Grantee Research Project Results
Final Report: Evaluation of Sanitary Sewers as a Source of Pathogen Contamination of Municipal Water Supply Wells
EPA Grant Number: R834869Title: Evaluation of Sanitary Sewers as a Source of Pathogen Contamination of Municipal Water Supply Wells
Investigators: Bradbury, Kenneth R. , Borchardt, Mark , Gotkowitz, Madeline B
Institution: University of Wisconsin - Madison , Marshfield Clinic Research Foundation , Wisconsin Geological and Natural History Survey
EPA Project Officer: Page, Angela
Project Period: June 1, 2011 through May 31, 2013 (Extended to December 31, 2013)
Project Amount: $598,580
RFA: Advancing Public Health Protection through Water Infrastructure Sustainability (2009) RFA Text | Recipients Lists
Research Category: Drinking Water , Water
Objective:
This study investigated the spatial and temporal dimensions of pathogen transport from sanitary sewers to municipal water supply wells. Groundwater monitoring wells were installed at seven sites located in urban and suburban areas. The supply wells and sanitary sewers at the sites spanned a range of ages, materials and condition. Groundwater samples were collected over a one-year period from a total of 16 monitoring wells, six deep municipal wells, and raw wastewater. Viruses were detected in 3.7 percent (17 of 455) of samples. Samples positive for total coliform bacteria occurred more frequently during June through October and were not correlated with virus-positive samples. There was no significant difference in the rate of virus positive samples based on well depth, however shallow wells contained higher virus concentrations than deep wells. The seven study sites had similar rates of virus detections in groundwater samples, regardless of the sanitary sewer age and density, or the site proximity to force mains.
The similarity in the rate and timing of virus-positive samples at the seven field sites indicate that a regional driver common across the study area promotes leakage and transport of enteric viruses from sewers to groundwater. Virus-positive samples were collected during a period of moderate temperatures and precipitation following a prolonged drought, and higher virus concentrations were correlated with precipitation events. The high soil temperatures, low rainfall, and a large depth to the water table during the drought apparently limited virus transport from the presumed source, leaky sanitary sewers. Interestingly, viruses were not detected late in the study, during conditions considered favorable to infiltration and rebound of the water table. This suggests that a high virus inactivation rate in the vadose zone during the drought subsequently reduced pathogen transport to groundwater in the following months. Virus concentrations in wastewater samples were low during months with high groundwater recharge, likely due to dilution from inflow of clear water to the sanitary sewer system. These data demonstrate the high temporal variability associated with the transport of fecal pathogen contamination from leaky sewer systems and suggest it is related to multiple climatic and hydrologic conditions.
This study included an evaluation of transport pathways of pathogenic viruses to deep supply wells, including preferential flow through fractures and multi-aquifer wells. Results from a combination of geophysical, geochemical, and hydraulic testing suggest that bedrock fractures in siliciclastic aquifers are important transport pathways from the surface to the deep aquifer. High transmissivity values measured in fractured intervals facilitate rapid transport of sewer-derived colloidal pathogens such as viruses. Borehole flows measured in five multi-aquifer wells in the study area revealed high velocities (up to 13 meters/minute downward) and associated volumetric flux rates (up to 3,790 liters/minute) through these wells, from the water table to the deep aquifer. However, simulations of flow through multi-aquifer wells completed with a three-dimensional numerical model demonstrate that advective transport through such wells were not sufficiently fast to explain observed virus transport to confined aquifer wells in this setting. This suggests that horizontal fracture pathways in the deep confined aquifer also facilitate the transport of pathogens in the deep subsurface.
Summary/Accomplishments (Outputs/Outcomes):
1. Introduction
Leakage from municipal sanitary sewer systems poses a threat to groundwater quality in urban areas. In communities that rely on untreated groundwater, the transport of fecal pathogens is of particular concern. Although the vulnerability of groundwater to virus contamination is now recognized, the sources of viruses found in deep aquifers have rarely been explicitly investigated. Our research group (Borchardt et al. 2007; Bradbury et al. 2013) detected virus contamination of deep bedrock wells serving the city of Madison, Wisconsin. Results suggested rapid transport from the virus source to the wells. The Madison, Wisconsin wells are typical (in construction, depth, age, and material) of wells now in use in many cities throughout Wisconsin and the U.S. Understanding the sources and transport route(s) of viruses from a near-surface source (sanitary sewers) to deep bedrock wells is critical to assessing the magnitude of this problem, the human health risks, and to developing remedial actions.
This project investigated the sources and transport pathways of human viruses from a presumed source in near-surface sanitary sewers to both shallow groundwater and deeply cased municipal water-supply wells. The research objectives included 1) quantifying the presence of pathogenic viruses in groundwater near urban sewers; 2) establishing correlations between virus presence and sewer characteristics such as age, construction, materials, and overall condition; 3) evaluating the transport pathways of pathogenic viruses to deep supply wells; and 4) using numerical modeling to develop estimates of the amount of sewer exfiltrant reaching groundwater and the probability of contamination of nearby water supply wells.
2. Project Activities and Methods
Project activities included detailed hydrogeologic characterization of seven sites within the study area and an extensive field program to sample shallow and deep groundwater for human enteric viruses in groundwater. This report provides detailed results of the groundwater sampling project and summaries of recently published results.
2.1 Study Setting
The City of Madison, Wisconsin (Figure 1) and surrounding communities rely on groundwater for potable, public water supply. Subsurface infrastructure in Madison includes over 1,350 km of pressurized water mains, first installed in 1882, with about 2,123 km of gravity-drained sanitary sewer mains and laterals, dating from 1899. A regional sewer collection system, including gravity and pressurized mains, directs all wastewater to a single treatment plant (Figure 1). A separate storm water collection system routes runoff to area lakes.
Figure 1. Madison, Wisconsin (left) and study sites (right). Regional pressurized sewer mains (blue lines), gravity mains (brown lines), and waste water treatment plant (WWTP) also shown.
The groundwater system in the region consists of over 200 m of nearly flat-lying sedimentary rocks (Figure 2). Precambrian crystalline rock forms the base of the system and underlies a confined lower aquifer. The Eau Claire aquitard separates the deep aquifer from an upper bedrock aquifer, which in turn is overlain by unlithified material. Pumping from municipal supply wells causes downward vertical gradients from the water table across the aquitard.
Figure 2. Cross section including stratigraphy and typical municipal well construction
2.2 Site Development
Seven sites (Figure 1; Table 1) were developed in the study area. The sites include a total of sixteen monitoring wells and six municipal wells. Five sites are located in urban settings with dense networks of sanitary sewers; these sites were selected based in part on multiple virus detections in the municipal wells during previous sampling (Bradbury et al., 2013). The two sites added for the current study include FB11, located in a newly sewered suburban area, and the Lake Edge site, which provides a sampling location over 1 km distant from an active municipal well. Two monitoring wells at the Lake Edge site (LED and LEVD) were constructed in 2003 for an unrelated study; a third monitoring well (LES) was drilled for this study. The fracture network at the Well 7 site was characterized by Gellasch, et al. (2013).
Table 1. Site and monitoring well construction information. Monitoring wells were constructed of 5-cm diameter PVC; well depth is calculated as the mid-point of the screen and gravel pack; municipal well depth is calculated as the mid-point of the open (non-cased) portion of the borehole.
Site ID | Well ID | Screen Length (m) | Mid-point of screen + sand pack (m) | Formation screened | Distance to municipal well (m) |
---|---|---|---|---|---|
Well 13 | MW 13-S | 3.05 | 22.4 | Tunnel City | 205 |
MW 13-D | 1.52 | 32.2 | Tunnel City | 205 | |
Well 11 | MW 11-S | 3.05 | 13.1 | Tunnel City | 53 |
MW 11-D | 1.50 | 23.5 | Wonewoc | 53 | |
Well 19 | MW 19-S | 4.00 | 15.7 | Tunnel City | 23 |
MW 19-D | 1.50 | 45.7 | Wonewoc | 26 | |
Well 7 | MW 7-S | 4.57 | 12.0 | Tunnel City | 8 |
MW 7-D | 3.05 | 28.7 | Wonewoc | 9 | |
MW 7-VD | 3.05 | 63.6 | Wonewoc | 10 | |
Well 30 | MW 30-S | 4.57 | 18.6 | Tunnel City | 38 |
MW 30-D | 1.52 | 41.0 | Tunnel City | 39 | |
FB 11 | FB 11-S | 1.52 | 31.1 | Tunnel City | 168 |
FB 11-D | 1.52 | 52.6 | Tunnel City | 168 | |
Lake Edge | LES | 3.05 | 16.9 | Tunnel City | 1006 |
LED | 1.52 | 71.3 | Wonewoc | 1006 | |
LEVD | 1.52 | 81.9 | Mt. Simon | 1006 |
2.3 Groundwater Sampling Program
Virus and total coliform samples were collected from sixteen monitoring wells approximately every two weeks between June 2012 and May 2013 (24 samples from each well). A 4-L sample of clarified and settled influent (72-hour composite) was collected every two weeks from the regional Madison Metropolitan Sewerage District (MMSD) Nine Springs waste water treatment plant. The six municipal wells were sampled every month (12 samples from each well). The municipal well and monitoring wells at each site were sampled on the same day. Samples for total coliform and Escherichia coli bacteria were collected aseptically in 100 mL sterile bottles. Viruses were concentrated from groundwater on glass wool filters (Lambertini, et al., 2008). Samples were collected at a rate of approximately 5 L per minute. The average sample volume was 920 L, with a range of 814 to 1,117 L.
Glass wool filters, sewage samples, and samples collected for total coliform and E.coli were processed within 24 to 48 hours at the US Department of Agriculture−Agricultural Research Service (ARS) Laboratory in Marshfield, Wisconsin. Coliform bacteria and E. Coli were evaluated with the Colilert® Quanti-Tray test, which is a most probable number (MPN) method. Results are expressed as MPN per 100 ml. Glass wool filters were eluted with beef extract/glycine and the eluate flocculated and concentrated with polyethylene glycol following Borchardt et al. (2012). Sewage influent samples (4 liters) were concentrated using the same secondary concentration procedure as for the filter eluates. Concentrated samples and sewage influent were stored at −80 °C prior to analysis. Samples were analyzed for seven virus groups including adenovirus A, adenovirus B, adenoviruses C, D, and F, enterovirus, norovirus genogroup I, norovirus genogroup II, and human polyomavirus (HPV). Viruses were detected by real-time quantitative PCR (qPCR) or reversetranscription qPCR (RT-qPCR) Assays were quantitative and in duplicate, using the LightCycler 480 (Roche Diagnostics, Mannheim, Germany) and TaqMan probes. Additional detail on qPCR and RT reaction conditions, primers, probes, standard curve preparation and quality assurance parameters, and the calculations for sample virus concentrations are described by Borchardt, et al., 2012.
Wells were sampled for water and isotope chemistry. Groundwater was collected on a quarterly basis and analyzed at the Water and Environmental Analysis Lab, University of Wisconsin-Stevens Point for major ions, nitrate and chloride. Wells were sampled once for tritium, with analyses performed at the University of Miami Tritium Laboratory at the Rosenstiel School of Marine and Atmospheric Science by enrichment and counting of H2 gas; samples expected to be greater than 3 tritium units (TU) were counted without enrichment. Wells were sampled three times over the course of the study for environmental isotopes of hydrogen. Analyses were performed at the Iowa State University Stable Isotope Lab; deuterium was determined by manganese reduction and oxygen-18 was determined by mass spectrometry on CO2 gas.
2.4 Virus Analysis Quality Control and Assurance Measures
Quality control procedures associated with analysis of viruses in groundwater samples included evaluation of groundwater matrix effects on sample collection methods, equipment disinfection methods, and RT and qPCR inhibition by naturally occurring compounds. The groundwater matrix effect refers to potential impacts of groundwater quality on the recovery efficiency of glass wool filters. Groundwater samples from study wells were spiked with Poliovorus Sabin 3 and recovery efficiency measured as described in Lambertini et al. (2008). Recoveries ranged from 12 to 99% (Table 2). These recoveries are within an acceptable range and suggest that matrix effects on virus recovery were not a significant factor in sample results.
Table 5.
Sample | Poliovirus Sabin 3, % recovery |
---|---|
19-D | 36 |
30-D | 17 |
7-D | 57 |
11-D | 99 |
LE-D | 15 |
FB11 | 12 |
Prior to each use, non-disposable sampling equipment (submersible pumps, discharge hose, and hose fittings) was cleaned in a 50 mg/L chlorine solution with a contact time resulting in a 6.0 log inactivation of viruses, rinsed in a chlorine neutralizing sodium thiosulfate solution, and a final deionized water rinse. To ensure adequate de-chlorination, equipment was tested for the presence of chlorine residual by use of a low-level colorimetric method (CHEMets™) sensitive to less than 0.06 mg/L. Groundwater pumped through decontaminated equipment was also tested to ensure chlorine residual below the detection limit prior to sampling for viruses.
Equipment field blanks were collected on a quarterly basis. A sterilized, phosphatebuffered saline solution was prepared and transported to a field site where a sterilized submersible pump and tubing were used to pump solution through a glass wool filter. The filter was transported with other filters to the ARS lab. Each of four equipment blanks was negative for viruses, suggesting field and equipment-related procedures did not lead to sample contamination. Quality control measures in place at the ARS laboratory for sample processing included negative controls for nucleic acid extraction, PCR amplification, and cell culture procedures. All controls were negative throughout the study.
2.5 Data Compilation and Analysis
Groundwater levels in monitoring wells were recorded throughout the study period. Field personnel measured water levels in each monitoring well prior to sampling. Recording pressure transducers measured water levels at twenty-minute intervals between sampling events. Madison Metropolitan Sewer District provided records of the daily volumetric flow of waste water from three pumping stations. Records of daily precipitation at the Dane County Regional Airport (Truax Field) are maintained by the National Oceanic and Atmospheric Administration National Climatic Data Center. Soil temperatures were recorded at a 15-cm depth in near-by Verona, Wisconsin by the University of Wisconsin-Madison Department of Soil Science.
Characterization of the sanitary sewer system at each field site was accomplished with a combination of Geographic Information System (GIS) analyses and numerical modeling of groundwater flow. The City of Madison Engineering Department, City of Fitchburg Public Works Department and the Sewer District provided information on age, material, location, and depth of sewer networks. The GIS analyses incorporated model-simulated zones of contribution to municipal wells sampled in this study. These zones were delineated by applying a one-year time of travel to results from a steady-state, three-dimensional regional groundwater flow model developed by the U.S. Geological Survey and the Wisconsin Geological and Natural History Survey.
Graphical analyses of data include scatter and bar charts to evaluate the occurrence of viruses in groundwater and potential relationships with well and site characteristics, chemical and biological constituents, and indicators of groundwater quality. Where graphical analysis suggested differences may be present within or between groups of samples, data were evaluated with the nonparametric 9 Kruskal-Wallis test (Helsel and Hirsch, 2002). Total coliform and virus concentrations below their analytic detection levels were assigned a value of zero for this evaluation. Relationships between virus detections in groundwater and viruses in wastewater were evaluated with linear or logistic regression applying a significance level of ≤ 0.10. Where indicated, enteric virus concentrations were summed for samples positive for more than one virus type. A two-week interval was applied to the data to aggregate results by time period. Two weeks was the typical time required to sample all wells in the study. Challenging field conditions resulted in only four wells sampled during the two week period in late December 2012.
Conclusions:
Virus Occurrence in Wells and Sewage
Groundwater was positive for viruses in 17 of 455, or 3.7% of the samples collected, and concentrations ranged from non-detectable to 12.7 gc/L. The mean concentration of virus-positive samples was 1.6 gc/L, with a median of 0.9 gc/L. Of the 16 monitoring wells in the study, 11 (69%) were virus-positive at least once (Table 3). Two of the six municipal wells sampled, 33%, were viruspositive during this study. The four municipal wells with no virus detections in this study were viruspositive in a previous study (Bradbury, et al., 2013). The percent of virus positive samples collected during each 2-week period varied through time, from zero to 21 % (Figure 3). With the exception of one sample collected in June 2012, virus occurrence was limited to the months of September through February.
Figure 3. Percent virus-positive samples during two-week periods.
Symbol label indicates number of groundwater samples collected each period.
Table 3. Pathogen- positive samples by well
Well | Times sampled | Virus positive samples | percent virus positive samples | Total Coliform positive samples | Percent Total Coliform positive samples |
---|---|---|---|---|---|
FB11-S | 24 | 1 | 4 | 9 | 38 |
FB11-D | 24 | 1 | 4 | 8 | 33 |
FB Well 11 | 12 | 1 | 8 | 0 | 0 |
MW11-S | 24 | 1 | 4 | 9 | 38 |
MW11-D | 24 | 1 | 4 | 5 | 21 |
Well 11 | 12 | 0 | 0 | 0 | 0 |
MW13-S | 24 | 2 | 8 | 5 | 21 |
MW13-D | 24 | 1 | 4 | 1 | 4 |
Well 13 | 12 | 1 | 8 | 0 | 0 |
MW19-S | 24 | 0 | 0 | 7 | 29 |
MW19-D | 24 | 1 | 4 | 5 | 21 |
Well 19 | 12 | 1 | 8 | 0 | 0 |
MW30-S | 24 | 0 | 0 | 9 | 38 |
MW30-D | 24 | 2 | 8 | 6 | 25 |
Well 30 | 12 | 0 | 0 | 1 | 8 |
MW7-S | 24 | 0 | 0 | 8 | 33 |
MW7-D | 24 | 1 | 4 | 3 | 13 |
MW7-VD | 24 | 2 | 8 | 3 | 13 |
Well 7 | 12 | 0 | 0 | 0 | 0 |
LE-S | 24 | 2 | 8 | 10 | 42 |
LED | 24 | 0 | 0 | 12 | 50 |
LE-VD1 | 24 | 0 | 0 | 10 | 43 |
1 LE-VD was sampled 24 times for viruses and 23 times for total coliform. |
Of the seven viruses enumerated in this study, five were detected in at least one groundwater sample. All seven were present on multiple occasions in wastewater (Table 4). Adenovirus B and norovirus GI were never detected in groundwater. Their absence is attributed to relatively low concentrations and irregular presence over time in wastewater (discussed below).
Sewage influent contained high concentrations of viruses. All samples were positive, with concentrations ranging from 2.4 x 105 to 3.2 x 106 gc/L, with a mean of 1.2 x 106 and a median of 9.8 x 105. Although each of the seven enumerated viruses was present multiple times throughout the study (Table 4), the total concentration of viruses in wastewater decreased over time (linear regression, p-value < 0.001) (Figure 4). Inflow and infiltration in the sanitary sewers increased from February 2013 through the end of the study (Figure 4), generally coincident with a decrease in wastewater virus concentrations. This suggests that viruses in wastewater are diluted during periods of high infiltration of clear water (groundwater or surface water) into sewers. Presumably, dilution from infiltration of clear water would in turn reduce the concentration of viruses in sewer exfiltration to groundwater, if exfiltration occurs primarily when sewers are at full (or exceeding) capacity. Although low virus concentrations in sewer exfiltration may contribute to the lack of virus-positive wells in March through May 2013, there was no statistically significant relationship identified over the duration of the project in the rate of virus detections as a function of virus concentration in wastewater (Figure 5).
Table 4. Number of groundwater and wastewater samples positive for viruses
virus | groudwater detections n = 455 | waste water detecctions n = 24 |
---|---|---|
Adenovirus A | 6 | 20 |
Adenovirus B | 0 | 9 |
Adenovirus CDF | 5 | 23 |
Tnterovirus | 1 | 17 |
G1 Norovirus | 0 | 17 |
G2 Norovirus | 5 | 23 |
Human Polyoma Virus | 2 | 24 |
Figure 4. Virus concentrations in wastewater and calculated infiltration and inflow to sanitary sewer.
Figure 5. Virus concentration in wastewater and percent of virus-positive groundwater samples.
Bacterial Indicators
Study wells were sampled for total coliform bacteria to investigate potential correlations between the indicator bacteria and the presence of human-specific viruses. Total coliform bacteria were present in 27 % (122 of the 453) of groundwater samples. One of the 122 coliform positive samples, collected on August 16, 2012 from monitoring well MW19D, was also positive for E. coli.,but the well was not virus positive on this date. Twenty-four percent (4 of 17) of the groundwatersamples positive for enteric viruses were also positive for total coliform, each at concentrations close to the detection limit of 1 MPN/ 100 mL. Total coliform positive samples occurred more frequently in June through October, during months with few virus detections (Figure 6). Coliform-positive samples were infrequent from March – May 2013, during an extended period with no virus detections. All wastewater samples were positive for total coliform and E. coli bacteria.
Figure 6. Groundwater samples, coliform detections, and virus detections during two-week sampling periods
The small number of coliform-positive samples also positive for viruses is not surprising. Coliform bacteria are common in soil and surface water, but they are not necessarily an indicator of fecal contamination. Other studies show little to no association between bacterial indicators and viruspositive samples in groundwater or surface water (Lambertini et al., 2011; Jiang et al., 2001). A number of factors likely contribute to the lack of correlation between the coliform bacteria and enteric viruses in water samples, including the smaller size of viruses. Environmental conditions such as soil temperature and moisture also affect bacteria and viruses differently (John and Rose, 2005). An additional challenge in comparing the occurrence of coliform bacteria to viruses in this study is related to sample collection methods. The presence of coliform bacteria was evaluated on a 100 mL groundwater sample whereas a much larger volume of water (approximately 920 L) was tested for viruses.
Virus Occurrence at Study Sites
The rates and locations of virus-positive samples were examined for relationships between pathogen susceptibility and characteristics of wells and sites. When categorized by well depth, the Kruskal-Wallis test indicated that there was no significant difference in the rate of virus occurrence by depth (p-value = 0.97) (Figure 7). The six municipal wells constitute the deepest class of wells, screened at depths of 100 – 200 m. Thus, the test also supports the conclusion that monitoring wells and municipal wells were virus positive at similar rates. Comparison of virus-positive percentages for monitoring wells (3.9%, 15 of 384 samples) with percentages for municipal wells (2.8 %, 2 of 71 samples) supports this conclusion. Regression analysis of log concentration data provides moderate evidence that shallower wells had higher virus concentrations than deep wells (p = 0.07) (Figure 8).
Virus occurrence is similar across the seven study sites. Virus-positive rates ranged from about 3 to 5 % at each site (Table 8; Figure 7). The Kruskal-Wallis test (no significant difference in virus occurrence; p-value = 0.99) supports the conclusion that there is no difference between the sites. Additional evaluations confirmed that virus detection rates were not attributable to specific site characteristics, such as sewer age, density of sewer pipes, and proximity to force mains.
Figure 7. Number of samples and virus detections by well depth (left) and site (right). The number of
wells in each depth interval are shown in parenthese The 100-200 m depth class consists of solely of
the six municipal wells in the study.
Figure 8. Virus concentration and well depth. Note log scale on x-axis. Maximum virus cncentration is
shown for wells with more than one virust-positive sample. Wells with no virus detectins are ploted on
the y-axis.
Relationships Between Virus Occurrence and Hydrologic Conditions
Although the over-all detection rate was low during this study, viruses were present at some point in time in 59% (13 of 22) of all sampled wells, and at 69% (11 of 16) of monitoring wells. These data exhibit a coincidence in the timing of virus presence, and in the timing of highest concentrations, 15 in wells (Figures 3 and 5). The similarity in the rate of detections and in the timing of virus occurrence at the seven sites supports the hypothesis offered by Bradbury, et al. (2013) that a regional driver, common to all of the study sites, promotes leakage and transport of enteric viruses in the subsurface.
Table 8. Virus occurrence in groundwater at seven sites
site | Percent virus positive samples | Virus positive samples | Number of samples | Sanitary sewer density in 1-year capture zone, m/m3 |
---|---|---|---|---|
Lae Edge | 2.8 | 2 | 72 | 1.9E-02 |
11 | 3.3 | 2 | 60 | 1.8E-02 |
13 | 5.1 | 3 | 59 | 1.6E-02 |
19 | 3.3 | 2 | 60 | 5.0E-03 |
30 | 3.3 | 2 | 60 | 8.0E-03 |
7 | 3.6 | 3 | 84 | 2.2E-03 |
FB111 | 5.0 | 3 | 60 | 5.0E-03 |
1 FB11 site is down-gradient of homes wiht on-site septic systems. |
Virus detections in groundwater were apparently affected by several hydrologic and climatic factors, including precipitation, infiltration, depth to the water table, and soil temperature. The study began during an extended drought, with little rainfall in May through mid-July 2012 (Figure 9). Drought conditions led to a large decline in the water table elevation and a lack of infiltrating flow through the unsaturated zone, effectively extending the residence time of any sewage exfiltration in the vadose zone. Increased virus sorption and inactivation during unsaturated transport compared to saturated conditions is well-documented (Lance and Gerba, 1984; Powelson, et al., 1990; Wan, et al., 1994). Several mechanisms have been postulated to explain enhanced retention of colloids in the vadose zone, including increased sorption at the solid-water interface or inactivation at the air-water interface (Chu, et al 2001; Denovio, et al., 2004). As discussed previously, exposure to soil temperatures on the order of 30°C would not have been conducive to virus survival (Hurst, et al.,1980). Wastewater samples contained elevated concentrations of viruses during this period, but less than 2% (1 of 54) of groundwater samples collected under these conditions were virus positive.
The virus-positive rate of groundwater samples increased to 7.5% (16 of 212) from mid-September through February. This period coincided with a marked decrease in soil temperatures, recovery of the water table, and some limited precipitation. Inflow to sanitary sewers remained low. The final six rounds of virus sampling occurred during a period of frequent precipitation and large water table rise with periods of high inflow and infiltration to the sanitary sewer. However, no groundwater samples were virus-positive during these months. We hypothesize that dilution of raw wastewater by inflow likely reduced virus concentrations in sewer exfiltration during this period. When examined over the year-long study period, regression analyses suggest a relationship between precipitation and virus concentrations. There is strong evidence of a positive relationship between log virus concentration and precipitation events six days prior to sample collection (p-value =0.0002).
Figure 9. Virus detections in wells, daily precipitation, soil temperature, virus concentration in wastewater, and infiltration and inflow to sewer. Groundwater levels from LES at Lake Edge site. Labels above wells symbol indicate the number of wells sampled during each event.
Discussion
Research objectives of this project included a field study to quantify the presence of pathogenic viruses in groundwater near urban sewers and to establish correlations between virus 17presence and sewer characteristics such as age, construction, materials, and overall condition. The relatively low virus detection rate during the study was similar at all sites, suggesting that under the conditions experienced in 2012, virus transport was not affected by site-specific conditions.
Results from this study support the conclusion that short term climatic conditions affect virus transport from sanitary sewers to the groundwater system. Figure 10 illustrates the contrast between hydrologic conditions during this study and those during a previous, related study with much higher virus detection rates (Bradbury, et al., 2013). The first study took place during a wet period, with greater magnitude and frequency of precipitation and sewer infiltration and inflows, lower summer soil temperatures, and higher water table conditions. Comparison of the two studies indicates that temporal variability in virus transport in this setting is a result of short term changes in climate.
Figure 10. Hydrologic conditions in 2007-2009 compared to 2012-2013.Virus concentrations in wastewater, infiltration and inflow to sewer; virus detection rates in wells and depth to water table; daily precipitation and soil temperature.
These results suggest that more than one temporally variable physical process affects sewer exfiltration and resulting presence of enteric viruses in near-by groundwater. Multivariate statistical analyses of the data presented in this report and data from the 2008 study (Bradbury, et al., 2013) may offer additional insight into virus transport to the water table.
An additional objective of this study was evaluation of transport pathways of pathogenic viruses to deep supply wells. As reported by Gellasch, et al. (2013), one such pathway evaluated during this project included flow through fractures. The study area is underlain in large part by relatively permeable sandstone aquifer, with some interbedded dolomitic layers. A combination of geophysical, geochemical, and hydraulic testing carried out at the Well 7 site suggested that bedrock fractures are important transport pathways from the surface to the deep aquifer. At this site, fractured intervals have transmissivity values several orders of magnitude higher than nonfractured intervals. With respect to rapid transport of colloidal contaminants, such as viruses, the high transmissivity values of individual fractures make them a likely preferential flow pathway to wells. Results suggest that fractures have an important role in the transport of sewer-derived wastewater contaminants in siliciclastic aquifers. The hydrogeologic susceptibility to viral pathogens in this setting is a significant finding, given that fracture pathways are typically associated with granitic or karstified aquifers.
A second hypothesis developed from the study of the fracture network at the Well 7 site involves the propagation of reverse water-level fluctuations, or RWFs (Gellasch, et al., 2014). RWFs can occur when pumping from confined aquifer causes piezometric head in adjacent aquifers or aquitards to temporarily rise prior to falling. Such responses were detected in monitoring wells at the Well 7 site in response to the start-up of the pump in municipal well 7. A RWF in a well is normally attributed to poroelastic coupling between the solid and fluid components in an aquifer system. In addition to revealing classical pumping-induced poroelastic RWFs, the pressure transducers installed in monitoring wells located at varying depths and distances from the public supply well suggest that the RWFs propagate rapidly through fractures to influence wells hundreds of meters from the pumping well. The rate and cycling frequency of pumping is an important factor in the magnitude of RWFs. The pattern of RWF propagation can be used to better define fracture connectivity in an aquifer system. Analysis of the Well 7 data indicate that rapid, cyclic head changes due to RWFs may also serve as a mechanism for contaminant transport.
Multi-aquifer wells (Figure 2) are also a potential pathway for wastewater contaminants to reach public supply wells in urban areas. During the course of this project, we measured vertical borehole flow in five such wells in the study area (Gotkowitz, et al., 2013; Gotkowitz, 2015). The well 13 site, located in the regional discharge area, showed no measureable flow across the aquitard. Velocity measurements from the four multi-aquifer wells located in regional recharge areas ranged from 1 to 13 meters/minute downward, from the upper aquifer to the deep, Mt Simon formation. Volumetric flux, which depends on well bore diameter, ranged from 190 to over 3,790 liters/minute. Numerical modeling, completed with a three-dimensional MODFLOW groundwater model of the Madison area demonstrated that measured flow rates are reasonable in these wells. However, the simulations demonstrated that while the multi-aquifer wells provide rapid transport from the water table to the deep aquifer, the subsequent simulated travel times (based on porous media flow through the aquifer) to near-by wells completed in the confined aquifer were not sufficiently fast to explain observed virus transport in this setting. Additional work on this topic will include numerical simulations of advective transport through multiaquifer wells and discrete fracture networks.
References:
Bradbury KR, et al., Source and Transport of Human Enteric Viruses in Deep Municipal Water Supply Wells. Environmental Science & Technology 2013;47(9):4096-4103.
Borchardt MA, Bradbury KR, Gotkowitz MB, Cherry JA, Parker BL. Human enteric viruses in groundwater from a confined bedrock aquifer. Environmental Science & Technology 2007;41(18):6606-6612.
Hurst CJ, Gerba CP, Cech I. Effects of environmental variables and soil characteristics on virus survival in soil. Applied and Environmental Microbiology 1980;40(6):1067-1079.
John DE, Rose JB. Review of Factors Affecting Microbial Survival in Groundwater. Environmental Science & Technology 2005;39(19):7345-7356.
Lambertini E, et al. Concentration of enteroviruses, adenoviruses, and noroviruses from drinking water by use of glass wool filters. Applied and Environmental Microbiology 2008;74(10):2990-2996.
Borchardt MA, Spencer SK, Kieke BA, Lambertini E, Loge FJ. Viruses in nondisinfected drinking water from municipal wells and community incidence of acute gastrointestinal illness. Environmental Health Perspectives 2012;120(9):1272-1279.
Lambertini E, Spencer SK, Kieke Jr. BA, Loge FJ, Borchardt MA. Virus contamination from operation and maintenance events in small drinking water distribution systems. Journal of Water and Health 2011;9(4):799-812.
Jiang S, Noble R, Chu W. Human Adenoviruses and Coliphages in Urban Runoff- Impacted Coastal Waters of Southern California. Applied and Environmental Microbiology 2001;67:179-184.
Powelson DK, Simpson JR, Gerba CP. Virus Transport and Survival in Saturated and Unsaturated Flow through Soil Columns. Journal of Environmental Quality 1990;19(3):396-401.
Wan JM, Wilson JL, Kieft TL. Influence of the Gas-Water Interface on Transport of Microorganisms through Unsaturated Porous-Media. Applied and Environmental Microbiology 1994;60(2):509-516.
Lance JC, Gerba CP. Virus Movement in Soil During Saturated and Unsaturated Flow. Applied and Environmental Microbiology 1984;47(2):335-337.
Chu Y, Jin Y, Flury M, Yates MV. Mechanisms of virus removal during transport in unsaturated porous media. Water Resources Research 2001;37(2):253-263.
DeNovio NM, Saiers JE, Ryan JN. Colloid movement in unsaturated porous media: Recent advances and future directions. Vadose Zone Journal 2004;3(2):338-351.
Journal Articles on this Report : 1 Displayed | Download in RIS Format
Other project views: | All 6 publications | 3 publications in selected types | All 3 journal articles |
---|
Type | Citation | ||
---|---|---|---|
|
Gotkowitz M, Bradbury K, Borchardt M, Zhu J, Spencer S. Effects of Climate and Sewer Condition on Virus Transport to Groundwater. ENVIRONMENTAL SCIENCE & TECHNOLOGY 2016;50(16):8497-8504. |
R834869 (Final) |
Exit Exit |
Progress and Final Reports:
Original AbstractThe perspectives, information and conclusions conveyed in research project abstracts, progress reports, final reports, journal abstracts and journal publications convey the viewpoints of the principal investigator and may not represent the views and policies of ORD and EPA. Conclusions drawn by the principal investigators have not been reviewed by the Agency.