Ambient Ground Water Monitoring – 2005 Sampling Summary and

Ambient Ground Water Monitoring – 2005 Sampling Summary and 2006 Sampling Plan Project Description and Objectives Summary of 2005 Field Activities 2.1. Sampling 2.2. Drilling 2.3. Installation of Water Level Recorders 2.4. Data Analysis 3. Results 3.1. Condition Monitoring 3.2. Trend Monitoring 3.3. Changing Land Use Monitoring 4. Monitoring Plan for 2006 4.1. Sampling 4.2. Data Management Appendix 1 - Collection and handling of samples for analysis of pfoa, pfos and other fluorinated compounds Appendix 2 – Land use assessment for monitoring wells and domestic wells sampling by the MPCA’s ambient ground water monitoring program 1. Project Description and Objectives The objectives of the Minnesota Pollution Control Agency’s (MPCA) ambient ground water monitoring activity are to assess impacts of non-agricultural chemicals on the condition of ground water in Minnesota, evaluate trends in ambient ground water quality, and identify factors that affect ambient ground water quality. Our monitoring network includes three components. The first component, condition monitoring, includes annual, random sampling of domestic and monitoring wells. All domestic wells exist in sensitive hydrogeologic settings. These settings include areas with surficial sand aquifers or where the Prairie du Chien, Jordan, or Galena formations are the first bedrock encountered (see Figure 1). Monitoring wells occur in coarse-textured soils. Sampling occurs in non-agricultural areas. Analytes include chloride, nitrate, and volatile organic compounds (VOCs). At some monitoring wells we may also collect pesticide samples for the Minnesota Department of Agriculture (MDA), samples for emerging ground water contaminants, or samples to support other MPCA water programs. The second monitoring component consists of long-term trend monitoring. This includes wells from condition monitoring that have a nitrate concentration exceeding 3.0 mg/L, a VOC detection, or a chloride concentration exceeding 150 mg/L. These wells will be sampled annually for five years, then every other year. Sampling in an individual well occurs only for the chemical(s) of concern. The third component consists of monitoring ground water quality in urban areas that have undergone a recent transition in land use. Examples of transition include conversion of non-urban to urban land, Brownfield development, conversion to municipal services, and urban in-growth. 1. 2. wq-gw2-04 Surficial sand aquifers Bedrock aquifers Figure 1: Location of potential sampling areas based on geology. Bedrock aquifers include the Prairie du Chien, Galena and Jordan aquifers. 2. Summary of 2005 Field Activities 2.1. Sampling In 2005 we sampled 113 wells as part of our condition monitoring, 37 wells as part of our trend monitoring, and 11 wells as part of our changing land use monitoring. Table 1 summarizes the breakdown of these wells and includes wells sampled in 2004. In addition to samples associated with routine sampling, we collected 20 samples for pesticides and 17 samples for perfluorooctane sulfonate (PFOS) and perfluoro-octanoate (PFOA) in 2005. Figure 2 illustrates 2004 and 2005 sampling locations. Figure 3 illustrates the location of trend monitoring samples collected in 2005, changing land use wells, wells sampled for pesticides, and wells sampled for pfos/pfoa. Field sampling methods are described in Data Analysis Protocol for the Ground Water Monitoring and Assessment Program (GWMAP) – Version 1.1, February, 1998 ( http://www.pca.state.mn.us/water/groundwater/gwmap/fieldguidance.pdf ). Appendix 1 describes additional sampling procedures for pfos and pfoa. At each well, following stabilization, we collected field measurements of dissolved oxygen, oxidation-reduction potential, temperature, pH, specific conductance, alkalinity, chloride, reduced iron, and nitrate. We then collected samples for laboratory analysis. Pesticide samples were delivered to the Minnesota Department of Agriculture laboratory for analysis. Samples for chloride, nitrate, VOCs, pfos, and pfoa were delivered to the Minnesota Department of Health Environmental Laboratory. Analysis wq-gw2-04 methods included SM4500-Cl-E for chloride, SM4500-NO3-F for nitrate, MDH 498 for VOCs, and MDH 553 for pfos/pfoa. Wells sampled - 2005 Wells sampled – 2004 Network Monitoring Domestic Monitoring Domestic Condition monitoring Cl, NO3, VOCs 39 74 32 58 Pfos/Pfoa 17 0 0 0 Urban Pesticides 20 0 0 0 Trend monitoring Nitrate 13 15 0 0 Chloride 7 0 0 0 VOCs 9 7 0 0 Changing land use 11 0 4 0 Table 1: Summary of 2004 and 2005 sampling. Some wells for trend sampling may fall into more than one category. Monitoring Domestic 2004 2005 Figure 2: Sampling locations for condition monitoring, by year and well type. wq-gw2-04 . . ! Monitoring . . . . ... .! ! .. .. VOC ! . . . !. ! . .!! .. . .. .. ! !! . ! .. . !.. ! . .! . . ! .. . !! . . ! . ! . ! . ! . ..! . ! . ! .. !! # ## # # )) # #))) # # )# )) )# ) )) ) #) PFC Pesticide Changing Land Use Nitrate ) Domestic VOC . . ! Nitrate ) Figure 3: Location of trend wells for VOCs and nitrate, wells sampled for PFCs (pfos and pfoa) or pesticides, and wells sampled as part of the changing land use network. 2.2. Drilling In May we drilled a shallow monitoring well at Benilde-St. Margaret’s High School in St. Louis Park. The monitoring well fills in a gap in our shallow monitoring network and was completed in cooperation with Benilde-St. Margaret’s, who will use the well to monitor ground water levels under different land uses. In August, we drilled two shallow monitoring wells in Linwood Township, Anoka County. One well was completed in an area transitioning from pasture to nonsewered residential land use. The second well was completed in a nearby pasture area. The two wells form a pair, with the well completed in the pasture area representing the control. In September we completed several geoprobes in a row crop agriculture field near Farmington in Dakota County. The area is transitioning to commercial development. The temporary wells helped define ground water flow direction and provide baseline water chemistry for the area. Permanent monitoring wells will be completed once the development is finished. 2.3. Installation of Water Level Recorders We installed six permanent water level recorders in shallow monitoring wells. The recorders are programmed to take daily water level and temperature measurements. Section 4 of this report describes the use of water level recorders in greater detail. 2.4. Data Analysis We used the Kruskal-Wallis statistical test to compare concentrations of chloride and nitrate. Grouping variables included year (2004 versus 2005), type of well (monitoring versus domestic), and dominant land use. Appendix B includes a short summary of our method for classifying land use. All tests were conducted at a significance level of 0.05. wq-gw2-04 3. Results 3.1. Condition Monitoring Table 2 summarizes 2005 sampling results from our condition monitoring network. Nitrate and chloride concentrations did not differ between 2004 and 2005 sampling years for either monitoring (chloride p = 0.964; nitrate p = 0.692) or domestic wells (chloride p = 0.273; nitrate p = 0.365). For both 2004 and 2005, nitrate and chloride concentrations were significantly higher in monitoring wells compared to domestic wells. Median (mg/L) % exceeding standard Sampling Monitoring Domestic pMonitoring Domestic Chemical year wells wells value wells wells Nitrate 2.40 a < 0.05 b 0.003 15.79 1.54 2005 Chloride 62 a 8.4 b < 0.001 22.58 0.00 Nitrate 2.40 a 0.38 b 0.014 3.57 0.00 2004 Chloride 39 a 11 b 0.005 34.78 0.00 Table 2: 2005 summary results from condition monitoring wells. Different letters within a row indicate median concentrations that differed at a significance of 0.05. Table 3 summarizes results by land use. In general, nitrate concentrations were highest under agriculture and residential land use. The analysis, however, is limited at this time by small sample sizes for many of the land uses and only partial descriptions of land use. In coming years we expect to define land use more specifically, such as whether an area contains municipal sewers or not, assessing effects of land use history, and calculating percentages of different land use around each sampling point. When 2004 and 2005 data were combined, we found significantly higher concentrations of chloride in commercial and industrial settings than under other land uses. Using combined data, we also observed significantly higher concentrations of nitrate in agricultural and some residential land use settings. Monitoring wells having green space as the dominant land use showed lower concentrations of chloride and nitrate compared to other land uses. Table 4 summarizes VOC detections in monitoring and domestic wells for 2004 and 2005. The detection frequency in monitoring wells has been about 30 percent over the first two years of the study, compared to about 15 percent in domestic wells. There have been three exceedances of drinking water standards. Solvents and fuel oils account for most detections in monitoring wells, compared to trihalomethanes and solvents in domestic wells. Figure 4 illustrates the location of wells with VOC detections. Detection of VOCs appears to be greatest in the core areas of the Twin Cities and Rochester. wq-gw2-04 Monitoring wells 2004 Chloride (mg/L) 11 60 22 110 98 61 7 Nitrate (mg/L) 11.0 2.50 2.15 0.26 3.30 3.90 3.80 2004 Chloride (mg/L) 7.5 10 7.0 71 2.0 11 23 2004 Nitrate (mg/L) 3.00 a 4.66 a < 0.05 b 0.37 b < 0.05 b < 0.05 b < 0.05 b n 3 13 6 5 1 6 1 n 3 13 6 5 1 6 1 2005 Chloride (mg/L) 160 30 151 58 1081 Nitrate (mg/L) 1.21 b 0.18 b 1.50 b 5.75 a 1.21 b 2005 Chloride (mg/L) 14 18 72 6 27 22 2005 Nitrate (mg/L) 1.30 3.00 0.51 7.70 < 0.05 3.66 n 7 12 2 14 2 n 8 14 4 16 2 Agriculture Commercial Green Space Industrial Mixed Urban Residential Transportation Agriculture Commercial Green Space Industrial Mixed Urban Residential Transportation Domestic wells Agriculture Commercial Green Space Industrial Mixed Urban Residential Water n 16 2 36 1 1 35 8 n 15 2 33 1 1 35 7 n 7 11 1 1 14 2 n 7 11 1 1 14 2 Agriculture Commercial Green Space Industrial Mixed Urban Residential Water Table 3: Median concentrations of chloride and nitrate, by land use. Different letters within a column and for a well type (monitoring or domestic) indicate median concentrations that differed at a significance of 0.05. Land uses with sample sizes of two or less were excluded from the statistical analysis. wq-gw2-04 Year Monitoring Domestic % wells with VOC detected 2004 28.1 15.5 2005 35.9 14.9 % Exceeding HRL 2004 0 1 2005 2 0 No. chemicals found 2004 36 21 2005 56 27 Breakdown of detections (2204-2005 combined) Chlorofluorocarbons 0 2 Fuel oils 52 1 Solvents 34 22 Trihalomethanes 6 23 Table 4: VOC detections in monitoring and domestic wells. Monitoring well Domestic well ! !! ! !! !!!! ! ! ! ! ! !!! ! ! ! Figure 4: Location of sampled wells with VOC detections. The inset illustrates results for the core area of the Metro. All wells sampled in our initial reconnaissance screening for pfos/pfoa had concentrations below the reporting limits. One monitoring well, located in Golden Valley, had an estimated pfos concentration of 0.18 ug/L. The reporting limits for these chemicals (0.5 ug/L for pfos and 1.0 ug/L for pfoa) were more than 10 percent of the wq-gw2-04 Health Based Values (1 ug/L for pfos and 7 ug/L for pfoa). Since we had a partial peak for one well and because of the relatively high reporting limits, we have decided to wait until we identify a laboratory that can achieve desired reporting limits before conducting a new round of sampling in monitoring wells. 3.2. Trend Monitoring For wells with only two years of data, trend analysis is not possible. We compared concentrations of chloride, nitrate, and VOCs in wells sampled in both 2004 and 2005. Table 5 summarizes the results of these comparisons. Nitrate concentrations were lower in 18 of 21 domestic wells in 2005 compared to 2004, with no change in two wells. There was a tendency for lower nitrate concentrations in monitoring wells in 2005 compared to 2004. Not all wells included in Table 5 are part of our trend monitoring network, which explains the differences in number of wells compared to Table 1. We sampled 12 wells that were part of our St. Cloud study from 1997 through 2001. During the St. Cloud study, these wells were sampled quarterly. We found no significant difference in chemical concentration by sampling quarter. For this analysis, we used linear regression to assess trends in these wells. Table 6 summarizes p-values from the regression analysis. Although there were significant trends in several wells, we are viewing the data cautiously because we used a different lab in 2004 and 2005 than we did for the St. Cloud study and because linear regression may not be the most appropriate method for assessing trend. Domestic wells Monitoring wells Change Chloride Nitrate VOCs Chloride Nitrate VOCs Lower in 2005 2 18 3 4 10 6 No change 0 2 1 1 1 3 Higher in 2005 7 1 3 4 9 10 Table 5: Comparison of chloride and nitrate concentrations for 2005 compared to 2004. Well ID 1094 1095 1096 1097 1098 1099 1100 1101 1102 1169 1170 2533 Chloride p-value 0.027 0.723 0.128 0.583 0.497 0.572 0.688 0.001 0.863 0.121 0.162 0.824 Nitrate p-value 0.51 0.44 0.003 0.05 0.07 0.41 0.002 0.76 0.45 0.071 0.018 0.013 Comment Downward chloride trend. Land use is university campus. Upward nitrate trend. Land use is sewered residential. Upward nitrate trend. Land use is CRP. Upward nitrate trend. Land use transitioning from agriculture to residential Upward chloride trend. Land use is sewered residential. Downward nitrate trend. Land use is sewered residential. Downward nitrate trend. Land use is nonsewered residential/green space. Table 6: Summary of p-values for nitrate and chloride for the model: concentration = a + b*sampling date, where a is the intercept and b is the slope of the regression line. Data wq-gw2-04 for these wells was collected from 1997 through 2000, and then as part of our current monitoring effort. 3.3. Changing Land Use Monitoring The focus of sampling for our changing land use network is trend analysis. Trend analysis is not possible with two years of data. Regression analysis for select wells was discussed in Section 3.2. Several of the wells listed in Table 6 are part of the changing land use network. We collected samples for ICP-MS analysis in all of our changing land use wells. We also included fluoride, orthophosphate, and sulfate in our sampling. Results are not shown in this report. The concentrations of these chemicals were low, with concentrations below reporting limits for many of the metals. All concentrations were less than 10 percent of the respective water quality standard. 4. Monitoring Plan for 2006 4.1. Sampling Table 7 summarizes our sampling plan for 2006. The table shows an increase in the number of trend and changing land use samples compared to 2005. This reflects the gradual shift away from random condition monitoring and towards trend monitoring and monitoring to assess human impacts on ground water quality. Table 7 includes our target sampling goals for 2010, at which time we expect to have fully implemented our trend and changing land use networks. Eventually we expect to have a trend monitoring network consisting of 150 wells sampled for nitrate, 150 for chloride, and 100 for VOCs, in addition to 50 wells completed in areas with changing land use. Table 7 also lists Special Studies as one of the sampling components. Examples of a special study include our sampling for perchlorate in 2000 (as part of the St. Cloud study) and pfos/pfoa in 2005. One special study we anticipate participating in is a study of impacts of stormwater retention basins on temperatures in the Vermillion River. This is a cooperative study being conducted with the University of Minnesota, the City of Farmington and the developer. We anticipate involvement in other stormwater related studies. We have also had discussions about additional monitoring at feedlots, in communities serviced by large sewage treatment systems, and for chemicals that may impact surface water assessments (e.g. phosphorus). Our involvement in these studies depends on our ability to partner with other groups or agencies. The recharge study listed in Table 7 is a cooperative effort with Benilde-St. Margaret’s High School in St. Louis Park. Four monitoring wells located within four miles of the school will be sampled for chloride. The four wells occur in different land use settings – green space (open field), residential, commercial, and industrial. Each well will be equipped with devices to record water level and temperature every 24 hours. Sampling and downloading of the recorded data will occur in February, April, July, and October. wq-gw2-04 Type 2006 Goal Condition Domestic Condition Monitoring Trend Nitrate Trend VOCs Trend Chloride Changing Land Use Recharge Study Special studies Eventual Goal Condition Domestic Condition Monitoring Trend Nitrate1 Trend Vocs1 Trend Chloride1 Changing Land Use1 Special studies 1 2 Number 70 40 57 40 32 13 42 ? 25 25 150 100 150 50 ? Wells are sampled annually for five years, then biennially Four wells sampled quarterly Table 7: Sampling plan for 2006 and eventual goals for the sampling network. Condition monitoring of domestic wells in 2006 will be biased toward those areas where we have done limited sampling. These include portions of southwest, northwest, and southeast Minnesota. Figure 5 illustrates our initial domestic well selection for 2006 and the location of wells that belong to our trend network. Typically we get about 50 to 70 percent approval from well owners for sampling. In addition to areas located further from the core areas of St. Cloud, Rochester, Brainerd, and the Metro, we selected a large number of wells on lakes that are undergoing development. We have less flexibility in selecting monitoring wells for condition sampling. These wells are primarily drilled through the MPCA’s remediation programs. At these remediation sites, we select wells that are up-gradient of the known source of contamination. Currently there are about 50 sites with monitoring wells that may meet our sampling requirements. We anticipate more wells being available prior to the beginning of field season. wq-gw2-04 Figure 5: Location of domestic wells targeted for sampling in 2006. 4.2. Data Management From 2004 to the present, we have worked on development of a data management system that will ultimately make our data available via a web-based GIS application. Our data are currently managed in an Access database. The data are periodically downloaded to STORET, the national repository operated by EPA for water quality, biological, and physical data. From STORET, the data can be accessed via MPCA’s Web-based Environmental Data Access Initiative (EDA), which should be operational for ground water by the end of 2006. We initiated three pilot studies in an attempt to make all incoming data from MPCA’s remediation programs available through EDA. Through the first pilot we have begun storing individual well locations in STORET. The second pilot involved direct transfer of chemistry data from our ambient activity to STORET. Pilots 1 and 2 are complete. Currently we have about 12000 well locations loaded into STORET, plus the wq-gw2-04 chemistry data for all wells sampled as part of our ambient activity. The third pilot is intended to develop a mechanism for direct transfer to STORET of chemistry data collected through the remediation programs. This pilot is underway but has met with limited success thus far. wq-gw2-04 Appendix 1 - COLLECTION AND HANDLING OF SAMPLES FOR ANALYSIS OF PFOA, PFOS AND OTHER FLUORINATED COMPOUNDS Avoid all contact of samples with Teflon, sticky labels and adhesive tape. Sampling personnel should avoid wearing jackets and other outer clothing that is new or that hasn’t been washed a few times. Water Samples. Collect one liter of water in a narrow mouth bottle with a screw cap lid. The bottle should be made of polypropylene. Neither glass nor Teflon bottles should be used. Before use the polypropylene bottle should be well rinsed with methanol and deionized water and air dried. Once filled each bottle of water should be sealed individually in a Ziploc or whirlpak type bag and stored in a cooler at 4 ºC. Water samples (and an empty bottle and any field blanks) should be shipped to the laboratory in a cooler with sufficient blue ice paks to maintain the cool temperature. wq-gw2-04 Appendix 2 – LAND USE ASSESSMENT FOR MONITORING WELLS AND DOMESTIC WELLS SAMPLING BY THE MPCA’S AMBIENT GROUND WATER MONITORING PROGRAM The land use surrounding each well sampled by the MPCA’s Ambient Ground Water Monitoring Program is assessed in order help understand factors affecting variability in our water quality data. Using the 2003 Digital Orthorectified Quadrangles (DOQs) as a background in GIS, each sampling location is buffered based on the well type. A 300 foot buffering radius is used for monitoring wells and a half mile buffering radius is used for domestic wells. The land use within these buffers is assessed. There are four general land use categories residential, commercial/industrial, undeveloped/green space, or agriculture. For wells located in areas of mixed land uses, we estimate the approximate percent of each land use within the buffer. If the ground water flow direction is know for a well, only the land use of the upgradient half of the buffer is assessed. During sampling, field verification of the land use surrounding a well is done. Our land use estimates are further refined at this time. Residential land use in divided into sewered or non-sewered and approximate density and lot size is determined. Commercial land use is divided into sales, service, or other. Industrial is divided into manufacturing or other. Crop type and irrigation status is noted for wells in agricultural settings. wq-gw2-04