Wisconsin Geological and Natural History Survey 3817 Mineral Point Road Madison, Wisconsin 53705-5100 TEL 608/262.7389 FAX 608/262.8086 www.uwex.edu/wgnhs/ James M. Robertson, Director and State Geologist Assessment of Virus Presence and Potential Virus Pathways In Deep Municipal Wells Kenneth R. Bradbury Wisconsin Geological and Natural History Survey Mark A. Borchardt Marshfield Clinic Research Foundation Madeline Gotkowitz Wisconsin Geological and Natural History Survey Randall J. Hunt U.S. Geological Survey 2008 Open-File Report 2008-08 48p. This report represents work performed by the Wisconsin Geological and Natural History Survey or colleagues and is released to the open files in the interest of making the information readily available. This report has not been edited or reviewed for conformity with Wisconsin geological and Natural History Survey standards and nomenclature. Assessment of Virus Presence and Potential Virus Pathways in Deep Municipal Wells Final Report to the Wisconsin Department of Natural Resources Kenneth R. Bradbury Wisconsin Geological and Natural History Survey University of Wisconsin-Extension Mark A. Borchardt Marshfield Clinic Research Foundation Madeline Gotkowitz Wisconsin Geological and Natural History Survey University of Wisconsin-Extension Randall J. Hunt U.S. Geological Survey August, 2008 Revised October, 2008 1 Assessment of Virus Presence and Potential Virus Pathways in Deep Municipal Wells Abstract Among the many waterborne pathogens of humans, enteric viruses have the greatest potential to move deeply through the subsurface environment, penetrate aquitards, and reach confined aquifers. Previous research revealed the presence of viruses in water from two of three deep bedrock wells sampled in Madison, WI. Virus presence in these wells was particularly surprising because the wells were cased through a regional aquitard thought to provide protection for the wells. This present study is a follow-up to the previous work and is intended to (1) obtain a time series of virus, isotopic, and geochemical data from several municipal wells completed in a deep bedrock aquifer, (2) use these data sets to evaluate virus presence and, if present, the potential sources of the viruses and pathways to the wells, and (3) evaluate the possibility that virus transport occurs through the well casing, grout or annular space. During 2007 and 2008 we sampled six deep municipal wells for viruses on an approximately monthly basis. Three of these wells had shallow casings, and three were cased through a regional aquitard. We also collected virus samples from local lakes and from untreated sewage and sampled groundwater and lake water for major inorganic ions and isotopes of hydrogen and oxygen. Viruses were detected at least twice in every one of the six wells, but no well was viruspositive in every sampling round. Overall, 43 percent of the samples were virus-positive, and virus concentrations ranged from 0.00 to 6.15 genomic copies per liter (gc/l), with a mean of 0.47 gc/l. Samples from three wells were positive for virus infectivity. Lake samples were positive 78 percent of the time, and ranged from 0.00 to 27.6 gc/l, with a mean of 5.8 gc/l. Not surprisingly, Madison sewage was extremely high in viruses, with all samples positive, and concentrations ranging from about 50,000 to over two million gc/l, with a mean of 581,000 gc/l. Virus results varied significantly with time, and there is apparent correlation between virus levels in sewage, lakes, and groundwater. Several different species (serotypes) of viruses were identified in wells, sewage, and lake water during this study, and in many cases wells and sewage contained identical virus serotypes. Detected viruses include Enteroviruses echovirus 3, echovirus 6, echovirus 11, Coxsackie A16 and B4, Adenoviruses 2, 6, 7, 41, as well as G1 norovirus and Rotovirus. The apparent correlation between viral serotypes found in sewage, lakes, and groundwater suggests very rapid transport from the sources to wells. Viral serotypes vary seasonally and annually, and so correlation between surface and subsurface serotypes would be unexpected if transport times from the surface to groundwater exceed many months. The Madison Lakes are probably not the main source of the viruses found in the wells as lake water contained some but not all of the serotypes found in the wells, and wells without lake-derived water had viruses present. Furthermore, the 18O/2H signature of water produced by these wells is not consistent with a significant lake water 2 component of recharge to most of the wells sampled. Virus levels in surface water were much lower than in sewage, thus significant volumes of lake water would be required to produce the virus levels measured in the wells. The most likely source of the viruses in the wells is the leakage of untreated sewage from the Madison sewer system. Given the high concentrations (millions of genomic copies per liter) of viruses in sewage, it would take very little sewage to produce the virus concentrations observed in the wells. Human enteric viruses might be excellent tracers of recently recharged groundwater in urban settings if virus sources exist. They have the desirable tracer characteristics of detectability over several orders of magnitude, high mobility, and are time-specific due to constantly changing serotypes. Although the presence of detectable tritium in a well is almost always an indicator of recent recharge to the well, the absence of tritium (at a detection limit of 0.8 TU) does not necessarily indicate that the well will be virus-free. In fact detection of viruses many be a far more sensitive indicator than tritium of a proportion of "young" groundwater in a well. 3 Table of Contents Abstract ............................................................................................................................... 2 Table of Contents................................................................................................................ 4 Introduction......................................................................................................................... 5 Virus contamination of groundwater .............................................................................. 5 Previous virus sampling in the Madison area ................................................................. 5 Project objectives and scope ........................................................................................... 8 Acknowledgments........................................................................................................... 8 Procedures and Methods ..................................................................................................... 9 Selection of wells for sampling....................................................................................... 9 Sampling procedure at municipal wells........................................................................ 12 Sampling procedure at lakes ......................................................................................... 13 Sewage influent sampling ............................................................................................. 13 Virus analyses and sequencing ..................................................................................... 13 Isotopic and geochemical sampling and analysis ......................................................... 14 Results............................................................................................................................... 15 Precipitation, climate, and water levels during the study period .................................. 15 Viruses .......................................................................................................................... 18 Well-by-well virus results......................................................................................... 20 Virus speciation and Infectivity ................................................................................ 21 Groundwater and lake water chemistry ........................................................................ 23 Environmental Isotopes in virus study wells ................................................................ 23 Discussion ......................................................................................................................... 28 Significance of virus detections .................................................................................... 28 Potential virus pathways to wells.................................................................................. 28 Lakes as a source of viruses.......................................................................................... 28 Sanitary sewers as a source of groundwater contamination ......................................... 28 Relationships between sewer leakage and the hydrogeologic setting .......................... 29 Calculated source volumes of viral contaminants ........................................................ 30 Conclusions and Recommendations ................................................................................. 33 Conclusions................................................................................................................... 33 Recommendations......................................................................................................... 34 References......................................................................................................................... 35 Appendices........................................................................................................................ 37 Appendix A: Virus results. ........................................................................................... 37 Appendix B: Geochemical results. ............................................................................... 40 Appendix C: Field measurements................................................................................. 46 4 Introduction Virus contamination of groundwater Among the many waterborne pathogens of humans, enteric viruses have the greatest potential to move deeply through the subsurface environment, penetrate aquitards, and reach confined aquifers. Enteric viruses are extremely small (27-75 nm), readily passing through sediment pores that would trap much larger pathogenic bacteria and protozoa. Viruses have been found in groundwater at depths of 67 m (Keswick and Gerba 1980; Robertson and Edberg 1997) and 52 m (Borchardt et al 2003) and lateral transport has been reported as far as 408 m in glacial till and 1600 m in fractured limestone (Keswick and Gerba 1980). Several recent studies have demonstrated widespread occurrence of viruses in domestic and municipal wells in the United States (Abbaszadegan et al 2003; Borchardt et al 2003; Fout et al 2003; Borchardt et al 2004), and approximately half of waterborne disease outbreaks attributable to groundwater consumption in the United States have a viral etiology (National Primary Drinking Water Regulations, 2006) . The US Environmental Protection Agency has listed several viruses on its drinking water Contaminant Candidate List, emphasizing that waterborne viruses are a research priority (http://www.epa.gov/safewater/ccl/index.html). Although the vulnerability of groundwater to virus contamination is now recognized, the occurrence of viruses in confined aquifers has rarely been explicitly investigated. In the most comprehensive groundwater-virus study to date, Abbaszadegan et al (2003) sampled 448 groundwater sites in 35 states and found 141 sites (31.5%) were positive for at least one virus type. Previous virus sampling in the Madison area During 2005 and 2006 we undertook initial virus sampling of three deep bedrock wells serving the city of Madison, Wisconsin (Borchardt et al. 2007a). Each of these highcapacity wells is over 700 feet deep and cased to at least 220 feet below the surface. The vertical hydraulic gradient is downward due to a major cone of depression beneath Madison. Two of the wells (wells 7 and 24) are cased through the Eau Claire shale, a regional aquitard described by Bradbury and others (1999) and thought to provide excellent protection to the underlying sandstone aquifer. A third well (well 5, now abandoned) was open both above and below the shale. Conventional wisdom suggested that viruses would not be detected in any of the three wells due to the probable long travel times from the surface to the wells, the depths of the wells, and the assumed short (six months to two years) lifetime of the viruses. The surprising result of the study was that viruses were repeatedly detected in the two wells thought to have greatest protection due to their deep casings (wells 7 and 24). Viruses were detected in 4 of 10 samples from well 7 and 3 of 10 samples from well 24 (Borchardt et al. 2007a). Moreover, five of the seven positive samples tested positive for infectivity, suggesting relatively rapid transport from the virus source to the wells. Replicate sampling and careful laboratory procedures have ruled out laboratory contamination as a source for the viruses. The human enteric viruses detected include serogroups coxsackieviruses and echoviruses as wells as poliovirus vaccine strain Sabin 1. The Madison, Wisconsin wells are typical of 5 wells now in use in many cities throughout Wisconsin and the United States. These highcapacity wells range in age from less than five to over 50 years and were constructed according to accepted well drilling practices, which include grouted well casing to depth. The wells produce water from one or both of two aquifers. The shallow bedrock aquifer is composed of sandstone and dolomite. The deeper bedrock aquifer is composed of sandstone. A regional aquitard, the Eau Claire aquitard, is composed of shale and siltstone, and separates the two aquifers, but may contain fractures or be absent beneath the nearby Madison lakes. Although the water utility samples the wells regularly for a long list of organic and inorganic contaminants, including bacteria, the wells are not tested for viruses, presumably because viruses have not been thought to be present in the subsurface. Our previous work in Madison shows that this assumption is false. Understanding how the viruses moved from a near-surface source (humans) to the deep bedrock wells is critical to assessing the magnitude of the virus problem, the human health risks, and to developing remedial actions. However, based on the limited sampling to date it was difficult to elucidate a pathway or mechanism to deliver the viruses to the wells. Given that the viruses originated near the land surface there are four conceptual models of virus transport to the confined aquifer: (1) transport through the aquitard by porous-media flow; (2) transport by porous-media flow around the edge of the aquitard or through nearby "windows" or breaches in the aquitard, including local lakes; (3) transport by rapid flow through fractures in the aquitard or through cross-connecting nearby wells; and (4) transport by rapid flow along the well annulus through damaged, deteriorated, or poorly installed grout or breaches in the well casing. Knowledge about the local hydrogeologic system and virus survival time makes some of these conceptual models more probable than others. The only environmental source of human enteric viruses is human fecal waste, and within the city limits of Madison human fecal waste is presumably only present in sanitary sewers. From this presumed point of entry, viruses must travel viruses downward over 200 feet though the upper sandstone aquifer, an additional 10 to 30 feet downward through the Eau Claire aquitard to reach the top of the Mount Simon aquifer. Once in the Mt Simon aquifer the viruses must move laterally some unknown distance to the production wells. Based on such a travel path, pathway 1 seems very unlikely because travel times would likely be far longer than the six months to two years these viruses can survive in the environment (Yates et al 1985, John and Rose 2005, Schijven et al 2006). Transport pathways 2 and 3, through breaches in the aquitard or through fracture pathways, are more probable, but one must still account for the long travel distance through the upper sandstone aquifer above the aquitard. Pathway 4, transport down the annulus of the well itself through deteriorated or poorly installed grout or through breaches in the well casing, seems the most likely mechanism for virus transport. This pathway could produce rapid downward movement of water with delivery directly to the well bore. Although the three wells tested in the previous study were drilled, cased, and grouted according to accepted practice it is impossible to confirm that the grout has remained intact over the entire length of the casing in wells that are now 27 years (Well 24) and 41 years old (Well 7). 6 During the previous virus study in Madison (Borchardt and others, 2007a) we collected limited samples for analysis of environmental isotopes. Tritium, deuterium, and oxygen18 have long been used in hydrogeologic studies to help distinguish groundwater age and source areas (Clark and Fritz, 1997). Previous tritium data suggested that Madison wells 5 and 24 produce relatively "old" groundwater (little or no tritium content), while well 7 produces "younger" water (tritium near the levels in modern precipitation). We hoped that oxygen-18/deuterium data would be useful in confirming or discarding flow paths that include surface water contributions from the nearby Madison lakes. However, the oxygen-18/deuterium data were not definitive, possibly due to subsurface mixing and or seasonal variations in the 18O concentrations in precipitation. Hunt and others (2005) showed that a time series of 18O/deuterium rations is necessary to unambiguously distinguish surface-water inputs from terrestrial recharge; the previous study obtained only single isotope samples from each well. In a population, like that of Madison, various viruses have a temporal signature, arriving and disappearing from the population over the course of a year. For example, late summer and autumn is the time of year for enterovirus infections in Wisconsin. Infected people in Madison shed enteroviruses, which are flushed through the sanitary sewers to the sewage treatment plant. There are 64 serotypes of enteroviruses and only a couple of serotypes are present in the population at any given time. One enterovirus strain will be dominant in Madison in August and a different strain dominant in October, which will differ from the strains present the following year. These temporal patterns and changes in the relative abundance of viruses and virus serotypes have been documented in wastewater for enteroviruses and adenoviruses (Sedmak et al. 2003; Sedmak et al. 2005; Carducci et al. 2006). Add in all the other human enteric viruses that can be detected and sequenced, and the viruses in the wastewater shed by the population become a "virus signature" for that point in time. The signatures can be used as a tracer of virus movement from source(s) (presumably leaking sanitary sewers or lake water) to the study wells. Using deuterium and O-18 as an isotope signature, Hunt et al. (2005) used a similar conceptual approach for estimating the time of travel of river water through the riverbank to adjacent wells. The virus signature has several information components: (1) the general type of virus (e.g., norovirus or enterovirus), which gives information on the size, charge, and "lifespan" of the virus particle; (2) the quantity of virus (e.g. genomic copies/liter), which provides a time-varying signal whose amplitude may be observed along the suspected transport route and well; and (3) the virus serotype or nucleic acid molecular fingerprint, which can be tracked over time in wastewater and well water and, in conjunction with virus quantity, gives information on transport time. For example, the presence of echovirus 18 in wastewater in October followed by its detection in a well in December might suggest a 2 month time of travel from the source(s) to the well, but could also suggest a 14-month travel time if echovirus 18 had been present the previous October. This is why obtaining a measure of virus variation in the source water is critical. Of course, one would want to base time estimates on multiple virus detections and samples. Working with these virus signature components as separate lines of evidence, or perhaps combining them using multivariate techniques such as cluster analysis or 7 multidimensional scaling, and corroborated with isotope and chloride data, we believe will allow powerful inferences about virus transport routes to the drinking water wells. One limitation of this approach is that for reasons not well understood among environmental virologists, there is substantial spatial and temporal variability in virus occurrence in groundwater. One approach to compensate for spatial variability is to take large sample volumes (~ 1000 liters) as commonly practiced. An approach to compensate for temporal variability is to increase sampling frequency, which is now affordable. The benefit of collecting numerous large sample volumes is that, spatial and temporal variability notwithstanding, the underlying biological and hydrogeologic patterns begin to emerge. A similar approach was recently reported by Borchardt et al (2007b) where several hundred water samples for viruses allowed the study team to quantify virus intrusions into municipal drinking water distribution systems. Project objectives and scope The objectives of this project are (1) to obtain a time series of virus, isotopic, and geochemical data from several municipal wells completed in a deep bedrock aquifer, (2) to use these data sets to evaluate virus presence and, if present, the potential sources of the viruses and pathways to the wells, and (3) to evaluate the possibility that virus transport occurs through the well casing, grout or annular space. This one-year project was entirely conducted in Madison WI, using wells owned and operated by the Madison Water Utility. Acknowledgments This project was funded by the Wisconsin Department of Natural Resources through the State of Wisconsin Groundwater Joint Solicitation Program. The Madison Water Utility participated in the project by providing information about wells and access to wells for sampling. WGNHS Geotechnician Peter Chase spent many hours in sample collection. Susan Spencer and Phil Bertz of the Marshfield Clinic performed the virus analyses. The Madison Metropolitan Sewerage District provided samples of sewage influent. 8 Procedures and Methods Selection of wells for sampling The Madison Water Utility currently operates 27 deep high-capacity wells completed in bedrock aquifers. The wells draw from a Cambrian-age sandstone aquifer underlying the city (Bradbury and others, 1999). This aquifer lies beneath 30 to 100 feet of glaciallydeposited sand and gravel and lake sediment. Most of these high-capacity wells are over 700 feet deep and cased to about 200 feet below the surface. Water enters the wells through open boreholes in the rock below the casing. Although the well casings are supposed to be sealed to the surrounding geologic materials using cement grout, the integrity of these grout seals is often suspect and nearly impossible to test. About onethird of the wells are cased through the Eau Claire shale, a regional aquitard described by Bradbury and others (1999) and thought to provide excellent protection to the underlying sandstone aquifer. The other two-thirds of the wells, most of which are the older wells, are "cross-connecting"; open both above and below the shale or drilled in places where the shale is thin or absent. These wells are more vulnerable to contamination than the deeply cased wells. The funding level for this project prohibited sampling of all 27 Madison wells. In order to understand the scope of the virus problem we decided initially to sample 11 wells and then sample fewer wells in subsequent rounds. Our rationale was to insure that we were working with some virus-positive wells and that we had a variety of well construction and well locations. We chose six wells reported to be multi-aquifer wells (open both above and below the Eau Claire aquitard) and five wells reported to be cased through the aquitard. We sampled surface water from Lakes Mendota, Monona, and Wingra as well as clarified sewage influent at the Madison Metropolitan Sewage District. Samples were also collected for inorganic chemistry and isotope analyses. Following the initial sampling rounds we chose six wells for repeated monthly sampling. Figure 1 shows the spatial distribution of wells, and figures 2 and 3 show the construction of the sampled wells. Figure 2 also shows the typical conceptualization of subsurface hydrostratigraphy in Madison. The complex geologic stratigraphy is simplified to consist of upper glacial materials (till, sand and gravel, or lake sediment) covering a shallow bedrock aquifer composed of sandstone and dolomite. Shale of the Eau Claire Formation forms a regional aquitard and separates the upper bedrock aquifer from a deep bedrock aquifer composed of sandstone. Crystalline PreCambrian rock bounds the bottom of the system. Vertical hydraulic gradients in groundwater beneath the city are downward due to a regional cone of depression beneath the Madison metropolitan area (Bradbury and others, 1999). Figure 2 shows this diagrammatically - the potentiometric surface of the deep sandstone aquifer is lower than the water table in the shallow aquifer. In this situation water and any contaminants in the upper aquifer have the hydraulic potential to move vertically downward and reach the underlying deep aquifer. Wells are typically cased and grouted through the upper geologic units and consist of open holes below the casing. 9 Construction diagrams of individual wells (figures 2 and 3) show the variation in well construction, thickness of layers, and presence and thickness of the aquitard. Wells 11, 12, 13, 16, and 17 are termed "cross-connected" wells because either the aquitard is missing completely (wells 11, 13) or the well casings do not extend through the aquitard (wells 12, 16, 17) and the open hole provides a vertical conduit between the upper and lower aquifers. These cross-connecting wells are much more susceptible to contamination than "confined" wells (wells, 7, 8, 19, 24, 28, and 30), in which the casing extends through the aquitard. Figure 1. Location of sampled wells and virus detections. "Positive" denotes a well testing positive for viruses on at least one date. Numbers refer to Madison Water Utility well numbers. 10 well glacial materials 50 casing well 11 18 water table shallow aquifer potentiometric surface casing well 12 35 well 13 casing 60 relative depths, feet below surface 111 relative depths, feet below surface 260 370 404 418 open hole casing 260 300 aquitard 320 deep aquifer open hole open hole 752 800 crystalline rock not to scale 780 986 well 7 30 casing well 19 15 casing well 30 23 casing relative depths, feet below surface 215 225 237.6 245 255 260 266 277 312 open hole open hole 725 718 800 not to scale Figure 2. Construction details of the municipal wells sampled throughout the project. Diagram at upper left shows typical hydrostratigraphy and well construction for the Madison area. open hole open hole 11 well 8 well 16 well 17 well 24 28 well 28 casing casing casing relative depths, feet below surface 135 255 264 280 300 225 435 465 open hole open hole open hole 201 240 200 365 230 open hole 235 395 casing 110 ? 115 casing 400 725 774 797 1005 1005 not to scale Figure 3 . Construction details of the additional five wells sampled during the initial phase of the project. Sampling procedure at municipal wells All well samples were collected at the wellhead while the well pumps were running. Viruses were concentrated using glass wool filters, a method that has been fully validated (Lambertini et al. 2008). Samples were obtained from a sampling tap on the well discharge line prior to discharge to the well reservoir. At wells where the pH exceeded 7.5, the pH was adjusted to between 6.5 and 7.0 using an acid injection ahead of the filter. The Madison wells are plumbed so that there is zero back pressure between the reservoir and the well discharge line; this lack of pressure required the use of a booster pump to force the sample through the glass wool filter. We used a portable heavy-duty peristaltic pump and food-grade tubing for this purpose; the pump and tubing were sterilized with a chlorine solution between each sample. Sampling each well required several hours of pumping; between 700 and 1000 liters of water were passed through the filter and the filtered volume was measured using a flow accumulator. A field blank was collected by pumping nineteen liters of reverse-osmosis water through a glass wool filter, using decontaminated field equipment. The filters were stored, transported and analyzed as described below. open hole 12 Sampling procedure at lakes The procedure for sampling lakes was similar to that for sampling the wells. A decontaminated pump and tubing were submerged in the lake, approximately 10 feet offshore. The water was pumped through a pre-filter to remove particulate matter. The sample stream was then acidified to a pH between 6.5 and 7.0, because the lake water was typically above pH 7.5. The acidified influent was split between two glass wool filters used in parallel. Filter effluent was directed onto the lake shore. Lake water was pumped at a rate of approximately 4 liters/minute until a total sample volume of about 1000 liters was passed through the filters. The pre-filter and two glass wool filters were transported on ice to Marshfield for analysis. The field equipment was decontaminated according to Marshfield standard procedures prior to re-use. Sewage influent sampling Clarified and settled sewage influent was collected and provided by the staff of the Madison Metropolitan Sewerage District at the Nine Springs sewage treatment plant. The influent was transferred to four-liter containers and shipped to Marshfield for analysis. Virus analyses and sequencing Pre-filters and glass wool filters were transported to the laboratory on ice and processed the next day after sampling. Filters were eluted with beef extract/glycine and the eluate flocculated and concentrated with polyethylene glycol following the methods described in Borchardt et al (2004) and Lambertini et al (2008). Samples were analyzed for six virus groups: enteroviruses, adenoviruses, rotavirus, hepatitis A virus (HAV), and norovirus genogroups 1 and 2. Viruses were detected by real-time quantitative reverse-transcription polymerase chain reaction (RT-PCR) and TaqMan probe using the LightCycler (Roche Inc.) platform. The procedures, primers, and probes are described in Borchardt et al (2003, 2004) and Lambertini et al (2008). Standard curves were established by treating stocks of each virus type with Benzonase (Novagen, Madison, WI) for 30 min at 37?C, followed by incubation for 2 days at 4?C, leaving only the nucleic acid contained within intact capsid-protected virions, and removing extraneous viral nucleic acid that would have inflated the estimate of genomic copy number. Viral RNA or DNA mass was measured fluorometrically using RiboGreen (Molecular Probes, Eugene, OR) or PicoGreen (Molecular Probes) and a CytoFluor Series 4000 fluorimeter (Applied Biosystems, Framingham, MA), then converted to genomic copies based on the nucleic acid molecular weight of that virus. Intact viruses were serially diluted, and each dilution was seeded into separate 0.14 ml volumes of negative final concentrated sample volume (FCSV) and extracted using the QIAamp DNA Blood Mini Extraction Kit (Qiagen). Therefore, the standard curves represent the entire quantitation process and include any matrix effects from the elution and flocculation procedures. Crossing points were calculated automatically by the 13 LightCycler with the second derivative maximum method, and plotted against the decimal logarithm of viral RNA or DNA concentration. RT-PCR controls for each batch of reactions included an extraction negative control (unseeded FCSV), negative controls for the RT and PCR cocktails, and a positive control of known low viral concentration seeded into an FCSV matrix. This positive control also served as the LightCycler reference control, validating the use of the standard curves. qRT-PCR inhibition was evaluated by seeding 800 copies of hepatitis G virus (HGV) Armored RNA(R) (Asuragen Inc., Austin, TX) into the RT reaction of every sample. qRTPCR was performed using HGV primers provided by the manufacturer and a laboratorydesigned probe. Inhibition was considered absent when the crossing point of the HGV seeded samples was less than one cycle higher than the inhibition reference control (crossing point = 32). Samples that were qRT-PCR-positive for enteroviruses were further evaluated for virus infectivity by cell culture using three cell lines (BGMK, RD, and Caco-2). Infectivity was gauged by two outcome measures: 1) observation of cytopathic effect (CPE) in cultures held six weeks; 2) a >= 10-fold increase in virus genomic copies in cell lysates from 2 week or 6 week cultures compared to the initial virus quantity in the FCSV cell culture inoculum. All enterovirus and adenovirus positive samples were identified to serotype by sequencing using the ABI Prism 3100 Genetic Analyzer and previously described methods (Borchardt et al 2004 and 2007). Isotopic and geochemical sampling and analysis Samples for major ions and isotopes were collected at the municipal wells from the sampling tap while the wells were running. Field collection followed standard procedures for collection of field parameters (pH, temperature, dissolved oxygen), filtration, and acidification of metals (e.g. Karklins, 1996). Surface water samples were collected from open water along the shoreline during periods when the lakes were fully mixed. Samples were analyzed for the following parameters: Ca, Mg, Na, K, Fe, Mn, HCO3, SO4, NO3, Cl. Analyses were conduced at the Wisconsin State Laboratory of Hygiene, a certified water analysis laboratory. Isotope samples were analyzed at the University of Waterloo (Ontario) Environmental Isotope Laboratory or at the US Geological Survey Isotope Laboratory. Deuterium was determined by manganese reduction. Oxygen-18 was determined by mass spectrometry on CO2 gas. Tritium was determined by liquid scintillation counting on enriched samples. 14 Results Precipitation, climate, and water levels during the study period The Madison area received unusually high precipitation during the study period. Figure 4 shows the distribution of precipitation and air temperature between July, 2007 and September, 2008. Intense rainfall during August, 2007 caused minor flooding during that Fall. Record snowfall (over 100 inches) occurred during the winter of 2007-2008. Finally, June, 2008 was the second wettest month on record, with a rainfall of 10.9 inches in the Madison area (MMSD, 2008). Very intense rainfall between June 9 and 12, 2008 cause major flooding across southern Wisconsin. Surface-water and groundwater levels and storm sewer flows responded to the precipitation events. Figure 5 summarizes storm sewer flows, the elevation of Lake Mendota, and groundwater levels in two local monitoring wells. Rapid increases in groundwater levels show that rapid recharge occurred after storm events. The Spring Harbor storm sewer drains street runoff from west Madison and discharges into Lake Mendota. It is one of several such storm sewers in the Madison area. Maximum storm flows occurred after the heavy rains in August 2007 and June 2008. A significant flow event also occurred during early January, 2008 following an unusually warm "January thaw". The June, 2008 precipitation event is also important because it resulted in extremely high flows in the Madison sanitary sewers (MMSD, 2008). Sewage flows often increase during precipitation evens due to stormwater infiltration through leaky sewers and basements. The average flow to the Nine Springs Wastewater Treatment Plant is about 41 million gallons per day (MGD). During the fist significant rains on June 8, flows increased to 122 MGD, and then declined to about 80 MGD for several days. Several discharges of sewage diluted with rainwater in the system occurred during this rain event, on June 9. The largest discharge was into the Cherokee Marsh and the Yahara River upstream of the Highway 113 bridge (1,080,000 gallons). There was a smaller discharge into the Cherokee Marsh on the south side on Golf Road (17,200 gallons). There were also two discharges that would have entered Starkweather Creek (245,000 gallons on the east side of the Dane County Regional Airport and 48,000 gallons near Milwaukee Street), a small discharge into Lake Mendota at Carroll Street, and two small discharges into Squaw Bay on Lake Monona; one on the south shore (50,000 gallons) and one on the east shore (4,000 gallons) (Jon Schellpfeffer, MMSD, written communication). 15 Dane County Airport - Precipitation and air temperature 40 avg daily temp, deg C 20 0 -20 -40 5 precipitation (inches of water) annual cumulative precip (in) avg air temp, deg C 50 4 daily precip, inches 40 3 30 2 20 1 10 0 7/1/07 8/1/07 9/1/07 1/1/08 4/1/08 5/1/08 6/1/08 7/1/08 8/1/08 10/1/07 11/1/07 12/1/07 9/1/08 2/1/08 3/1/08 0 Figure 4. Precipitation and air temperature in the Madison area 16 736 water table elev, DN-83 ft above msl 734 732 730 728 726 80 70 60 50 40 30 20 10 0 7/1/07 8/1/07 9/1/07 852 850 848 846 844 842 856 flow at Spring Harbor storm sewer, CFS 852 848 844 840 1/1/08 2/1/08 3/1/08 4/1/08 5/1/08 6/1/08 7/1/08 8/1/08 10/1/07 11/1/07 12/1/07 9/1/08 date Figure 5. Stormwater flows, lake levels, and groundwater levels during the study period. Wastewater flows are from the Spring Harbor Storm Sewer (USGS site ID 05427965). Groundwater levels are from observations well DN-83 (USGS) and DN-1464 (unpublished data). elevation of Lake Mendota feet above msl hydraulic head, upper sandstone Nine Springs site, ft above msl DN-83 water table upper aquifer water level Lake Mendota elev Spring Harbor storm sewer flows 17 Viruses The overall virus sampling consisted of 95 samples from wells, lakes, and sewage influent (Appendix A). Well samples included 76 samples from 11 different wells. The three Madison lakes (Mendota, Monona, and Wingra) were each sampled three times. Sewage influent was sampled at ten different dates. The initial sampling rounds (September and October, 2007) consisted of eleven wells (wells 7, 8, 11, 12, 13, 16, 17, 19, 24, 28, and 30). Following the October round we selected six wells (7, 11, 12, 13, 19, and 30) for regular monthly sampling. Our selection was based on initial virus detection, well construction, and some wells being off-line during the winter months. Table 1 summarizes the overall virus results by sample source. Overall, water samples from wells were positive for viruses in 43 percent of the samples, and virus concentrations ranged from 0.00 to 6.15 gc/l (genomic copies per liter), with a mean of 0.47 gc/l. Lake samples were positive 78 percent of the time, and ranged from 0.00 to 27.6 gc/l, with a mean of 5.8 gc/l. Not surprisingly, Madison sewage was extremely high in viruses, with all samples positive, and concentrations ranging from about 50,000 to over two million gc/l, with a mean of 581,000 gc/l. Virus results varied significantly with time, and there is apparent correlation between virus levels in sewage, lakes, and groundwater. Figure 6 shows the percentage of virus detections in wells along with virus concentrations in sewage and lake water. During the fall and winter of 2007, the wells were about 50 percent virus-positive. The positive percentage declined to about 20 percent in early 2008, and to zero in late May, 2008 before jumping to over 80 percent in July, 2008. Virus concentrations in sewage, while always in the thousands of gc/l, peaked in November, 2007, declined through May, 2008, and then rose in July, 2008. Although the lakes were only sampled three times, these samples are consistent with the apparent temporal trend. All three lakes contained viruses in September, 2007. Only lake Mendota contained detectable viruses in May, 2008, but all three lakes were positive in July, 2008. It is interesting to note that the July increases in virus detections followed the extreme rainfall events in June, 2008. Table 1. Summary of virus detections by water source Virus detection (gc/l) Water source Percent positive min max mean Wells 43.4 0.00 6.15 0.47 Lakes 77.8 0.00 27.6 5.80 Sewage 100.0 48,600 2,078,000 581,000 18 overall percentage of virus-positive wells 100 80 60 40 20 0 10 7 wells (number sampled) 6 11 11 6 6 6 6 sewage 6 6 6 6 106 105 104 30 20 10 0 30 20 10 0 30 20 10 0 9/1/07 1/1/08 2/1/08 3/1/08 4/1/08 5/1/08 6/1/08 7/1/08 8/1/08 10/1/07 11/1/07 12/1/07 9/1/08 virus concentration (gc/l) L Wingra 1.26 0.00 0.01 8.91 L Monona 0.00 3.07 27.60 L Mendota 9.05 2.31 sample date Figure 6. Overall virus detections in wells and concentrations in lakes and sewage. 19 Well-by-well virus results Viruses were detected at least twice in every one of the six wells repeatedly sampled for this study, but no well was virus-positive in every sampling round. Figure 7 shows virus concentrations through time for each well, along with the overall percentage of detections in each well. Note that each well had a spike in virus concentrations in June and July 2008. 6 4 2 0 6 4 2 0 well 7, 60% well 11, 60% virus concentration in wells (gc/l) 3 2 1 0 0.8 0.6 0.4 0.2 0 3 2 1 0 well 12, 55% well 13, 27% well 19, 45% 0.1 0.08 0.06 0.04 0.02 0 9/1/07 well 30, 18% 10/1/07 11/1/07 12/1/07 1/1/08 2/1/08 3/1/08 4/1/08 5/1/08 6/1/08 7/1/08 8/1/08 sample date Figure 7. Virus concentrations through time for each of the six long-term wells. Percentages next to well labels show percent virus-positive samples. 9/1/08 20 Virus speciation and Infectivity Several different species (serotypes) of viruses were identified in wells, sewage, and lake water during this study, and in many cases wells and sewage contained identical virus serotypes (table 2). Detected viruses include Enteroviruses echovirus 3, echovirus 6, echovirus 11, Coxsackie A16 and B4, Adenoviruses 2, 6, 7, 41, as well as G1 norovirus and rotovirus. The apparent correlation between viral serotypes found in sewage, lakes, and groundwater is important because it suggests very rapid transport from the surface to groundwater. Viral serotypes vary seasonally and annually, and so correlation between surface and subsurface serotypes would be unexpected if transport times from the surface to groundwater exceed many months. Although some viruses (A41, A2, echovirus 3, echovirus 11) were found in both lakes and wells, other viruses found in wells (A7, echovirus 6, CoxA16) were never found in lakes, suggesting that the lakes are not a source for these viruses in groundwater. With the exception of A7, all viruses found in wells were also detected in Madison sewage. Infectious enteroviruses were found in wells 7, 11, and 19 in some, but not all, samples tested from these wells (infectivity testing on all samples was not completed in time for this report). 21 Table 2. Speciation of viruses detected. Numbers and letters refer to virus serotypes; E6 (echovirus 6) enterovirus, Adneovirus 41, etc. sample period Enterovirus well 7 Adenovirus well 11 long-term wells well 12 well 13 well19 well 30 well 8 well 16 well 28 well 24 L Mendota lakes L Monona L Wingra sewage initial wells Enterovirus Adenovirus Enterovirus Adenovirus Enterovirus Adenovirus Enterovirus Adenovirus Enterovirus Adenovirus Adenovirus Adenovirus Adenovirus Adenovirus Enterovirus Adenovirus Enterovirus Adenovirus Enterovirus Adenovirus Enterovirus Adenovirus other Sep-07 7 Oct-07 E3 41 Nov-07 Dec-07 E6 Jan-08 E6 7 Feb-08 Mar-08 Apr-08 May-08 Jun-08 CoxA16 7 E3 41 E3 41 41 CoxA16 41 2 2 2 E11, 41 41 41 2 E3, E30 41 E30 2 E3 CoxA16 G1 41 E11 41 E3 41 CoxA16 41 E3 Jul-08 CoxB3 7 E11 41 E11 41 41 41 41 41 41 41 E11 41 G1 E11 E3 6 2 E11 CoxB4 41 G1 E11 E6 41 2 G1 2 G1, R E11 Sewage 41 2 G1, R 22 Groundwater and lake water chemistry Samples were collected twice from the six study wells and once from the three lakes for analysis of major ions and tritium (Appendix B). Measurements of pH, specific conductance and dissolved oxygen were collected during each sampling event (Appendix C). Results are available from six sets of well samples and one round of lake samples for the stable isotopes of water, O18 and deuterium. Groundwater from all of the six wells has similar composition, and all six are higher in calcium and lower in chloride than lake water. As shown by a Piper diagram of major ion concentrations (Figure 8), lake water and groundwater are bicarbonate type. Nitrate and chloride are naturally occurring constituents of groundwater, however elevated concentrations of these constituents may be attributed to contamination from septic systems or fertilizer, and road salt. Background nitrate levels in Wisconsin aquifers are generally less than 2 mg/L, and average chloride concentrations in Dane County wells are about 8 mg/L (Kammerer 1981). Chloride and nitrate concentrations in the six study wells are compiled in Table 4. Well 11 has elevated nitrate and chloride, indicating that it receives a relatively large amount of shallow, or recently recharged, groundwater. This conclusion is consistent with its shallow casing depth (111 ft.) and elevated tritium level (Table 3). Well 7 contains elevated chloride and tritium, and although cased through the Eau Claire aquitard, well 7 apparently receives a significant proportion of recent recharge and is vulnerable to contamination from the ground surface. Although Well 13 has a shallow casing depth (128 ft), it has less tritium and lower nitrate and chloride than Well 11, suggesting that it receives a smaller proportion of shallow or recently recharged groundwater. Wells 30 and 19 are cased through the Eau Claire aquitard and are low in nitrate and chloride, indicating little vulnerability to shallow contaminants. However, the appreciable tritium measured in well 19 samples suggests that the well receives a significant proportion of recent recharge. Well 12 is low in nitrate and chloride, and similar to well 30, has no detectable tritium. The apparently small volume of young groundwater that reaches well 12 is surprising it is open to 120 feet of the upper aquifer (figure 2). Environmental Isotopes in virus study wells Tritium (3H) contents and the deuterium (2H) and oxygen-18 (18O) contents of water help discriminate wells and show which wells are most vulnerable to surface-water recharge. Wells 19 and 7 are reportedly cased through the Eau Claire aquitard (Table 4). These wells are located close to lakes (Fig. 2), and oxygen isotope ratios plot to the right of other samples (Fig. 9), shown with a local meteoric water line (LMWL) from Dane County (Swanson and others, 2006). This lighter water suggests some contribution of lake water to these wells. As discussed above, both wells have tritium levels that indicate a significant volume of recent recharge reaches these wells. In contrast, the third confined aquifer well in the study, well 30, has tritium at less than detection (<0.8 TUs) and a ?18O 23 composition lower than reported for modern groundwater by Bradbury and others (1999), Hunt and Steuer (2000), and Kurtz et al. (2007). Lower compositions are indicative of cooler climates; thus, it is likely that well 30 pumps appreciable amounts of glacial melt water from the Pleistocene - amounts not seen in the other study wells. The multi aquifer wells in the study are located further from the Madison lakes, however well 12 is near a retention basin (at Odana Hills) and well 13 is close to Cherokee Marsh. Their oxygen isotope signatures plot to the left of wells 19 and 7, indicating little to no contribution of fractionated surface water at these wells. The variability in a well's isotopic composition can also help identify wells with surface water contributions (Hunt et al. 2005). The median and standard deviation of ?18O collected in wells over the study period form a direct relation because the isotopic composition of terrestrially derived groundwater should reflect little to no surface evaporation and is expected to be less variable than surface water (Hunt et al. 2005). Well 30 has little variability in contrast to well 19 (Fig. 10), supporting the conclusion that terrestrially derived water dominates flow to well 30 (separated from modern water by the aquitard, away from lakes) whereas well 19 has some contribution from lake water. Well 12 also has a greater degree of variability than might be expected given its distance from the lakes and well 13 (which has a similar median isotopic composition). This could result from surface water contributions from the near-by retention basin. Alternatively, Hunt et al. (2005) identify changes to pumping schedule - both in the well of interest as well as nearby wells - as being a mechanism that can affect the variability in water isotope composition. 24 Table 3. Tritium results and well characteristics well or lake 7 30 19 12 11 13 Monona Wingra Mendota * Well 12 distance reported is to storm water retention pond; well is about two miles from Lake Wingra ** Well 13 distance reported is to Cherokee Marsh 1 Tritium reported in Borchardt et al. 2007;samples collected in June, 2003 & May, 2004 2 Tritium reported in Bradbury et al.1999;samples collected in June, 1995 aquifer year constructed total depth (ft) casing depth (ft) distance to surface water (ft) 3411.2 4526.4 836.4 1,115* 4739.6 2,510** Tritium (TUs) Sept., 2007 4.6 <0.8 4.4 <0.8 6.3 2.5 8.7 9.2 8.5 Tritium (TUs) April, 2008 5.40 <0.8 3.70 <0.8 5.40 1.30 Tritium ? 1? 0.6 0.4 0.5 0.4 0.6 0.4 15.12 13.72 11.42 8.91, 9.91, 19.62 ? 1? 0.5 0.3 0.5 0.6 0.9 0.7 0.7 0.8 0.7 confined confined confined multi-aquifer multi-aquifer multi-aquifer 1939 2003 1970 1957 1959 1959 736 800 710 529 752 780 238 312 260 260 111 128 25 80 % +C l 60 % % 80 % 60 Legend L L L I J K J M P MENDOTA MONONA WINGRA WELL 7 WELL 30 WELL 12 WELL 11 WELL 13 WELL 19 + Ca 4 Mg SO 40 % I M J KJ P LL L 20 % 80% Mg +C O 3 20 % % 20 % 40 + Na K HC O 40 % 3 % 20 80% 4 SO % 40 60% 60% 40% 20% J IM KJ P L L L 60 % % 60 40% 20% 80 % 20 % Ca Cl Figure 8. Piper diagram of well and lake water samples. Table 4. Dissolved chloride and nitrate in the study wells. Chloride (mg/L) Well Jun-071 Sep-07 Jan-08 Jun-081 Jun-071 Nitrate (mg/L) Sep-07 Jan-08 0.02 NS 0.73 1.74 ND ND Jun-081 ND 2.58 0.78 1.73 ND ND 7 12.49 10.40 10.50 5.91 ND 0.02 11 45.86 17.30 NS 45.19 2.66 1.14 12 2.49 1.07 1.08 2.62 0.72 0.77 13 8.36 7.74 8.06 8.50 1.71 1.76 19 5.72 3.60 3.70 5.89 ND ND 30 4.11 3.75 2.58 4.39 ND 0.04 1 Data from June 2007 and June 2008 provided by Dane County Health Department. 80 % 40 % 60 % J JI KM PJ % 80 LL L % 80 % 60 % 40 % 20 26 -20 Legend Well 7 Well 30 Well 19 Well 12 Well 11 Well 13 Mendota Monona Wingra LMWL -40 del 2H (?) -60 -80 -10 -8 -6 -4 del 18O (?) Figure 9. Deuterium/oxygen-18 results. N = 6 or 7 samples from wells, n = 1 from lakes (September only). -8 well 19 well 7 well 11 -9 median del 18 O well 13 well 12 -10 well 30 -11 0 0.02 0.04 0.06 del 18O standard deviation 0.08 0.1 Figure 10. Median ?18O compared to standard deviation, n = 6 or 7. 27 Discussion Significance of virus detections Viruses were detected in at least one sample from all but one of the municipal wells sampled for this project and in at least two samples from each of the six wells chosen for long-term sampling. These findings are consistent with our previous work (Borchardt and others, 2007a) and show that even deeply cased municipal wells in confined aquifer settings can be susceptible to pathogen contamination. Potential virus pathways to wells As stated in the introduction to this report, the four conceptual models of virus transport to the confined aquifer include (1) transport through the aquitard by porous-media flow; (2) transport by porous-media flow around the edge of the aquitard or through nearby "windows" or breaches in the aquitard, including local lakes; (3) transport by rapid flow through fractures in the aquitard or through cross-connecting nearby wells; and (4) transport by rapid flow along the well annulus through damaged, deteriorated, or poorly installed grout or breaches in the well casing. This current project has not been able to confirm or discount any of these potential flow paths. We had hoped to undertake in-well borehole sampling during this project in order to evaluate pathway 4 above, however logistical considerations prohibited this work during the past year. We intend to carry out the in-well sampling as part of a follow-up project during 2008-2009. Lakes as a source of viruses Although at first glance infiltrating lake water seems a plausible source for the viruses found in the municipal wells, several lines of evidence show that the lakes are probably not the primary virus source. First, the deuterium/oxygen-18 relationships (figure 9) suggest that only two wells (7 and 19) receive a significant proportion of lake-derived water, while all wells contained viruses. Second, with the exception of the July 2008 levels in Lake Mendota, virus concentrations in the lakes are generally as low as or lower than virus concentrations in the wells. Assuming significant mixing and dilution with virus-free water in the aquifer, the lake virus contents are likely too low to account for the virus levels in the wells. Third, the lakes contained only four of the six virus species detected in the wells. Sanitary sewers as a source of groundwater contamination Sanitary sewers are a major part of civic infrastructure in urban settings and represent a significant potential source of groundwater contamination. Sewer exfiltration, or outward leakage of sewage wastes, represents a potential source of pathogens, toxic chemicals, pharmaceutical compounds and other materials to the subsurface environment (Bishop et al. 1998). There have been two schools of thought on the significance of sewer exfiltration (Rutsch et al. 2008). Some investigators argue that the overall impact of 28 sewer exfiltration is insignificant due to the small volumes of leakage and to biodegradation and sorption of contaminants in the soil zone. Others (e.g. Leif Wolf 2004; Osenbr?ck et al. 2007) believe that exfiltration can be a major source of groundwater contamination. Most studies conclude that the impact of sewage exfiltration on groundwater is quite variable in time and space and there is currently a lack of knowledge about both the quantity of leakage and its consequences for the environment (Rutsch et al. 2008). Relationships between sewer leakage and the hydrogeologic setting Engineers commonly acknowledge that sanitary sewer systems leak. Most urban sewer systems consist of tens to hundreds of miles of buried pipes of a variety of ages, materials, and construction. Leaks can occur due to deteriorated materials, failed joints and junctions, damage from shifting soil or construction practices, tree roots, faulty construction, and many other natural and/or man-made sources. Historically, the overriding concern for sewage and wastewater management and treatment has been sewer infiltration - or groundwater leaking into sewers. Infiltration increases the volumes of sewage to be handled, treated, and disposed of, and can represent a major expense for communities. Sewer utilities usually inspect their lines for damage and infiltration leaks using remotely-operated television cameras, and it is not uncommon to see streams of water entering the sewers through joints or breaks in the pipes (B. Borelli, MMSD, personal communication, 2008). Exfiltration, on the other hand, is much more difficult to quantify. Outward-leaking sewage presents no obvious visual signal in televised pipe inspections, and mass-balance approaches to quantifying exfiltration are difficult because the rates of exfiltration may be below the uncertainty of flow measurements in the sewer system. Moreover, exfiltration is often thought not to pose a risk to the environment because it is expected to operate similar to a septic field whereby subsurface filtering and attenuation mitigates any adverse impact. The relationships between sewers and the local hydrogeologic setting controls the potential for sewer infiltration and exfiltration. Figure 11 shows, in cross section, the four possibilities for sewer construction relative to the water table. In the figure, H1 represents the hydraulic head inside the sewer, and H2 represents the hydraulic head in the adjacent aquifer. There are two types of sewers. Gravity-drain sewers operate along an elevation gradient, and are only occasionally completely full of liquid. More commonly these gravity-drain sewers are only one-third to one-half full (A and B on figure 11). Gravity-drain sewers can temporarily fill under conditions of heavy sewer discharge, or permanently fill at low points in the system. Pressurized or force main sewers (C and D on figure 11) are permanently full of liquid and are connected to booster pumps that maintain positive pressure in the lines. Where leaks exist, the relationship between H1 and H2 controls the flow direction between the sewer and the environment. From figure 11, the only situation where infiltration can occur (H1 < H2) is A, where a gravity-drain sewer lies below the water table. In each of the other three possibilities (B, C, D) the head in the sewer can be higher than the head in the aquifer, and exfiltration can occur. In the cases of pressurized force mains (C, D) the potential for outflow can be very large due to large head differentials (H1>>H2). 29 Wastewater for the entire Madison Metropolitan area is collected through laterals to individual homes and businesses and moved by gravity and force mains to treatment at the Nine Springs sewage treatment plant operated by the Madison Metropolitan Sewerage District (MMSD). Although the treatment plant itself is quite modern many of the wastewater mains and laterals are up to 50 years old, and some older lines are up to 90 years old. The sewer lines are constructed of a variety of materials, including (ranging in general from older to newer construction) vitrified clay, cast iron, ductile iron, reinforced Figure 11. Cross sections showing the possible locations of a sewer relative to the water table. A: gravity-drain sewer below water table; B: gravity-drain sewer above water table; C: force main above water table; D: force main below water table. H1 and H2 represent hydraulic head inside and outside the sewer. Arrows show directions of potential sewer leakage. concrete, asbestos cement, and PVC plastic. The sewers are generally placed in trenches 10 to 20 feet deep on top of a gravel bed backfilled with native material. The City of Madison manages nearly 800 miles of sewer lines that extend along each city street. The Madison Metropolitan Sewage District (MMSD) manages larger collector and connector sewers that extend from various locations in the city to the MMSD regional sewage treatment plant. The MMSD sewers include about 93 miles of gravityflow lines and 30 miles of force mains. Calculated source volumes of viral contaminants Calculation of potential mixing between groundwater and sanitary sewer leakage suggests that the sewer leakage is a likely source of viral contamination of groundwater. Very little sewage is needed to produce the concentrations seen in the wells, as would be 30 expected when a gram of feces from an infected person can contain over one trillion infectious viruses. A simple calculation using some results from the study is provided below to illustrate this point. This calculation is based on the amount of water a well pumps during one virus sampling event (around 4 hours of pumping). During a typical sampling event a well produces over one million liters of water (1500 gal/min x 4 hr pumping x 60 min/hr x 3.78 l/gal). We assume that the viral concentration of the pumped water is constant during the four-hour sampling period, and that the viral filter collects a representative sub-sample of this water. Also assuming complete mixing in the aquifer and well bore and that "background" groundwater contains no viruses, we can calculate the volume of sewage needed to produce the observed concentrations in the wells. Table 5 summarizes viral and tritium concentrations observed in this study. Table 5. Summary of virus and tritium observations Water source Madison lakes Madison sewage Madison wells Virus concentration, gc/l 0 - 27 49,000 - 2,100,000 0 - 6.2 Tritium content, TU 8.5-9.2 0-6 (assumed) 0-10 Assuming that all viruses originate in the source water, the basic conservation of mass equation is: Vs x Cs = Vgw x Cgw Where Vs = volume of sewer leakage, Cs = concentration of viruses in sewer leakage, Vgw = volume of groundwater, and Cgw = virus concentration in groundwater. If the source volume is the only unknown, the equation becomes: Vs = Vgw x Cgw / Cs For the minimum sewage concentration (49,000 gc/l) and maximum well concentration (6.2 gc/l): Vs = 1.36x106 l x 6.2 gc/l / 49,000 gc/l = 172 l For the maximum sewage concentration (2,100,000 gc/l) we have : Vs = 1.36x106 l x 6.2 gc/l / 2,100,000 gc/l = 4 l 31 Accordingly, about 4 to 170 liters (1 to 44 gallons) of sewage leakage into the recharge area could produce the maximum virus concentration observed in well water during the 4-hour long sampling event when the well was pumping. This analysis is only meant to be illustrative does not include the true contaminant transport processes. This disclaimer notwithstanding, such a minor sewer leak does not seem unreasonable given length of sewer pipe in a typical Madison well capture zone, The small amount of virus-contaminated leakage required to produce the viral concentrations seen in the well samples explains why a well (such as wells 12 and 30) can be virus positive but not contain detectable tritium. Ambient tritium concentrations in surface water are on the order of 10 TU. Sewage, which is mostly derived from locally-pumped groundwater, is assumed to have tritium concentrations in the same range as the wells (0-6 TU). Mixing these small amounts of tritiated water with "old" groundwater (assumed to contain less than 0.5 TU) would not raise the tritium content above the laboratory detection limit of 0.8 TU. For the example above, mixing 172 liters of virus-laden sewage with one million liters of uncontaminated water produces detectable virus concentrations in the water. However, mixing the same 172 liters of water having a tritium content of 10 TU with one million liters of tritium-free water would produce a mixed concentration of about 0.002 TU, far below the laboratory detection limit. The dilution calculations above demonstrate that human viruses have the potential to be used as very sensitive groundwater tracers. They possess several characteristics necessary for good tracer performance. First, when present, they are detectable over several orders of magnitude, from 1 gc/l to millions of gc/l. Second, they are extremely mobile. Third, virus speciation allows correlation of specific viral serotypes which vary through time, giving a temporal measure to tracer experiments. Finally, there has been much progress in reducing the time and cost of analyses, bringing such a tracer into the reach of more studies. However, they can only be used as tracers where there is a virus source, which limits their use to urban areas or areas affected by sewage treatment systems. Additional investigation of the use of viral tracers in groundwater study should be the focus of future research. 32 Conclusions and Recommendations Conclusions Human enteric viruses are a common contaminant in water produced by municipal wells in Madison, Wisconsin. Viruses were found in all wells sampled monthly, though not in every sample from every well. The percentage of virus-positive samples ranged from 60% in wells know to have multi-aquifer construction or shallow casings to 18 % in well 30, a new, deep well deeply cased across a regional aquitard. The presence of viruses in wells cased and grouted 200 to 300 feet below a regional aquitard raises disturbing questions about aquifer vulnerability in confined-aquifer settings usually thought to be well-protected from surface contaminants. Although we are unable at this time to elucidate the transport pathway for viruses from the surface to the wells, several lines of evidence suggest that transport is rapid - on the order of months or weeks rather than years. Because they require a human host, these viruses must originate at or just below the land surface. Identical viral serotypes were found in sewage and groundwater, and the mix of viral species varied with time through the project. Moreover, virus detections in wells, and virus concentrations in lakes and sewage varied together through time. This temporal correlation is consistent with relatively rapid transport. The Madison Lakes are probably not the main source of the viruses found in the Madison municipal wells. Lake water contained some but not all of the serotypes found in the wells, and virus levels in lake water are generally low. Furthermore, the 18O/2H signature of water produced by most Madison wells is not consistent with a significant lake water component of recharge. The most likely source of the viruses in the wells is the leakage of untreated sewage from the Madison sewer system. Untreated sewage sampled at the Madison sewage treatment plant contains virus concentrations several orders of magnitude higher than concentrations observed in wells or lakes. Review of sewer construction and location data, the shear total length of city sewers (hundreds of miles), and the evidence that sewers are not completely water-tight suggests that leakage of sewage to the subsurface environment probably occurs in at least some parts of Madison. Given the high concentrations (millions of genomic copies per liter) of viruses in sewage, it would take very little sewage to produce the virus concentrations observed in the wells. Human enteric viruses might be excellent tracers of recent groundwater. They have the desirable tracer characteristics of detectability over several orders of magnitude, high mobility, short analytic times and relatively reasonable cost, and are time-specific due to constantly changing serotypes. Although the presence of detectable tritium in a well is 33 almost always and indicator or recent recharge to the well, the absence of tritium (at a detection limit of 0.8 TU) does not necessarily indicate that the well will be virus-free. In fact detection of viruses many be a far more sensitive indicator than tritium of a proportion of "young" groundwater in a well if the well captures a virus source. Recommendations This study shows that human viruses can be commonly present in groundwater in deep bedrock wells. To protect human health, communities in Wisconsin and elsewhere that use groundwater for a drinking water source should consider using chlorination or other water treatment techniques to deactivate viruses, and work to ensure that these systems are operating correctly. Sampling for viruses requires a time series approach because virus concentrations, and virus species, vary with time in individual wells. Untreated sewage contains very high concentrations of viruses and should be considered a source of groundwater contamination. Wisconsin communities should evaluate sewer infrastructure to determine the potential for leakage of untreated sewage to the subsurface. For example, communities might wish to prioritize sewer repair or replacement within the contributing areas of municipal wells. Research on the impacts of sewers on groundwater quality should be encouraged. Human enteric viruses represent a potentially powerful new tracing tool for hydrogeologic studies. Both fundamental (theoretical and column studies) and applied (field evaluations) research on the use and effectiveness of viruses as tracers should be undertaken. 34 References Abbaszadegan, M., M. LeChevallier, and C. Gerba. 2003. Occurrence of viruses in US groundwaters. Journal of American Water Works Association. 95:107-120. Bishop, P. K., et al. (1998), Impacts of sewers on groundwater quality, Journal of the Chartered Institution of Water and Environmental Management, 12, 216-223. Borchardt MA, Bradbury KR, Gotkowitz MB, Cherry JA, and Parker BL. 2007a. Human enteric viruses in groundwater from a confined bedrock aquifer. Environmental Science and Technology 41:6606-6612. Borchardt MA, Spencer SK, Harrington GW, Bertz PD, Volenec MJ, Kieke BA Jr., Lambertini E, and Loge FJ. 2007b. Inputs of viruses into municipal drinking water: contaminated groundwater versus distribution system intrusions. Poster presented at the American Society for Microbiology, Toronto, Ontario, May 2125, 2007. Borchardt MA, Haas NL, Hunt RJ. 2004. Vulnerability of municipal wells in La Crosse, Wisconsin, to enteric virus contamination from surface water contributions. Applied and Environmental Microbiology 70: 5937-5946. Borchardt MA, Bertz PD, Spencer SK, Battigelli DA. 2003 Incidence of enteric viruses in groundwater from household wells in Wisconsin. Applied and Environmental Microbiology 69:1172-1180. Bradbury, K. R., S. K. Swanson, J. T. Krohelski, and A. K. Fritz. 1999. Hydrogeology of Dane County, Wisconsin. open-file report Open-File Report 1999-04, Wisconsin Geological and Natural History Survey, Madison, WI. Clark, I., and P. Fritz, 1997. Environmental Isotopes in Hydrology, CRC Press, Boca Raton, FL. Carducci A., Verani M., Battistini R., Pizzi F., Rovini E., Andreoli E., and Casini B. 2006. Epidemiological surveillance of human enteric viruses by monitoring of different environmental matrices. Water Sci. technol. 54:239-244. Fout, G.S., B.C. Martinson, M.W.N. Moyer, and D.R. Dahling. 2003. A multiplex reverse transcription-PCR method for detection of human enteric viruses in groundwater. Appl. Environ. Microbiol. 69:3158-3164. Hunt, R.A., T.B. Coplen, N.L. Haas, D.A. Saad, and M.A. Borchardt. 2005. Investigating surface water-well interaction using stable isotope ratios of water. Journal of Hydrology, 302 (2005). P 154-172. Hunt, R.J., and J.J. Steuer. 2000. Simulation of the recharge area for Frederick Springs, Dane County, Wisconsin. USGS Water-Resources Investigations Report 00-4172. John, D.E. and J.B. Rose. 2005. Review of factors affecting microbial survival in groundwater. Environ. Sci. Tech. 39:7345-7356. Kammerer, P.A. 1981. Ground-water-quality atlas of Wisconsin. Information Circular 39. Wisconsin Geological and Natural History Survey. Karklins, S. 1996. Groundwater sampling field manual. Wisconsin Department of Natural Resources. Publication PUBL-DG-038-96, 70 p. Keswick, B.H. and C.P. Gerba. 1980. Viruses in groundwater. Environ. Sci. Tech. 14:1290-1297. Kurtz, A. M., J.M. Bahr, Q.J. Carpenter, and R. J. Hunt. 2007. The importance of 35 subsurface geology for water source and vegetation communities in Cherokee Marsh, Wisconsin. Wetlands, 27: 189-202. Lambertini E., Spencer S.K., Bertz P.D., Loge F.J., Kieke B.A. and Borchardt M.A. 2008. Concentration of enteroviruses, adenoviruses, and noroviruses from drinking water with glass wool filters. Applied and Environmental Microbiology 74:29902996. Leif Wolf, I. H., Matthias Eiswirthdagger, Heinz H?tzl, (2004), Impact of Leaky Sewers on Groundwater Quality, Acta hydrochimica et hydrobiologica, 32, 361-373. Madison Metropolitan Sewerage District (MMSD), 2008. Metro Interceptor (newsletter). Summer/Fall 2008 issue. p 4. National Primary Drinking Water Regulations - Ground Water Rule, Final Rule. Federal Register, 71, 224 (21 November 2006) p. 67427 - 65660. Osenbr?ck, K., et al. (2007), Sources and transport of selected organic micropollutants in urban groundwater underlying the city of Halle (Saale), Germany, Water Research, 41, 3259-3270. Robertson JB, Edberg SC. 1997. Natural protection of spring and well drinking water against surface microbial contamination. 1. Hydrogeological parameters. Crit Rev Microbiol 23:143-178, 1997. Rutsch, M., et al. (2008), Towards a better understanding of sewer exfiltration, Water Research, 42, 2385-2394. Schijven, J.F., J.H.C. M?lschlegel, S.M. Hassanizadeh, P.F.M. Teunis and A.M. de Roda Husman. 2006. Determination of protection zones for Dutch groundwater wells against virus contamination - uncertainty and sensitivity analysis. J Water Health 4:297-312. Sedmak, G., D. Bina, and J. MacDonald. 2003. Assessment of an enterovirus sewage surveillance system by comparison of clinical isolates with sewage isolates from Milwaukee, Wisconsin, collected August 1994 to December 2002. Appl. Environ. Microbiol. 69:7181-7187. Sedmak, G., Bina D., MacDonald J., and Couillard L. 2005. Nine-year study of the occurrence of culturable viruses in source water for two drinking water treatment plants and the influent and effluent of a wastewater treatment plant in Milwaukee, Wisconsin (August 1994 through July 2003). Appl. Environ. Microbiol. 71:10421050. Swanson, S.K. J.M Bahr, and K.W. Potter. 2006. A local meteoric water line for Madison, Wisconsin. Wisconsin Geological and Natural History Survey, Openfile Report 2006-01. 5 p. Trowsdale, S. A., and D. N. Lerner (2007), A modelling approach to determine the origin of urban ground water, Journal of Contaminant Hydrology: Issues in urban hydrology: The emerging field of urban contaminant hydrology, 91, 171-183. Yates, M. V., C. P. Gerba, and L. M. Kelly. 1985. Virus persistence in groundwater. Appl. Environ. Microbiol. 49:778-781. 36 Appendices Appendix A: Virus results. Key: MC sample ID = Marshfield Clinic sample ID; Sample ID = field sample ID; type= well, lake, or wastewater; well ID = local number of well or lake name; Total virus concentration = virus concentration in genomic copies per liter; collection date = date of collection; filtration volume = volume of water filtered. MC Sample ID 32117 32427 32587 32605 32621 32637 32838 32865 32944 32970 32118 32540 32115 32457 32567 32608 32622 32639 32842 32876 32945 32971 32002 32425 32568 32588 32607 32618 32631 32831 32878 32948 32969 32072 32937 32972 Sample ID 7-1 7-2 7-4 7-5 7-6 7-7 7-8 7-9 7-10 7-11 8-1 8-2 11-1 11-2 11-3 11-5 11-6 11-7 11-8 11-9 11-10 11-11 12-1 12-2 12-3 12-4 12-5 12-6 12-7 12-8 12-9 12-10 12-11 13-1 13-10 13-11 Total Virus Conc (gc/L) 0.05 0.69 0.11 4.97 0.00 0.00 0.00 0.00 5.48 3.78 0.00 0.30 0.22 0.00 0.15 0.00 0.35 0.02 0.00 0.00 6.15 0.94 0.21 0.30 0.09 1.45 0.00 0.00 0.00 0.00 0.00 1.70 2.91 0.14 0.00 0.00 Collection Date 9/25/2007 10/25/2007 1/4/2008 1/28/2008 2/28/2008 3/26/2008 4/30/2008 5/27/2008 7/8/2008 7/29/2008 9/25/2007 11/2/2007 9/24/2007 10/30/2007 11/28/2007 1/30/2008 2/27/2008 3/27/2008 5/1/2008 6/2/2008 7/8/2008 7/29/2008 9/14/2007 10/24/2007 11/26/2007 1/3/2008 1/29/2008 2/26/2008 3/24/2008 4/28/2008 6/2/2008 7/9/2008 7/28/2008 9/20/2007 10/30/2007 11/28/2007 Filtration volume (l) 1139.4 817.6 556.5 919.9 813.9 829.0 829.0 916.0 950.1 817.6 817.6 832.8 836.6 1237.8 844.1 851.7 768.4 829.0 942.6 859.3 806.3 855.5 1018.3 806.3 1294.6 681.4 836.6 829.0 1449.8 806.3 810.1 878.2 817.6 863.1 874.4 810.1 type well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well ID 7 7 7 7 7 7 7 7 7 7 8 8 11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 13 13 13 37 MC Sample ID 32458 32572 32589 32604 32623 32638 32839 32866 32003 32460 32569 32584 32148 32455 32034 32938 32967 32459 32570 32586 32603 32624 32635 32833 32867 32428 32044 32035 32426 32116 32949 32968 32456 32571 32590 32606 32619 32636 32832 32877 32033 32043 Sample ID 13-2 13-3 13-4 13-5 13-6 13-7 13-8 13-9 16-1 16-2 16-3 16-4 17-1 17-2 19-1 19-10 19-11 19-2 19-3 19-4 19-5 19-6 19-7 19-8 19-9 24-2 27-1 28-1 28-2 30-1 30-10 30-11 30-2 30-3 30-4 30-5 30-6 30-7 30-8 30-9 Wingra-1GW Mendota1-GW type well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well well SW SW well ID 13 13 13 13 13 13 13 13 16 16 16 16 17 17 19 19 19 19 19 19 19 19 19 19 19 24 27 28 28 30 30 30 30 30 30 30 30 30 30 30 Wingra Mendota Total Virus Conc (gc/L) 0.00 0.02 0.00 0.00 0.00 0.00 0.38 0.65 0.07 0.00 0.00 0.00 0.00 0.00 0.11 0.00 0.40 0.82 0.00 0.00 0.00 0.09 0.00 2.83 0.00 0.06 0.00 0.00 0.09 0.00 0.00 0.00 0.00 0.10 0.00 0.00 0.00 0.00 0.00 0.05 1.26 9.05 Collection Date 1/7/2008 1/25/2008 2/27/2008 3/26/2008 4/30/2008 5/27/2008 7/7/2008 7/29/2008 9/14/2007 10/31/2007 11/26/2007 12/19/2007 9/26/2007 10/29/2007 9/18/2007 10/30/2007 11/27/2007 1/2/2008 1/24/2008 2/29/2008 3/24/2008 4/28/2008 5/29/2008 7/7/2008 7/30/2008 10/25/2007 9/19/2007 9/17/2007 10/24/2007 9/24/2007 10/29/2007 11/27/2007 1/9/2008 1/28/2008 2/26/2008 3/25/2008 4/28/2008 6/2/2008 7/9/2008 7/28/2008 9/17/2007 9/19/2007 Filtration volume (l) 829.0 806.3 908.5 280.1 1059.9 794.9 806.3 855.5 787.4 802.5 870.6 813.9 1362.7 802.5 840.4 810.1 972.8 813.9 863.1 878.2 969.1 806.3 806.3 851.7 855.5 836.6 931.2 829.0 836.6 844.1 893.4 1188.6 810.1 1052.3 1067.5 832.8 798.7 904.7 1048.6 923.6 832.8 889.6 38 MC Sample ID 32074 32834 32840 32843 32933 32935 32946 32334 32543 32573 32585 32609 32620 32640 32845 32864 32950 Sample ID Monona1-GW Wingra-8 Monona-8 Mendota8 Wingra10-P Monona10-P Mendota10 MMSW-1a MMSW-2 MMSW-3 MMSW-4 MMSW-5 MMSW-6 MMSW-7 MMSW-8 MMSW-9 MMSD-10 type SW SW SW SW SW SW SW WW WW WW WW WW WW WW WW WW WW well ID Monona Wingra Monona Mendota Wingra Monona Mendota Total Virus Conc (gc/L) 8.91 0.00 0.00 2.31 0.01 3.07 27.60 91569.00 2077558.00 1561945.00 558965.77 640625.00 184734.00 227578.00 48623.00 68482.14 348944.00 Collection Date 9/21/2007 4/29/2008 4/30/2008 5/1/2008 7/7/2008 7/8/2008 7/9/2008 10/15/2007 11/7/2007 11/29/2007 12/19/2007 2/5/2008 2/28/2008 3/31/2008 5/5/2008 5/28/2008 7/15/2008 Filtration volume (l) 821.4 847.9 1150.8 859.3 984.2 984.2 984.2 3.0 4.0 8.0 4.0 4.0 4.0 3.8 4.0 4.0 1.0 39 Appendix B: Geochemical results. Key: Field ID = field sample ID; well or lake = local number of well or lake name; LOD = laboratory limit of detection. Date Collected 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 1/4/2008 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 Field ID Well or Lake Parameter Result 311 74.3 10.4 0.4 42.4 29 0.02 7.42 1.4 6.2 33.9 <1.0 309 75 10.5 669 0.4 43.9 28 0.021 7.78 1.5 6.5 34.2 <1.0 301 67.9 14 0.5 41.1 54 ND 7.63 1.4 8.6 16.6 <1.0 317 Units MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L US/CM MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L LOD 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 7-1 7-1 7-1 7-1 7-1 7-1 7-1 7-1 7-1 7-1 7-1 7-1 7-4 7-4 7-4 7-4 7-4 7-4 7-4 7-4 7-4 7-4 7-4 7-4 7-4 8-1 8-1 8-1 8-1 8-1 8-1 8-1 8-1 8-1 8-1 8-1 8-1 11-1 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 7 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 8 WELL 11 ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS CONDUCTIVITY AT 25C IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 40 Date Collected 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 1/3/2008 9/20/2007 9/20/2007 9/20/2007 9/20/2007 9/20/2007 9/20/2007 9/20/2007 9/20/2007 9/20/2007 Field ID Well or Lake Parameter 11-1 11-1 11-1 11-1 11-1 11-1 11-1 11-1 11-1 11-1 11-1 12-1 12-1 12-1 12-1 12-1 12-1 12-1 12-1 12-1 12-1 12-1 12-1 12-4 12-4 12-4 12-4 12-4 12-4 12-4 12-4 12-4 12-4 12-4 12-4 12-4 13-1 13-1 13-1 13-1 13-1 13-1 13-1 13-1 13-1 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 11 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 12 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS CONDUCTIVITY AT 25C IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS Result 77.5 17.3 ND 48.9 9 1.14 7.44 1.3 15.2 11.4 <1.0 264 57.3 1.07 ND 31.8 1 0.774 7.5 1.2 2.3 9.96 <1.0 263 60.2 1.08 518 ND 32.7 ND 0.734 7.78 1.2 2.4 9.98 <1.0 284 63.4 7.74 ND 38.2 14 1.76 7.55 2 Units MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L US/CM MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L LOD 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 41 Date Collected 9/20/2007 9/20/2007 9/20/2007 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 1/8/2008 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 9/14/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 12/19/2007 9/26/2007 9/26/2007 9/26/2007 Field ID Well or Lake Parameter 13-1 13-1 13-1 13-4 13-4 13-4 13-4 13-4 13-4 13-4 13-4 13-4 13-4 13-4 13-4 13-4 16-1 16-1 16-1 16-1 16-1 16-1 16-1 16-1 16-1 16-1 16-1 16-1 16-4 16-4 16-4 16-4 16-4 16-4 16-4 16-4 16-4 16-4 16-4 16-4 16-4 17-1 17-1 17-1 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 13 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 16 WELL 17 WELL 17 WELL 17 SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS CONDUCTIVITY AT 25C IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS CONDUCTIVITY AT 25C IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS Result 5.5 14.1 <1.0 280 61.4 8.06 580 ND 37.9 12 1.74 7.74 1.7 5 13.5 <1.0 270 63.2 35.8 ND 38.2 ND 3.15 7.48 1.1 13.7 9.9 <1.0 269 67.6 40.2 664 ND 38.8 ND 2.7 7.77 1.1 13.7 18.6 <1.0 274 64.8 33.4 Units MG/L MG/L NTU MG/L MG/L MG/L US/CM MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L US/CM MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L LOD 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 42 Date Collected 9/26/2007 9/26/2007 9/26/2007 9/26/2007 9/26/2007 9/26/2007 9/26/2007 9/26/2007 9/26/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 9/18/2007 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 1/2/2008 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 Field ID Well or Lake Parameter 17-1 17-1 17-1 17-1 17-1 17-1 17-1 17-1 17-1 19-1 19-1 19-1 19-1 19-1 19-1 19-1 19-1 19-1 19-1 19-1 19-1 19-4 19-4 19-4 19-4 19-4 19-4 19-4 19-4 19-4 19-4 19-4 19-4 19-4 27-1 27-1 27-1 27-1 27-1 27-1 27-1 27-1 27-1 27-1 27-1 WELL 17 WELL 17 WELL 17 WELL 17 WELL 17 WELL 17 WELL 17 WELL 17 WELL 17 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 19 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 WELL 27 IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS CONDUCTIVITY AT 25C IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS Result ND 41.8 31 0.014 7.49 1.4 15.2 43.3 <1.0 276 59.6 3.6 0.2 31.9 50 ND 7.58 1.8 3.9 6.89 <1.0 274 62.2 3.7 534 0.2 33.9 52 ND 7.84 1.8 3.9 6.8 <1.0 300 77.6 39 0.1 42.3 35 0.397 7.34 1.6 17.5 41.5 Units MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L US/CM MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L LOD 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 43 Date Collected 9/19/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 9/24/2007 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 1/9/2008 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 Field ID Well or Lake Parameter 27-1 28-1 28-1 28-1 28-1 28-1 28-1 28-1 28-1 28-1 28-1 28-1 28-1 30-1 30-1 30-1 30-1 30-1 30-1 30-1 30-1 30-1 30-1 30-1 30-1 30-4 30-4 30-4 30-4 30-4 30-4 30-4 30-4 30-4 30-4 30-4 30-4 30-4 MENDOTA-1 MENDOTA-1 MENDOTA-1 MENDOTA-1 MENDOTA-1 MENDOTA-1 MENDOTA-1 WELL 27 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 28 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 WELL 30 MENDOTA MENDOTA MENDOTA MENDOTA MENDOTA MENDOTA MENDOTA TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS CONDUCTIVITY AT 25C IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS Result <1.0 266 59.9 0.887 0.2 32.8 24 0.022 7.42 1 2.4 17.8 <1.0 249 58.7 3.75 ND 34.2 3 0.04 7.55 1.7 3.9 16.8 <1.0 252 54.4 2.58 509 0.2 32.3 13 ND 7.81 1.6 3.6 16.8 <1.0 155 26.1 38 ND 30.4 3 0.023 Units NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L US/CM MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L LOD 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 44 Date Collected 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/19/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/21/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 9/17/2007 Field ID MENDOTA-1 MENDOTA-1 MENDOTA-1 MENDOTA-1 MENDOTA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 MONONA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 WINGRA-1 Well or Lake MENDOTA MENDOTA MENDOTA MENDOTA MENDOTA MONONA MONONA MONONA MONONA MONONA MONONA MONONA MONONA MONONA MONONA MONONA MONONA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA WINGRA Parameter PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY ALKALINITY TOTAL CACO3 CALCIUM DISS CHLORIDE DISS IRON DISS MAGNESIUM DISS MANGANESE DISS NITROGEN NO3-N DISS PH LAB POTASSIUM DISS SODIUM DISS SULFATE DISS TURBIDITY Result 8.55 3.3 18.9 20.1 <1.0 152 26.6 48.5 ND 27.7 1 ND 8.88 3 23.7 24.5 <1.0 142 30.4 69.1 ND 27.3 4 0.016 8.61 2.1 34.1 15.8 <1.0 Units SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU MG/L MG/L MG/L MG/L MG/L UG/L MG/L SU MG/L MG/L MG/L NTU LOD 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 2.5 0.1 0.026 0.1 0.1 0.5 0.006 0.1 0.1 0.02 45 Appendix C: Field measurements. Key: well id = local number of well or lake name; Sample id = field sample id; collection date and Marshfield Clinic sample ID provided in Appendix A. well id 7 7 7 7 7 7 7 7 7 7 8 8 11 11 11 11 11 11 11 11 11 11 12 12 12 12 12 12 12 12 12 12 12 13 13 13 13 13 sample id 7-1 7-2 7-4 7-5 7-6 7-7 7-8 7-9 7-10 7-11 8-1 8-2 11-1 11-2 11-3 11-5 11-6 11-7 11-8 11-9 11-10 11-11 12-1 12-2 12-3 12-4 12-5 12-6 12-7 12-8 12-9 12-10 12-11 13-1 13-10 13-11 13-2 13-3 pH 7.03 7.34 7.09 7.37 7.26 7.39 7.26 7.04 7.45 7.26 7.47 7.28 6.79 7.31 7.27 7.34 7.29 7.19 7.4 7.09 7.4 7.26 6.78 7.1 7.27 7.09 7.42 7.31 7.11 6.99 7.16 7.47 7.3 6.99 7.45 7.36 7.4 7.3 Dissolved oxygen (mg/L) 0.2 0.8 0.4 0.4 0.6 0.05 0.5 0.1 NA NA 0.5 0.6 3.5 0.4 3 2 2 3 2 3.5 3.5 3 3.5 3 3 4 2 5 2 2 1.5 3.5 3.5 1.5 2.5 2 2 2.5 Specific conductance (uhos/cm) 791 741 700 712 854 715 697 741 731 682 710 662 850 826 811 805 833 788 821 839 840 815 612 560 516 590 571 545 591 588 597 537 632 621 624 606 630 670 46 well id 13 13 13 13 13 13 16 16 16 16 17 17 19 19 19 19 19 19 19 19 19 19 19 24 27 28 28 30 30 30 30 30 30 30 30 30 30 30 lake lake lake lake lake sample id 13-4 13-5 13-6 13-7 13-8 13-9 16-1 16-2 16-3 16-4 17-1 17-2 19-1 19-10 19-11 19-2 19-3 19-4 19-5 19-6 19-7 19-8 19-9 24-2 27-1 28-1 28-2 30-1 30-10 30-11 30-2 30-3 30-4 30-5 30-6 30-7 30-8 30-9 Mendota 10 Mendota- 8 Mendota-1 Monona- 8 Monona-10 pH 6.97 7.39 7 7.36 7.37 7.43 7.2 7.26 7.36 7.2 7.5 7.76 6.84 7.3 7.19 7.35 7.38 7.42 7.32 7.28 7.44 7.32 7.01 7.52 6.52 6.47 7.14 6.7 7.46 7.29 7.44 7.38 7.07 7.44 7.48 7.45 7.49 7.06 8.43 8.18 8.25 7.82 8.43 Dissolved oxygen (mg/L) 1.5 2 2 1 3 NA 6.6 5 4 3.5 0.8 0.05 1 0.5 0.15 1.5 0 0.2 0.2 0.05 1 0 0.5 0.6 1 3 0.6 0.1 0.05 0.05 0.1 0 0.05 0.8 0.05 0.1 0 0.1 NA 9 6 9 NA Specific conductance (uhos/cm) 704 725 583 595 624 649 708 701 740 732 670 837 589 555 NA 583 538 547 660 538 549 570 596 554 778 NA 600 653 541 603 578 526 523 546 536 532 560 729 NA 546 472 715 NA 47 well id lake lake lake lake sample id Monona-1 Wingra- 8 Wingra-10 Wingra-1 pH 8.77 8.24 8.1 8.23 Dissolved oxygen (mg/L) 8 9 NA 6 Specific conductance (uhos/cm) 574 741 NA 552 48