were 2ºC warmer than in 2017. Meancombined temperature data for 2016 was 22.3( 3.5)ºC compared to 2017 of 20.2 ( 3.3)ºC. Meanwater temperatures for 2016 ranged from 16.82 ( 0.8)ºC at Mill Creek at Highway 7 (WHI0213)to 27.78 ( 1.81)ºC at Buffalo River at Pruitt (BUFR03). Mean 2017 temperatures ranged from16.66 ( 1.5)ºC at WHI0213 to 24.67 ( 3.1)ºC for BUFR03 (Table 3). Lowest mean pH valuesfor 2016 and 2017 were recorded at WHI0213 (Table 3). Highest mean dissolved oxygen wasobserved at WHI0213 for 2016 and second highest mean dissolved oxygen during 2017, onlybehind Upper Harp Creek (WHI0210) (Table 3). Specific conductance was variable across sitesand years; however, for both years highest and lowest mean specific conductance values were atWHI0212 (Unnamed tributary to Mill Creek) and BUFR03, respectively (Table 3).ADEQ Office of Water Quality staff collected 297 E. coli samples between 2016 and 2017primary contact seasons. Highest mean E. coli concentrations were observed at WHI0210 (UpperHarp Creek) (451 418) and BUFT0401 (Mill Creek Campground) (476 781) for 2016 and2017, respectively (Table 4; Figures 3, 4). Highest variability among E. coli concentrations wasobserved at BUFR03 (Buffalo River at Pruitt) in 2016 and WHI0211 (Mill Creek belowDogpatch) in 2017 (Table 4). Lowest mean E. coli was observed at BUFR03, BUFR0304, andWHI0211 (Buffalo River at Pruitt, Buffalo River at Crow Hole and Mill Creek below Dogpatch)through both years (Figures 3, 4).Evaluation of individual samples to the applicable water quality criteria indicate that only oneexceedance of ERW criterion of 298 cfu/100 mL was observed in 2016. In 2017, Mill Creek at

Mill Creek Campground (BUF0401) and Lower Harp Creek (BUFT0405) each had twoexceedances of the 410 cfu/100 mL criterion and BUFT0402 (Mill Creek at Spring Valley Rd.)and BUFT04 (Mill Creek at Pruitt) each had one exceedance above 410 cfu/100 mL.Table 3. Mean in situ water quality parameters collected during 2016 and 2017.ADEQ 27.417.277.467.647.677.757.837.842016WaterTemp 7.197.597.307.667.797.717.747.697.742017WaterTemp .92316.67455.30364.13179.03191.61

Figure 3. Box plots of E. coli concentrations from Mill Creek and Buffalo River during 2016monitoring period with quartile distributions representing minimum, first quartile, median, thirdquartile, and maximum.Figure 4. Box plots of E. coli concentrations from Mill Creek and Buffalo River during 2017monitoring period with quartile distributions representing minimum, first quartile, median, thirdquartile, and maximum.

Table 4. Calculated geometric and arithmetic means (cfu/100 mL SD) for 2016 and 2017 andexceedances of applicable bacteria criterion. Not applicable (NA) applies for stations withwatersheds less than 10mi2 (APC&EC Reg. 2.507).20162017ADEQ 4539001847730ConclusionsThe objective of this two year study was to determine whether Mill Creek was in compliancewith applicable state water quality criteria for bacteria and supporting its primary contactrecreation designated use. Data collected during 2016 and 2017 indicate an acceptable level ofexceedances of applicable bacteria criteria; thus, Mill Creek supports its primary contactrecreation designated use per APC&EC Reg. No. 2.507.However, the NPS has applied the definition of APC&EC Reg. No. 2.302(A) (ExtraordinaryResource Waterbody) to all tributaries of the Buffalo River and subsequently applied ERWspecific E. coli criteria for protection of primary contact recreation within NPS boundaries.Review of ADEQ’s 2016 and 2017 data compared to the individual sample ERW criterion of298 cfu/100mL, only six of the ten monitoring location met applicable watershed threshold of 10 mi2. In the main stem of the Buffalo River, one ERW E. coli criterion exceedance wasobserved at BUFR03 in 2016, and no exceedances were observed at BUFR03 in 2017 or atBUFR0304 in 2016 or 2017. In 2016, BUFT0401 (Mill Creek at the campground) was the onlysite with one exceedance outside of the main stem Buffalo River site. In 2017 BUFT0401 (MillCreek at the campground), BUFT0402 (Mill Creek downstream of Dogpatch), and BUFT0405(lower Harp Creek) all had at least two samples exceed the criteria. Comparisons of E. coli datato the applicable water criterion as well as the more stringent NPS approach suggests that the

number of exceedances observed during 2016 and 2017 support attainment of Mill Creek’sprimary contact recreation designated use.In 2010, the NPS reported highest arithmetic mean concentrations (273cfu/100mL) from HarpCreek (BUFT0405) and lowest mean (72 cfu/100mL) from Spring Valley Road (BUFT0402).Interestingly, NPS reports that the Mill Creek Campground site (BUFT0401) and Harp Creek(BUFT0405) had higher E. coli concentrations during base flow events than storm flow,indicating more localized bacteria sources.Since 1985, the NPS has routinely collected bacteria from Mill Creek near the mouth (BUFT04).Arkansas Watershed Conservation Resource Center (AWCRC) (2017) summarized 27 years offecal coliform data for Buffalo River and tributaries. Arithmetic mean and geometric mean werecalculated for 62 samples during base flow conditions, which were 43.4 and 8.7 cfu/100mL,respectively. The 2017 AWCRC report also compared fecal coliform geometric mean from1985-1994 (Mott 1997) and 1995-2011. Mott (1997) reported fecal coliform geometric meanfrom 1985-1994 as 14 cfu/100mL, while the 1995-2011 geometric mean increased by 3.7 cfu to17.7cfu/100mL. The individual sample criterion for fecal coliform in non-ERW waterbodies is400 for primary contact season and 2,000 cfu/100mL for secondary contact (APC&EC 2017).Fecal coliform and E. coli concentrations are generally correlated in freshwaters (Francy et al.1993); however, the NPS dataset lacks paired bacteriological samples to evaluate the relationshipbetween the two indicators in Mill Creek or the Buffalo River.Findings from ADEQ’s two year study are consistent with Maner and Mott (1991) and NPS(2010), in which highest bacteria concentrations were observed in the vicinity of the Mill CreekCampground and the confluence of Harp Creek. The NPS reported that bacteria sources in MillCreek are still unclear, but speculate that bedrock features within this vicinity may precludeseptic infiltration into the soils, thus leading to chronic bacteria contamination. In an effort todetermine bacteria as well as nutrient sources, ADEQ and the United States Geological Survey(USGS) have partnered to study what, if any, anthropogenic or agricultural sources can bedetected in Mill Creek. The study is designed to sample base and storm flow events whileanalyzing for multiple constituents indicative of both forms of non-point source contributions. Indoing so, multiple lines of supporting evidence can be provided to document bacteria andnutrient source contributions to Mill Creek.

ReferencesAley, T. and C. Aley. 2000. Inventory and delineation of karst features, Buffalo National River,Arkansas. Report on Phase 2 investigations and final project report. Ozark UndergroundLaboratory, Protem, Missouri.Aley, T, 2010. Assessment of the Impacts of Leaks in the Marble Falls Sewage System onWater Quality in Mill Creek and the Buffalo River. Ozark Underground Laboratory, Inc.1572 Aley Lane, Protem, MO 65733. (417) 785-4289.Arkansas Department of Environmental Quality. 2016. Responsiveness Summary to f Accessed September 22, 2017.Arkansas Department of Environmental Quality. Consent Administrative Order Case No. CV2010-10-1Arkansas Department of Environmental Quality. 2014. Arkansas’ Water Quality and ComplianceMonitoring Quality Assurance Project Plan (QTRAK #14-351). 76pp.Arkansas Pollution Control and Ecology Commission. 2017. Regulation No. 2. RegulationEstablishing Water Quality Standards for Surface Waters of the State of Arkansas. August 25,2017. 129pp.Arkansas Watershed Conservation Resource Center. 2017. Surface-Water Quality in the BuffaloNational River (1985-2011). Prepared for: National Park Service-Buffalo National River. 71pp.Cheri, Kevin G., 2015. National Park Service letter to ADEQ Director Becky Keogh. October 6,2015. 3 pp.Cheri, Kevin G. 2016. National Park Service letter, “Arkansas 2016 list of impaired, 303(d) .pdfFrancy, D.S., D.N. Myers, K.V. Metzker. 1993. Escherichia coli and fecal-coliform bacteria asindicators of recreational water quality. Water-Resources Investigations Report 93-4083. 38pp.Great Lakes-Upper Mississippi River Board of State and Provincial Public Health andEnvironmental Managers. 2012. Recommended Standards for Water Works (Ten StateStandards). Policies for the Review and Approval of Plans and Specifications for Public WaterSupplies. Health Research Inc., Albany, NY. 166pp.Maner, M. and D. Mott. 1991. Mill Creek Survey. Arkansas Pollution Control and EcologyCommission and National Park Service. Unpublished report. 38pp.Mott, David N., 1997, Ten Years of Water Quality Monitoring, Buffalo National River,Arkansas: Harrison, Arkansas, 34 pp.

Mott, D.N., Hudson, M.R., and Aley, T., 1999, Nutrient loads traced to interbasin groundwatertransport at Buffalo National River, Arkansas, in Harmon, D., ed., Proceedings of the 10thConference on Research and Resource Management in Parks and on Public Lands, GeorgeWright Society, Hancock, Michigan, p. 114-121.Mott, D.N., Hudson, M.R., Aley, T. 2000. Hydrogeologic investigations reveal interbasinrecharge contributes significantly to detrimental nutrient loads at Buffalo National River,Arkansas. In Steele, K.N., ed., Proceedings of the Arkansas Water Resources Center AnnualConference: Environmental Hydrology. Publication No. MSC-284. University of Arkansas,Fayetteville, Arkansas. p. 13-120Usrey, F.D. 2011. Bacteria Monitoring for Mill Creek and Associated Reaches of Buffalo RiverStatus and Trends for 2010. Division of Fire and Resource Management. Buffalo National River.38pp.U. S. Environmental Protection Agency, 1986. Ambient Water Quality Criteria for Bacteria.EPA s/drinking/upload/2009 04 13 beaches 1986crit.pdfU.S. Environmental Protection Agency. 2006. Method 1603: Escherichia coli (E. coli) in Waterby Membrane Filtration Using Modified membrane-Thermotolerant Escherichia coli Agar(Modified mTec). 42pp.U.S. Environmental Protection Agency. 2017. EPA Action on Arkansas’s 2010, 2012, 2014, and2016 § 303(d) Lists. ated/303d/pdfs/2017/epadecision-7192017.pdf

In situ water quality parameters were measured with a YSI ProDSS, YSI 550A, YSI 100, or YSI 556 meters. Meters were calibrated per manufacturer’s instructions and calibration values were recorded. All in situ measurements were collected and recorded after decanting a 100mL